This is the unperfected transcript of proceedings on Day 18 of the Horizon High Court trial, part of the Bates and others v Post Office group litigation.
Today was the first day of Dr Robert Worden's cross-examination. Dr Worden is the independent IT expert contracted by the Post Office to look into the Horizon issues.
Transcript follows:
Tuesday, 11 June 2019
(10.30 am)
MR GREEN: My Lord, my learned friend is about to call
Dr Worden. Just before he does, your Lordship asked us
about which days were which on the transcript.
MR JUSTICE FRASER: Yes.
MR GREEN: I have shown my learned friend this because
accidentally we only got one copy, but it is Days 9, 10,
11 and 13 that are not part of this trial, and we have
set them out on --
MR JUSTICE FRASER: I'm not sure that's necessarily correct
in the sense that they are -- but that's fine. So which
day is today?
MR GREEN: Today is Day 18, my Lord.
MR JUSTICE FRASER: Day 18, all right. That's excellent.
Thank you very much. Are you going to give a copy of
that to Mr de Garr Robinson?
MR GREEN: Yes, I have shown him it already and we will give
him a copy.
MR JUSTICE FRASER: Thank you very much.
MR DE GARR ROBINSON: My Lord, I call Dr Robert Worden.
DR ROBERT PEEL WORDEN (affirmed)
Examination-in-chief by MR DE GARR ROBINSON
MR JUSTICE FRASER: Do have a seat, Dr Worden.
A. Thank you.
MR DE GARR ROBINSON: Dr Worden, I see that you have
a formidable box there which I believe contains hard
copies of your reports. You may or may not want to put
the box on the ground; that's a matter for you.
What I'm going to do is take you to your reports and
ask you a couple of questions about them. First of all,
your first report on the trial bundle is {D3/1/1}. Is
that your first report?
A. Let me just go to the very beginning. That is.
Q. If you could go to page {D3/1/260}.
A. Could you tell me what tab that is?
Q. It should be at the end of the first tab.
A. I see, it is the affirmation bit. I'm sorry, the
signature bit. Got it.
Q. Is that your signature?
A. That is my signature.
Q. Then if I could ask you -- this may not be in the
bundles -- to go to {D3/1.1/1}.
A. Sorry, how do I correlate this thing with D3? I don't
think I do.
Q. It may not be in the hard copy bundles. I'm afraid I
don't know what's in those bundles. But you will see
a list of corrections there, yes?
A. The list of corrections I think is at the front.
Q. Are these some corrections to your first report that you
identified that needed to be made?
A. Yes.
Q. And this was added to the trial bundle back in April?
A. Yes.
Q. Now let's go to your second report. That's {D3/6/1}.
A. Let me just find my way around.
Q. I suspect it will be in a different bundle but it may
not.
A. I think it is right at the back of this one actually.
Let's have a look.
Q. Gosh, that's a very big bundle, isn't it?
A. Sorry, my fault.
Q. I wonder, my Lord, in due course whether it might be
appropriate to split them up.
A. I think it is a different bundle.
Q. Yes. I see that there are other bundles in the box.
A. Sorry, I will find my way around this soon.
Supplemental report. Got it.
Q. For the trial bundle it is {D3/6/1}. Is this your
second report?
A. This is my second report.
Q. Dr Worden, I have not found your signature page or,
indeed, an expert declaration. Do you give that expert
declaration in relation to your second report?
A. I do.
Q. And there are some appendices and financial calculations
which add to or correct some of the points made, and
calculations contained in your first report, aren't
there?
A. There's a series of corrections. I mean, the
calculations made in the first report were first
corrected in the second report and then a further
correction was made with the second expert joint
statement.
Q. I see. So subject to the corrections we have just
discussed, do you believe that these two reports, their
appendices and their attached financial impact
calculations to be true?
A. I do. They are my opinion.
Q. Joint statements. A point has arisen on the joint
statements that you and Mr Coyne have agreed. I'm told
there is an amendment that has been agreed to
paragraph 11.1 of the fourth joint statement. Is that
right?
A. That is correct.
Q. Do you have a copy of the amendment that has been
agreed?
A. Well, I had a few minutes ago when I signed it. I'm not
sure where it is now. It is probably in one of these,
isn't it? The problem is it is not amended in this one.
Q. Could you tell me what you did with the piece of paper
you signed?
A. I gave it to my assistant who gave it, I think, to
somebody else.
Q. Do we have a copy of the signed version of the amended
JS4?
MR GREEN: We have got copies if that helps.
MR DE GARR ROBINSON: Thank you. If that could be handed to
Dr Worden.
MR JUSTICE FRASER: I would like a copy, please. Is there
only one?
MR GREEN: My Lord, I think this is the one that's signed.
MR JUSTICE FRASER: Give it to the witness first.
MR GREEN: We will produce copies.
A. I think two were signed. I signed two.
MR DE GARR ROBINSON: That's the fourth statement which --
for the trial bundle, the current version of which is in
{D1/5/1}. I believe the amendment may be at page
{D1/5/10}. Is that right?
A. The amendment is 11.1, which is at page 10.
Q. Could you read out what the original version says and
what the changed version now says.
A. Do you want the full statement each time?
Q. It's just a sentence, isn't it, in 11.1?
A. Yes, it is a longish sentence. Start with the original:
"Evidence from several Peaks indicates that whenever
Fujitsu needed to make any change to data which impacted
branch accounts, they were concerned to seek permission
from PO to do so, and to ensure that PO took
responsibility for the resulting change."
{D1/5/10}
Q. Right. And what is the change that you have agreed?
A. The change that Mr Coyne and I agreed out there the
other day was that the word "whenever" is replaced by
"usually when".
Q. Just to explain, could you explain what you mean -- I'm
not asking you to delve into Mr Coyne's head -- by
"usually when"?
A. Yes. I mean, I agreed to the replacement of "whenever"
because "whenever" implies that the opposite never
happens, and I felt I didn't know strongly enough to say
that categorical statement. So "usually when" to me
means the great majority of cases, but I can't say with
my hand on my heart all cases.
MR DE GARR ROBINSON: Very good.
My Lord, I have no further questions.
MR JUSTICE FRASER: Just before -- and it might not be a big
point but it might be, the note that I made of when this
point was put to Mr Coyne was that the change was to
"usually" rather than "usually when". But your
understanding --
A. I noted that problem. I didn't say anything about it at
the time.
MR JUSTICE FRASER: This is last week?
A. Yes. I noted people said "usually" instead of
"whenever". I thought that is a problem, but I didn't
actually say anything.
MR JUSTICE FRASER: Right. But the change that you consider
should be made is that the single word "whenever" is
changed to "usually when"?
A. That I believe is the words we agreed.
MR DE GARR ROBINSON: That has been agreed, my Lord.
MR JUSTICE FRASER: The point that was put to Mr Coyne that
was agreed though was that the change was from
"whenever" to the word "usually".
MR DE GARR ROBINSON: Well, your Lordship will remember that
it was made clear that at that stage the actual wording
hadn't been --
MR JUSTICE FRASER: That's why I said it might not make
a difference, but my note of what was put was that the
change was to "usually". But there's now a signed
statement that has "usually when" on it; is that right?
MR DE GARR ROBINSON: Yes.
MR JUSTICE FRASER: Is that what the witness has been given?
MR DE GARR ROBINSON: Yes, my Lord.
MR JUSTICE FRASER: Can I have either a copy of that or that
at some point.
MR DE GARR ROBINSON: Yes.
MR JUSTICE FRASER: Thank you very much,
Mr de Garr Robinson.
Mr Green.
Cross-examination by MR GREEN
MR GREEN: Dr Worden, can we look at your CV, please. It
will come up on the screen. So when I call
a document --
A. Fine.
Q. It is {D3/4/1}.
A. Is it in this bundle, by the way? It should be,
I guess. Anyway, there we are.
MR JUSTICE FRASER: Just for the cross-examination generally
if you would like to see the hard copy in front of you
of whatever is on the screen please just say so.
A. Thank you, my Lord.
MR JUSTICE FRASER: You will be given the reference.
Right, Mr Green.
A. Here we are.
Q. So we have got your strong practical experience
explained there?
A. Yes.
Q. You have a PhD in theoretical particle physics?
A. Yes.
Q. That is right. And you have listed a number of
important and quite high value IT cases in which you
have been an expert?
A. That is right.
Q. You have a wealth of experience of being an expert?
A. I have. I have tried over the past 15 years to spend
not more than half of my time being an expert.
Q. I understand. Do you have any particular qualifications
in statistical techniques?
A. Well, I'm trained as a scientist and engineer, and as
such I have been using mathematics all my life. And
a part of that mathematics is probability theory and
statistics. I have a PhD which applies that
mathematical extensively. I have not any formal
qualification in statistics, but I don't regard
statistics as distinct from the other body of my
mathematical knowledge.
Q. Indeed if we look at page {D3/4/2} of that document --
it is just coming up now. There will sometimes be
a slight delay while the Opus operator loads the
page up.
A. Yes sure.
Q. -- if you look at the middle of the page you will see
July 1997 to July 2010 in bold. Do you see that?
A. Yes, right.
Q. The first paragraph begins:
"Acted for WPL ..."
Do you see that?
A. Yes.
Q. If you go to the right-hand side of that paragraph which
ends with "June 2010" and you come down two lines, you
see:
"Robert gave evidence on software development
methods, advanced statistical techniques ..."
Do you see that?
A. Yes.
Q. The phrase "advanced statistical techniques" reflects
the application of the expertise that you have just been
talking about?
A. Absolutely. I believe that the statistics I have
applied as a scientist particularly but also as
an engineer goes to the point of advanced statistical
techniques.
Q. And in statistics there are some matters which are
questions of approach, how should we go about looking at
something?
A. There are indeed.
Q. And there are other matters which are maths, which are
either right or wrong?
A. Yes, there is a lot of strain mathematics which dates
back to a long time.
Q. And you felt completely comfortable giving your evidence
on statistics and probability in this case?
A. Yes. I should say SAS is a statistics package, so it
has an advanced statistical method built into it and
I felt thoroughly comfortable examining those methods
and so on.
Q. Yes. And was it you who was doing the statistics
mostly, or Mr Emery, or a team effort?
A. No, Mr Emery was not involved in this dispute, it was
just solely me. Sorry, in this dispute I did the maths.
Q. You did the maths?
A. In SAS v WPL it was only me.
Q. I understand. But in this dispute, you and Mr Emery
worked together?
A. That is right.
Q. You did the maths?
A. Yes.
Q. Did Mr Emery check the maths?
A. No, he didn't actually. The maths is there in the
spreadsheets for anybody to check. It is not advanced
statistics, it is basic multiplication and division, and
it is laid out in the spreadsheets.
Q. Did you check it yourself?
A. Absolutely.
Q. Probably several times given you have checked for other
errors and made a correction?
A. Well, the correction was not a correction to the maths,
the correction was adjustments to the input assumptions.
Q. Let's just pause. I want to take you to one correction
you have made. It is at {D3/1.1/1}. This is
a corrections document to your report, isn't it?
A. Well, I haven't got -- I mean, the screen is not showing
me anything I can see is relevant at the moment.
Q. No, I'm just identifying the documents.
A. Right. This is corrections to my first report, yes.
Q. It says there just underneath what we call the tramlines
where it says "Corrections to Dr Worden's Expert
Report", it says you have identified some minor
corrections and clarifications?
A. Yes.
Q. If we go to page {D3/1.1/2} and if we look down at 761
towards the bottom, you say:
"the chances of the bug occurring in a Claimant's
branch would be about 2 in 10 million"?
A. Yes.
Q. And it should be:
"the chances of the bug occurring in a Claimants'
branch would be about 2 in a million."
A. That is right.
Q. So what was stated there was out by a factor of 10, and
that's why you've corrected it?
A. A correction had to be made. Yes, that is right.
Q. Did you notice any other errors?
A. No, I didn't because there were quite a few statements
of this "in a million" nature in my report. I made
careful checks of the main thread that led to my key
result in section 8.7, and that was my focus. And
similarly on the transaction corrections I had made
checks to that spreadsheet and, you know, I included the
spreadsheets with the reports so that you can see how
the arithmetic is done from the spreadsheets.
Q. And some of it you point out is basic arithmetic?
A. I think all of it is basic arithmetic.
Q. I'm very grateful.
Now, have you been involved in any group litigation
as an expert?
A. No, this is a new experience for me.
Q. I understand.
A. I think I haven't. No, I'm pretty sure I haven't.
Q. It is right, isn't it, you are not a behavioural
economist by training?
A. Absolutely, I'm not a behavioural economist.
Q. You know who Richard Thaler is?
A. No, I don't.
Q. He won the Nobel prize for economics in 2017 for his
work in behavioural economics. That is not your field
of expertise?
A. No, it is not my field of expertise. My knowledge of
economics is shallow and really my knowledge is much
more about business finance because I have been involved
in managing parts of various businesses.
Q. But you have referred in your report -- if we look at
{D3/1/107} and if we look at paragraph 415 there.
A. Yes.
Q. You have said there:
"I can go a little further than this ..."
This is referring to approaches to challenging
discrepancies:
" ... by making weak inferences about how a manager
of a small business, such as a Subpostmaster, needs to
prioritise his time in monthly balancing and other
evidence."
A. Yes.
Q. "The Second Witness Statement of Ms Angela Van Den
Bogerd at paragraph 187 says: 'Generally, when
discrepancies are of a value of several hundreds of
pounds, I would expect Subpostmasters to contact NBSC.'"
A. Yes.
Q. Now, what then follows is this statement:
"If this is correct, it is consistent with my
estimates."
That suggests you have made your own estimate of the
value at which a subpostmaster would contact NBSC?
A. Yes, I have made my own estimates, and you can see they
are rather more detailed than Mrs Van Den Bogard's
witness statement. In other words, she's put a point on
the spectrum. I have made some estimates about what the
spectrum is.
Q. Can we just be careful with what we are talking about,
because it does matter. You understand, particularly
with your background as a scientist, the difference
between an estimate and an assumption, don't you?
A. Absolutely. I think I do, yes.
Q. It is pretty fundamental. An estimate is based on some
facts --
A. I have made a number of -- well, have I been loose in
terminology between --
Q. Dr Worden, at present I'm just trying to clarify with
you what you have and have not done. I'm not going to
tie you to the words, but I'm going to show you the
words you use to clarify what you have in fact done or
not done.
A. "I assume the following, as best --"
Q. Dr Worden, would you allow me to take you through it?
A. I'm sorry.
Q. It will just help you to identify to show what you I'm
asking --
A. I'm just reminding myself that it is actually
assumptions I'm making in this paragraph.
Q. That's what I'm asking you, because there are a number
of ways of reading that paragraph. Because after the
quote of Mrs Van Den Bogard's witness statement, which
is in bold and ends with the words "paragraph 187)",
there's then this sentence:
"If this is correct, it is consistent with my
estimates."
Then there is a separate sentence which says:
"I assume the following, as best assumptions ..."
Now, the first question for you is this: when you
use the phrase, which we have also seen elsewhere, or
similar, "if this is correct, it is consistent with my
estimates", now, the use of the word "estimate" there,
I think we have agreed, tends to suggest something
different to "assumption", doesn't it?
A. Absolutely, yes.
Q. So can you please just tell the court did you make your
own free-standing estimate without regard to Angela Van
Den Bogard's witness statement of the levels at which
a subpostmaster would or would not challenge
a discrepancy by ringing the NBSC?
A. To clarify wordings, in my reports --
Q. Could you answer the question first?
A. In my report I made various estimates following from
an assumption. In other words, I made assumptions to
put into the maths which then constituted my estimates.
So that's the distinction in my mind between assumption
and estimate. And when I say it is consistent with my
estimates, it is consistent with estimates following
from my assumptions.
Q. Right. To clarify that sentence I'm asking you about,
where it says "if this is correct, it is consistent with
my estimates", you are not saying that you have any
independent basis upon which to support what Angela Van
Den Bogard says, are you? You didn't make any estimates
that confirm her witness evidence?
A. Well, there is a process you use in engineering of
ballpark estimates, which are imprecise -- or ballpark
assumptions -- I'm sorry if I get confused about the
wording here -- which are very approximate assumptions
you make in order to drive a calculation through. And
then you look back at the calculation and you say: is
the precision I have achieved in that calculation
sufficient or do I need to refine/revisit some
assumptions?
So you are always looking at the precision you need
to achieve in the result when you are concerned with how
precise should my assumptions or my estimates be. How
precise should my assumptions be in order to drive this
set of estimates which arrives at a number. How precise
do I need that number to be? I felt that given the
precision I needed or the court needed in the final
result, the precision of these assumptions, you know,
whether it is £300 or £400, for instance, was
sufficient.
Q. I will just take you back to the question for the
moment. We understand you made assumptions.
A. Yes.
Q. And you have set them out.
A. Yes.
Q. And you have said they are assumptions.
A. That is right.
Q. But you didn't make any fact-based estimate of what
would in fact happen yourself, did you?
A. I didn't go and interview a postmaster, for instance, or
anything like that. They were assumptions and they are
advertised as assumptions, and so if the court wishes to
change the assumptions and drive them through the
calculation, for instance, the calculations were
intended to have that ability, so that if the court --
I make some assumptions, if the court decides to find
something different then the court can drive those
different findings through my calculations.
Q. Okay. I will put it one last time. The answer is no,
isn't it, you didn't make --
A. I think I said no. Yes, it is no. I didn't make
a fact-based --
Q. Yes. To do so you would have needed the relevant fact
background, wouldn't you, to make an assessment of that?
You would have had to understand the process, what sort
of headwinds people faced when they sought to report
things, how they valued their own time etc?
A. To get precise numbers --
Q. To make an estimate not an assumption?
A. To get precise numbers I would have had to do that.
Q. Or any number that was an estimate rather than
an assumption, you would need some facts upon which to
make an estimate, wouldn't you?
A. Well, as I say, my view of the word "estimate" is
an estimate is what I made following my assumptions.
Q. No, Dr Worden --
A. Is that unclear? That I made some assumptions and put
them into some maths and that resulted in some
estimates.
Q. So the estimates that you refer to are the outcome of
having made assumptions?
A. That is right.
Q. I understand. And you accept that you had neither the
facts nor the relevant expertise to make estimates which
effectively fall within the field of behavioural
economics?
A. I accept that entirely.
Q. Thank you.
A. I won't say more.
Q. Thank you.
Just stepping back a minute, you have got a PhD in
theoretical particle physics, yes?
A. Mm.
Q. You have this wealth of experience as acting as
an expert and so forth?
A. Mm.
Q. You have made very detailed statistical calculations in
the report?
A. I wouldn't call them very detailed. I think you know
there are a number of multiplications there and
divisions and so on, and where the more detailed
statistics came in was estimating what sort of sample
size I would need to get more precision in the result.
Q. I understand. So you went into more detail in those
areas?
A. Occasional footnotes addressed these questions, yes.
Q. And you spotted the correction of the typo when you went
through your report?
A. Yes.
Q. Order of magnitude of 10?
A. Yes.
Q. Would you say, Dr Worden, you are a details man?
A. Physics people tend to focus; they don't like loads and
loads of facts, they like to focus on the core analysis
and make sure that's very right.
Q. I understand.
A. So they are detailed people in that sense.
Q. I understand. When we look at Mr Emery at {D3/5/1}, we
have got his CV.
A. Yes.
Q. And we can see he has obviously had a lot of IT
experience over many years. And if you go over the
page, please, {D3/5/2} we can see that he has a degree
in computing science from Imperial College?
A. That is right.
Q. And if we go to page {D3/5/5}, we can see where he
worked and over what period.
A. Yes.
Q. He was obviously at Logica for many, many years,
wasn't he?
A. That is right. We worked together at Logica in 1976, or
something like that.
Q. I understand. Is that how you have ended up together?
A. Well, obviously, you can see from his resume he has been
at Logica for a fair old time and he's been at
Charteris, and Charteris was a kind of spin-off Logica,
and I've been there. So we've overlapped quite a lot.
Q. Does Charteris still do some consulting for Logica?
A. Not for Logica. Logica doesn't exist anymore.
Q. When did it cease to exist?
A. It was taken over by CGI, I think, and ceased to exist
as a name about ten years ago, I think.
Q. What I really meant was I think at that point quite
a lot of people left?
A. Well, there have been various watersheds, yes.
Q. Was Credence originally a Logica solution?
A. That's a very good question. I haven't really looked
into that. Logica obviously had some role to do with
Credence but certainly I was nothing to do with
Logica at the time.
Q. Had you had experience of data warehousing when you were
working for Logica?
A. Yes, we -- a team that worked for me was involved in
business intelligence and that involved data
warehousing. And generally I have been doing databases
since the year dot and that includes data warehouses.
Q. Because data warehousing was quite big in the late
1990s, wasn't it?
A. It has been big since then, yes.
Q. In fact, Logica was quite strong in that area in the
early 2000s?
A. Well, I had left them by that point.
Q. I understand. That is right though, isn't it?
A. Logica was a bit of a jack-of-all-trades, actually.
Q. Okay. And just going back to Mr Emery, in your
experience of Mr Emery, I don't want to be unfair, but
is he a details man?
A. I think he is more than me, yes.
Q. And are there any parts of the expert report that he has
not checked? I think you suggested earlier he might not
have checked the maths?
A. That is right.
Q. Is that right?
A. I don't think he has checked the maths.
Q. But otherwise would he have read through --
A. He certainly read through the whole report and the
appendices.
Q. And you have probably read through it several times as
well?
A. I'm afraid so.
Q. I'm afraid so. How many times?
A. How many times I read through my report? Yesterday
and -- about five times, I suppose.
Q. Five times. So with your eyes on it and Mr Emery's eyes
on it, we have four eyes, haven't we?
A. Yes.
Q. Probably other people who I want ask you about have
looked at it as well, lots, and some of the people who
have looked at it looked at it many times, yes?
A. Well, I have looked at it many times, Mr Emery has
looked at it many times and I suspect the lawyers have
as well.
Q. Can I ask you about how you have approached the
assumptions upon which you have given your opinions in
your report.
Could we look, please, at page {D3/1/153}. We are
going to look, if we may, at paragraph 650. Do you have
that?
A. Yes. "Receipts/Payments Mismatch".
Q. Now, this is looking at the receipts and payments
mismatch bug at the bottom of page 153.
A. Mm.
Q. You see the heading there "8.6.1"?
A. Yes.
Q. And you say:
"This issue is cited in paragraph 5.6 of Mr Coyne's
report."
A. Mm.
Q. "It involved a bug in Horizon which was triggered by a
rare circumstance (which one would not expect to be
exercised in testing) and which had an effect on branch
accounts."
A. Yes.
Q. "If Mr Godeseth's evidence about this bug is not
accepted, I shall revise my opinions accordingly. They
are based on written evidence - particularly on a
written analysis by Gareth Jenkins ..."
Yes?
A. Yes.
Q. "... as well as the Second Witness Statement of
Mr Godeseth."
A. Yes.
Q. You make a consistency point there. Can we just go back
now you have seen the whole paragraph?
A. Yes.
Q. I just want to ask you about three lines down in that
paragraph, 650, on the right-hand side the following
sentence:
"If Mr Godeseth's evidence about this bug is not
accepted, I shall revise my opinions accordingly."
Now, pausing there. Let's just only look at this
example in your report. Let's just focus on this.
In this example what you have done is premised the
opinion that you give on the court accepting
Mr Godeseth's evidence.
A. That is right.
Q. But you have fairly said that if the court does not
accept that evidence "I will revise my opinions
accordingly."
A. Yes.
Q. The content of this report in this respect, we will come
to other examples later, focuses on what the
consequences would be if the defendant's factual
evidence is accepted?
A. Generally, if findings -- if things happen in oral
evidence which go against witness statements then
I would need to come back and say what's the impact.
Q. Yes. You understand, don't you, that the moment at
which everyone finds out whether factual evidence is
accepted or not is when the judgment is handed down?
A. Absolutely, yes.
Q. So you are proposing to revise your opinions after that,
are you?
A. No. That's a good point. What I was saying there is if
it becomes evident, all I can do is make assumptions.
The court will find findings and all I can do is make
assumptions and drive them through my opinions and try
and assist the court that way.
So all of this is assumptions I have made based on
the evidence I have seen, and the court may find
differently.
Q. Can I pause there. You were assuming there that
Mr Godeseth's version was true?
A. I was assuming mainly Mr Jenkins' written analysis which
Mr Godeseth's evidence confirmed.
Q. Let's just go with that for the moment. Have a look, if
you would, please, at {D3/1/260}. It is paragraph 1194.
A. Here we go, yes.
Q. You recognise this page because you were shown it --
A. Yes.
Q. Now, come down four lines and look at the right-hand
side about three-quarters of the way across.
A. Yes.
Q. You see "I have not assumed that any particular version
of events is true ..."
A. Yes.
Q. Yes? Now, what you say, if we look at the transcript,
please, at the bottom of page 26 of today's transcript,
just a few moments ago --
A. Can I get that? Yes.
Q. See the bottom of 26, the question is:
"So you are proposing to revise your opinions after
that, are you?"
And you say?
"Answer: --
A. Sorry, I haven't got the right line.
Q. It is the bottom of page 26, halfway down, line 23. Do
you have that?
A. Yes, I have got that.
Q. Dr Worden, if I'm taking you too fast at any point let
me know --
A. No, it is just me finding my way, that is all.
Q. Not at all, it is not always straightforward.
Line 23:
"Question: So you are proposing to revise your
opinions after that, are you?
Your answer is:
"Answer: No. That's a good point. What I was
saying there is if it becomes evident, all I can do is
make assumptions. The court will [make] findings and
all I can do is make assumptions and drive them through
my opinions and try and assist the court that way.
"So all of this is assumptions I have made based on
the evidence I have seen, and the court may find
differently."
A. Yes.
Q. Do you feel that overall in your report you have
provided opinions on the footing that the claimants'
evidence is accepted as often as you have provided
opinions on the footing that the defendant's evidence is
accepted?
A. Well, my analysis of the claimants' evidence is mainly
contained in my supplemental report, and I explained
there that individual claimants' evidence, particularly
individual subpostmaster evidence, I did not feel able
to make strong use of that and I gave the reasons in my
supplemental report.
So my opinions have little dependence on that. And
the core of my opinions, the numerical estimates I make,
those estimates have been designed -- or the process and
the method has been designed so that if -- so that to
make my assumptions evident where those assumptions come
in, and if the court finds something different from my
assumptions the court can go to the spreadsheets and
re-do the method for itself.
So I have tried to make the dependence of what the
court may take from my reports as little dependent on my
assumptions as possible. I'm not sure if I'm answering
the question.
Q. Let me give you another example. This is a slightly
different situation; let's look at that as well. It is
on page {D3/1/206}. It is where you are dealing with
transaction corrections.
A. Yes.
Q. They were set out, you understood, in Mr Smith's witness
statement?
A. Evidence was in that witness statement which I thought
bore on the issue of erroneous transaction corrections,
yes.
Q. Let's have a look. If we go over the page, please,
{D3/1/207} we can see that you say, 934:
"I proceed on the assumption that these figures
(which are the only ones available to me) are accepted
by the court. If they are not, a different calculation
along the same lines may possibly be appropriate."
A. Yes.
Q. We will come back to what available evidence there was
about TCs later. For now, you will accept that again is
proceeding on the footing that the evidence of the
defendant will be accepted?
A. Yes, I don't see that evidence of the defendants is very
relevant to this calculation.
Q. I understand.
Now, you have been in court, haven't you, for the
whole of the trial, I think?
A. Not the whole trial.
Q. Did you miss any days?
A. I think I must have missed some days, yes.
Q. Did you hear the defendant's evidence?
A. I certainly heard Mr Rolls' evidence.
Q. No, the defendant's evidence.
A. Sorry, the defendant's --
Q. I will say Post Office.
A. I heard all of that, yes.
Q. You heard all of that?
A. Pretty sure I did.
Q. Have you had any changes of heart about anything in your
report having heard it?
A. Changes of heart?
Q. Anything you want to change about what you said?
A. Well, had there been some fundamental change I would
have felt obliged to communicate it to the court.
Q. Of course, because you take your duty very seriously?
A. Yes, I do. I can't think of any major change of heart,
but there may well be things that you bring me to and
I say there's an adjustment here. But to come back to
the point about precision, I am conscious of how precise
my numerical results have to be in order to be of
assistance to the court.
Q. Well, that's appreciated.
A. And therefore that level of precision is my yardstick
for saying: has my opinion changed? Does the court need
to be informed about it? And so on and so forth.
Q. I see. Let's look at some particular facets of your
evidence. Let's look, if we may, please, at {D3/1/239},
which is paragraph 1086 in your report. Do you have the
page there? It is the bottom paragraph.
A. Yes.
Q. You say, second line, do you have halfway through the
second line the words "when Post Office"?
A. Yes.
Q. "When Post Office is investigating anomalies reported by
subpostmasters, they use Credence and their other
management information systems in the first instance
..."
Just pausing there, it is pretty important to
identify with care and clarity what systems and
information are available to Post Office, isn't it, in
this case because it is one of the issues?
A. As far as the experts can find out it is important, yes.
Q. Exactly, so far as the experts can find out. We will
come back to that in a minute, but let's just focus on
this.
So you agree it is important to identify what
information and systems is available as far as the
experts can find out. Just recapping on that sentence
in the third line of that paragraph, so just remind
yourself what I was asking you:
" ... they use Credence and their other management
information systems in the first instance ..."
Yes?
A. Yes.
Q. There is a footnote there, footnote 41?
A. Yes.
Q. There you say:
"The Witness Statement of Ms Tracy Mather,
16 November 2018, is consistent with my understanding."
A. Yes.
Q. Pausing there, did you get your understanding materially
from her witness statement or did you form
an understanding from something else that was consistent
with her witness statement?
A. I formed an understanding from something else, which is
largely my understanding of things like POLSAP, things
like Credence and my experience of how --
Q. And the opportunities you have had to look at that since
you were instructed?
A. To look at what?
Q. To look at the overall system, documents and so forth?
A. Yes, that was combined with my experience of how
businesses worked.
Q. Of course. So let's look at what Ms Mather said in that
witness statement to which you have referred.
It is at {E2/8/3} and it is paragraph 12.
A. This is not one I have looked at very recently.
Q. Don't worry, I will take you through what it says. It
is a short paragraph.
A. (Reads to himself) Yes.
Q. Let's take it is this stages, firstly as an information
tool, I think that's agreed?
A. Yes, it is also called POLMIS in certain documents.
Q. Yes. That's the same ...?
A. My understanding is that POLMIS and Credence are
different names at different times for the same system.
Q. Same system. Because "POLMIS" stands for Post Office
Limited Management Information System?
A. That is right.
Q. And if we look at the second sentence, it says it is
designed to work alongside other applications?
A. Yes.
Q. So there are a number of different things they can
look at?
A. Yes.
Q. "It is used to help understand what has happened in
a branch as it records all keystroke activity performed
in that branch by the user ID, date and time ..."
Do you see that?
A. Yes.
Q. That was your understanding when you wrote your report,
wasn't it?
A. Yes.
Q. Had you formed that understanding from any other
documents that you had seen or was that understanding
only formed from reading Ms Mather's witness statement?
A. No, I had formed it from other documents.
Q. Can you remember what other documents you had seen which
showed that Credence recorded keystrokes in the branch?
A. To that level of keystrokes in the branch -- no, to go
into that level of -- not level of misunderstanding,
level of understanding, to go to that level of
understanding I think Mrs Mather's witness statement was
useful to me. But the point that Credence is an MIS and
Post Office use it to determine what happened in the
branch and so on, those points were evident to me from
my reading of quite a lot of evidence.
Q. But you will forgive me, Dr Worden, if I ask you the
precise question again, which is this: in relation to
Credence recording "all keystroke activity performed in
that branch", was that your understanding when you wrote
your report or not?
A. I did not go as far as understanding all keystroke
activity. In other words, I felt that Credence was
a fairly comprehensive record of things that had gone on
in the branch, but the phrase "all keystroke activity"
would not have occurred to me before her witness
statement.
Q. So --
A. Because that is a very detailed level of information,
and my experience of MIS is that you are taking all
sorts of slices of information, and really drilling down
to the keystroke is very detailed.
Q. But you would accept, Dr Worden, that whether factually
you could look at the keystrokes' activity performed in
a branch or whether you could not, as Post Office, do
that, would be important, wouldn't it?
A. No, I do not think so. For the purpose of knowing that
Post Office could investigate events in the branch using
their MIS, I did not need to know that it went down to
the keystroke level.
Q. Why not?
A. Because MISs typically deal with more summary
information and they are designed based on the
requirements, and if the requirement doesn't require to
go to keystroke level that's not a part of the MIS. And
I didn't feel myself that going to keystroke level was
necessary for the level of investigation that I believed
needed to be done in Post Office.
Q. So you made a judgment about what needed to be done by
way of investigation and you made a judgment that
keystrokes were not necessary to the level of
investigation that you formed a judgment about?
A. Yes. To me, keystroke level is very, very detailed, and
from my experience of MISs, keystroke level is not often
required in an MIS. An MIS typically deals with more
summary information and ways of getting different slices
of information, and so on and so forth. Going down to
the keystroke level is, to my mind, very detailed.
Q. We will come back to this page in a moment, but can we
just go, please, to {C1/1/2}.
The document you are now looking at is not the
document that was on the screen before.
A. No, this is --
Q. These are the Horizon Issues which you probably
recognise?
A. Yes, that is right. Yes.
Q. And if you look halfway down you will see "Operation of
Horizon" in capitals and in bold?
A. Yes.
Q. Under that, there is one subheading, "Remote Access",
and there is another heading, "Availability of
Information and Report Writing". Do you see that?
A. Yes.
Q. Under that, there is issue 8, and issue 8 says:
"What transaction data and reporting functions were
available through Horizon to Post Office for identifying
the occurrence of alleged shortfalls and the causes of
alleged shortfalls in branches ..."
Yes?
A. Yes.
Q. So it was a specific Horizon Issue that you were
instructed to report on?
A. Yes.
Q. To identify what transaction data and reporting
functions were available through Horizon to Post Office?
A. Yes.
Q. And the reason that you weren't interested in whether or
not keystrokes were available was the one you have just
given to the court a moment ago: that you formed
a judgment about what would be necessary, and on the
basis of that judgment you formed a judgment that
keystrokes wouldn't be important?
A. Well, broadly that is correct. I should say also
there's a kind of prioritisation that the expert has to
do, how far he can drill down in particular areas and
what drilling down is used for.
Q. And you felt, if we go back to Ms Mather's report -- it
is {E2/8/3}, it is a short paragraph and it makes quite
a key point about keystroke activity, doesn't it?
A. It does.
Q. And she positively said that it records all keystroke
activity in that branch by the user?
A. Yes.
Q. And when I started asking about this it did sound as if
you were saying you had separately formed the view from
other documents that that was possible?
A. No, I had said I had separately formed a view from other
documents that Credence was used to investigate what
happened in the branch.
Q. Right. Let's go back, please, to the paragraph
I originally asked you about, which is {D3/1/239}. It
is paragraph 1086. You will remember it is at the
bottom of the page.
I asked you about the second sentence of that
paragraph:
"When Post Office is investigating anomalies
reported by Subpostmasters, they use Credence and their
other management information systems in the first
instance ..."
Yes?
A. Yes.
Q. You give the footnote at 41?
A. Yes.
Q. And you say:
"The Witness Statement of Ms Tracy Mather ... is
consistent with my understanding."
Yes?
A. Yes.
Q. And that's the witness statement we have just been
looking at.
A. Yes.
Q. Which specifically says that Credence records all
keystroke activity.
A. Yes.
Q. So you accept that this would suggest to anyone reading
it, whether it is the court or the claimants or anyone
else, that you agreed with her description of what
Credence could do?
A. No, that's not quite what I said.
Q. You say it is consistent --
A. Her description is consistent with my understanding and
went beyond it.
Q. You didn't spell out anything -- should we read
throughout your report where you say "consist within the
my understanding" as meaning "there is some aspect in
which my understanding coincides with what I'm referring
to"?
A. I think we read "consistent" in the ordinary sense of
"consistent".
Q. I see. But on the face of Mrs Mather's statement you
would agree, wouldn't you, that Post Office had access
to information that the SPM doesn't have if she is
right?
A. Well, whether or not she is right, Post Office has
information that the branch doesn't have.
Q. Of course we will come to that in more detail. I'm just
focusing on the keystrokes.
If she is right about the keystrokes being available
to Post Office, they have effectively a record of what
buttons were pressed in the branch?
A. Yes.
Q. And the SPM doesn't have them?
A. I believe that is right, yes.
Q. And so you were here I think when Mrs Van Den Bogard
gave her evidence, weren't you?
A. I certainly read the transcripts, but --
Q. Let's refresh your memory. Sorry, I interrupted you,
Dr Worden.
A. I do not think I was actually in court when Mrs Van Den
Bogard was --
Q. Okay, let me show you the transcript then, to be fair to
you. {Day6/69:1}, please.
(Pause)
We are going to look at lines 10 to 23. There is
a question:
"Question: ... the point is this, that the Credence
report would show the specific keystrokes by the
operator in branch, wouldn't they?"
Do you see that?
A. And she says yes.
Q. She says yes, and then:
"... so if one wanted to get to the truth about what
keystrokes Mr Patny had pressed, one could have obtained
a Credence report to identify that and that would
have --
"Answer: Have used a Credence report, yes.
"Question: And that would have shown all the
keystrokes pressed --
"Answer: Yes.
"Question: -- so you could follow exactly what he
had done in sequence?
"Answer: Yes."
Do you see that?
A. Yes.
Q. Were you here for the evidence of Mr Patny?
A. I do not think I was.
Q. Where we looked at the internal Post Office documents
where he had been promised a Credence report and it
didn't seem to come?
A. I do not think I was here for that evidence, no.
Q. Now if we look, please, at page {Day6/104:1} because
Mrs Van Den Bogard was re-examined about that.
(Pause)
We are going to look on this page at lines 2 to 10.
A. 104, 2 to 10.
Q. Just to orientate you, this is my learned friend
Mr de Garr Robinson, leading counsel for the
Post Office, re-examining Mrs Van Den Bogard about the
section of cross-examination I have just shown you. Do
you understand?
A. Yes.
Q. Okay.
"Question: Now, the reason why I'm putting this to
you is because Mr Green put to you that Credence shows
every single keystroke that's pressed by the postmaster
and you accepted what he said and he put it to you on
the basis that this is what Mrs Mather said. Could I
ask you this question: is it the case that Credence
gives an account of every single key that's pressed by
the postmaster in branch?
"Answer: That's not my understanding, not every
single stroke.
"Question: I'm grateful."
Do you see that?
A. Yes.
Q. So in re-examination there was a different picture.
A. Yes.
Q. But I think you said you were reading the transcripts.
Did you notice this?
A. I didn't notice this. I'm afraid I didn't, no. I can't
say I have read every line of the transcript because it
would take a day to do it for each --
Q. I understand.
Okay. Let's look at page {Day6/149:1}.
A. Right.
Q. Because Mrs Mather was the person whose witness
statement you had actually positively referred to,
wasn't she?
A. By me? Yes.
Q. Yes, it is in your report. You say look at that --
A. She's consistent with my understanding, yes.
Q. Yes. And if we look on page 149, look at lines 7 to 20.
Again, this is my learned friend, leading counsel for
Post Office, re-examination Mrs Mather this time.
Sorry, examining Mrs Mather in chief. That's before she
is cross-examined.
A. Right, okay.
Q. Yes?
A. Mm.
"Question: And now, Mrs Mather, I have to ask a
question which is born out of sheepishness on my part.
Everyone else in court will understand why, but you
won't. I would like to ask you about paragraph 12 of
your statement please ..."
That is the one we have just been looking at.
A. So that is --
Q. We saw what it said. It was perfectly clear, wasn't it?
A. The keystroke --
Q. Yes, that is right.
"Question: Here you are describing Credence and you
say:
"'Credence is used as an information tool.
It is designed to work alongside other
applications. It is used to help understand
what has happened in a branch as it records all
keystroke activity performed in that branch by
the user ID, date and time.'
"Could I ask you to explain what you mean
by the phrase 'It records all keystroke
activity'?"
And she says:
"Answer: What I actually meant was the
transactional data as in sales and non-sales."
A. Right.
Q. So that was a bit of a surprise.
A. Yes. I mean --
Q. Hence your reaction by saying the intonation of the word
"right" in your answer which doesn't always come out on
the transcript?
A. Yes. This, if you like, Mrs Mather's witness statement,
down to keystroke, that was consistent with my
understanding but it went beyond it. And now what's
happening here is it doesn't go beyond my understanding
as far as it did before.
Q. So it recedes back to what your unspecified --
A. It comes closer to what I would have expected from
an MIS.
Q. I don't want to be unfair, but were you sort of guessing
based on experience about what an MIS would have?
A. Yes, I have lots of experience of what an MIS has.
Q. So that is the understanding you are referring to?
A. That is right, yes.
Q. So --
A. Also what I can infer from the documents I have seen,
you know. I put those together with my experience and
one imagines what Post Office would need in order to
investigate what went on at the branch, and my
understanding is they would need what Mrs Mather says
here, transactional data. They would not need
individual keystrokes. That's my -- what my experience
tells me.
Q. So is it right, then, that you hadn't appreciated
whether or not keystrokes were recorded in branch either
way despite referring expressly to Mrs Mather's
statement when you made your first --
A. I said Mrs Mather's statement was consistent with my
understanding, and it was, and it went beyond my
understanding. And now we have seen that the extent to
which it went beyond my understanding varied at the
time.
Q. It is fair to say that the finesse that you are now
making was not expressly explained in your first report,
was it?
A. What I said in my first report was it is consistent with
my understanding, and it was and it is.
Q. But what she said about keystrokes couldn't have been
consistent with your understanding about that because
you had no idea?
A. If somebody tells you something that is consistent with
what you think and goes beyond it, then it is
consistent. That's all I'm saying.
Q. So the court should have that in mind for every use of
the word "consistent" in your report?
A. I think I'm using the word "consistent" in the ordinary
sense of the word.
Q. I understand.
A. That things can be consistent if they overlap in
a consistent way.
Q. Let's look at a different example. For all those
reasons you didn't feel there was any need to change any
aspect of your report even on Horizon Issue 8 which
asked about what information --
A. No, there was always this trade-off that one wanted to
write twice as much as one could write, and there was
always this editorial trade-off of, how much to drill
down and what was useful to the court.
Q. But isn't it quite useful just to list to the court what
the information they have is, in answer to the question
"what information do they have", scope: they have this,
they have this, they have this, they did have this, they
didn't have that and they did have the other?
A. I was making decisions about how much detail is needed
here and there, and I made that decision not to do
a detailed list of this and this and this, and perhaps
I was wrong. I'm always having to make these decisions
about how much detail do you go into here in order to
assist the court.
Q. But you would accept, Dr Worden, that it is a binary
issue: do you have access to the keystrokes in the
branch or do you not?
A. That is a binary issue.
Q. So that could have been accommodated in your report by
the addition of a few words in that footnote?
A. It could have been.
Q. Or one line in your report?
A. It could have been. Yes.
Q. Let's move to a different example. I'm not going to go
into the detail of this because we are going to deal
with remote access later. I want to ask you what you
say at {D3/1/244}. Do you see that? If we can go to
{D3/1/244}.
A. Paragraph number?
Q. Let's just get the correct page up. There we are. It
is paragraph 1114. Do you see that's under the heading
"Transaction injection in Old Horizon"?
A. Yes.
Q. You say there:
"Mr Godeseth says that, in the Old Horizon
system ..."
Which is Legacy Horizon?
A. Mm.
Q. "... The SSC could also inject transactions and that
those transactions were clearly distinguished from those
entered at the branch because they would have included a
counter position greater than 32 when no branches would
have had such a high number of counters ..."
Yes?
A. Yes.
Q. And if we go forward, please, to page {D3/1/245}, so
turn over the page, at paragraph 1119 you are
considering what Mr Roll's evidence was.
A. Yes.
Q. You say:
"Mr Roll worked in the SSC, and I established above
that (during his tenure with Fujitsu) certain SSC users
had the ability to transact injections, although these
would have become visible to Subpostmasters. So, in my
opinion, Mr Roll could not have made these changes to
branch accounts 'without the Subpostmaster knowing'."
Yes?
A. Yes.
Q. Now, we know in fact now that Mr Roll was right about
that and we will come back to that when we do the remote
access piece. But at the moment I just want to ask you
about this paragraph and how it is expressed.
A. There is a typo. "Transact injections" is a bit --
I think it is "inject transactions".
Q. Don't worry about that, no one is going to complain
about that. What you say is that Mr Roll worked in the
SSC, yes?
A. Yes.
Q. Service support centre?
A. Yes.
Q. And then these are the words I want to ask you about:
"... I established above that (during his tenure
with Fujitsu) certain SSC users had the ability to,"
inject transactions I think it can be fairly be --
A. Yes.
Q. "... although these would have become visible to
Subpostmasters."
Yes?
A. Yes.
Q. So in your opinion, Mr Roll could not have made these
changes to branch accounts without the subpostmaster
knowing?
A. Yes.
Q. Now, that was a disputed issue of fact between Mr Roll
and the Post Office, wasn't it?
A. Yes.
Q. And you say that you have established it above.
A. Yes. Can we go back to the above?
Q. Let's go back to the above and I will show you. It
seems to the reader possibly that it is paragraph 1114,
which is {D3/1/244} and following?
A. Yes.
Q. Because this is the section where you basically say what
Mr Godeseth says: a counter position greater than 32,
yes?
A. Yes.
Q. And you say it accords with your experience and so
forth. And then you point out at 1116 that it couldn't
be done without the subpostmaster's knowledge, yes?
A. I say "thus". Let me see what leads before "thus".
(Pause)
Q. So what you have actually done, as we can see between
1114 and 1118, and then over the page on 1119, is you
have looked at a disputed issue of fact, you have
accepted Mr Godeseth's factual evidence in the face of
Mr Roll's account, you say you have established over the
page {D3/1/245} that effectively what Mr Godeseth says
is correct because you used the word "established". And
on that footing you have given the court your opinion
that Mr Roll is wrong. That's what you have done.
A. Yes.
Q. And you shouldn't have done that, should you?
A. I think I shouldn't. I think the word "established" was
too strong.
Q. Well, there are two points. (1) It is misleading to say
you have established it because all you have done is you
have said "I accept Mr Godeseth's evidence" and proceed
on that footing as if it is fact. That's the first
problem.
A. As I say, I think "established" was too strong and I was
wrong to say that.
Q. Well, okay. Let's go back. I took you to
paragraph 114. It is on {D3/1/260}. We had a look at
this didn't we, 114?
A. What paragraph are we talking about?
Q. Sorry, I misspoke, paragraph 1194.
Right-hand side, four lines down. That declaration
is not true in respect of what you did where you said
"establish", is it?
A. I think there are occasions in my report I have made
mistakes and this was a mistake, yes.
Q. And the particular type of mistake that we are talking
about is one that you have expressly said at the end you
have not done?
A. That's what this says, I think. I think on that
occasion, on this counter 32 issue, I did, although
I said at the time Mr Godeseth's evidence is consistent
or accords with my experience in that paragraph, I did
later at 119, or whatever it was, go on to draw
a stronger inference which I should not have done.
Q. You have based your opinion accepting Mr Godeseth's
evidence and on that basis you have rejected Mr Roll's
evidence of fact on a disputed issue, hotly disputed and
central to these proceedings. The answer to that is
yes?
A. Well, this has developed over time and what I have said
is that my statement in the first report was too strong,
and --
Q. But pausing there, your approach, it is not just the use
of the word "established", I'm putting to you fairly and
squarely that your approach was wrong. There were two
witnesses saying opposite things. As an expert you
should have said: if Mr Godeseth is correct, this; if
Mr Roll is correct, it undermines the proper access
controls that I mention in my witness statement as my
final countermeasure and undermines the reliability or
robustness, or whatever it is, of the Horizon system, or
whatever your view was if Mr Roll was correct. That's
what you should have done, isn't it?
A. I have tried in my reports to --
Q. Can you just answer the question: do you accept that's
what you should have done?
A. I'm just prefacing an answer. I have tried in my
reports generally to acknowledge the limitations of
factual witness evidence. On this occasion I was wrong
and I should have done so.
Q. Can you think of any prominent example in your report
where you proceed on the footing that the claimants are
right about anything?
A. Well, this depends on the claimants' evidence being
relevant to my opinions, and I can't think of any
occasions where individual subpostmasters' evidence, for
instance, is relevant to my opinions.
Q. But we know from this example that Mr Roll's evidence
was the evidence that was put forward by the claimants.
And you rejected it.
A. Yes.
Q. And we know that that's an important example because we
now know the answer that you were wrong to reject it,
don't we?
A. We do.
Q. So it was both wrong in its approach and wrong in its
outcome, wasn't it?
A. The outcome was wrong, no doubt. As I say, my approach
has been to try wherever possible to qualify factual
witness evidence with the statement: this is my
assumption and this is why I'm using an assumption to
proceed.
I have tried wherever I can to make that
qualification and we agreed on this occasion that
I should have made it more careful.
Q. I can only put it one more time: on this your approach
was wrong and your conclusion was wrong?
A. Yes.
MR GREEN: My Lord, would that be a convenient moment?
MR JUSTICE FRASER: Yes, we will have 10 minutes.
Dr Worden, you know this, I know, because you are
experienced and you also heard me reminding Mr Coyne
many times, but now you are in the middle of your
cross-examination you are not allowed to talk to anyone
about the case. We are going to have a 10-minute break
for the shorthand writers. Please don't feel you have
to stay in the witness box. I always encourage
witnesses to move around and stretch their legs. But we
will come back in at 11.55 am.
(11.46 am)
(A short break)
(11.55 am)
MR GREEN: Now, Dr Worden, just picking up from where we
left off just before the break, we had been looking at
{D3/1/245}, which is your first report, at
paragraph 1119. And I'm going to ask you about that.
Yes? Do you remember?
A. Yes, I do indeed.
Q. I think you may have mentioned that you address that in
your second report?
A. Yes.
Q. Between the first and second reports there had been
a further statement from Mr Roll and a further statement
from Mr Parker, hadn't there, Mr Parker's second witness
statement?
A. That is right.
Q. And you would have read those?
A. Yes.
Q. To inform yourself about whether the approach you had
taken in your first report was right or not?
A. Yes.
Q. It is the only fair way to do it, isn't it?
A. Well, obviously I did --
Q. Yes, you must have done. And paragraph 27 of
Mr Parker's second witness statement explained that what
Mr Roll had described was possible. Do you remember
that?
A. If we go to that --
Q. Let's have a look at it. It is on {E2/12/9}.
There's Mr Parker's second witness statement. Go to
page {E2/12/9} of that. If we look at paragraph 27, and
this is Mr Parker referring to paragraph 20 of Mr Roll's
second witness statement?
A. Yes.
Q. "... Mr Roll describes a process by which transactions
could be inserted via individual branch counters by
using the correspondence server to piggy back through
the gateway."
Yes?
"He has not previously made this point clear. Now
that he has, following a discussion with colleagues who
performed such actions I can confirm that this was
possible."
Yes? So we are now in a position where a point
favourable to the claimants has not only been set out in
the claimants' evidence, but agreed by the Post Office.
That is correct, isn't it?
A. Well --
Q. On the evidence.
A. The difficulty I have with this, or one difficulty
I have, is that I don't understand the phrase "piggy
back".
Q. Well, let's pause there. Just look at what he actually
says:
"He has not previously made this point clear."
Line 3:
"Now that he has, following a discussion with
colleagues who performed such actions ..."
So it is not theoretically possible. I have spoken
to people who were actually doing this. Having done
that, I can confirm that this was possible.
A. Yes.
Q. Right. Now, pause there. The premise that you
proceeded upon, which we have looked at in your first
report, was that Mr Roll was wrong, that it was
possible. This is a yes or no answer.
A. No, it is not that simple. In other words, that
I believe that what I did not accept in my first report
was that it could be done without the subpostmaster
knowing --
Q. Precisely.
A. And I was never very clear on -- I never got to the
bottom of the issue of what the subpostmaster might
know. Now, piggy backing says this was possible, but
I was not clear whether that would make it evident to
the subpostmaster.
Q. No, Dr Worden, the way it went was that Mr Rolls said it
was possible without the knowledge of the subpostmaster?
A. Yes.
Q. And that's what you were specifically responding to in
your first report?
A. Yes.
Q. And that's what you rejected which we dealt with before
the break, yes?
A. Yes.
Q. And you rejected it on the basis of Mr Godeseth's
evidence that it would be done through a counter number
higher than 32?
A. Right.
Q. So that it would be visible to the postmaster if they
noticed the counter number?
A. Counter number higher than 32 was one way in which the
postmaster could find out.
Q. Well, that was the way --
A. No, not the way. That's not what I believe.
Q. Well, that's what Mr Godeseth was saying, wasn't it?
A. That's what Mr Godeseth said at the time, yes. But --
Q. Let's go back to --
A. -- fundamentally, I believe that whether the postmaster
knew or not is not a simple matter of counter 32 or not.
Q. Dr Worden, you know where this cross-examination is
going, don't you?
A. No, I don't.
Q. Let's have a look and then I will bring you back to what
you have just said.
Mr Parker in paragraph 27 specifically says that
what Mr Roll has been describing --
A. Can I read paragraph 27 carefully, then. {E2/12/9}
This was possible --
Q. Without using counter 32, yes?
A. But this paragraph, as far as I can read it, does not
refer to the subpostmaster's knowledge.
Q. Okay. Well, I will have to come back and trace through
everything in a minute. Let me take you forward to what
you say in your second report and we will do it all over
again carefully. But {D3/6/20} is your second expert's
report. Okay?
A. Yes.
Q. So you were commenting, yes, on paragraph 82 at the
bottom of the page?
A. Yes.
Q. You are specifically commenting on paragraph 20 of
Mr Roll's witness statement.
A. Mm.
Q. Yes?
A. Mm.
Q. And you say:
"[He] addresses a factual point about injection of
transactions."
So when you say that, you know that that is
a factual matter in dispute, don't you?
A. Yes.
Q. You do. Second line:
"He says: 'Sometimes we had to ask for a specific
person to log in to the counter before injecting
transactions so that the software would not detect any
discrepancies. A transaction inserted in this way would
appear to the subpostmaster as though it had been
carried out through the counter in branch."
A. Yes.
Q. Then you add the words {D3/6/21}:
"He ... goes on to disagree with my paragraph 1119."
Which is the one --
A. Can we just go back to that paragraph again so we can
compare them side by side?
Q. Yes, we can.
MR JUSTICE FRASER: It might be sensible to have that in the
hard copy.
A. Yes, that's what I'm trying to do.
MR JUSTICE FRASER: Is that what you are looking for? It
is --
MR GREEN: 245.
MR JUSTICE FRASER: I don't know if Dr Worden has
a paginated one.
MR GREEN: I understand.
A. Sorry about this.
MR JUSTICE FRASER: Don't worry.
(Pause)
A. Now, this is chapter 11, is it?
MR JUSTICE FRASER: Does that document have a bold number in
the bottom right-hand corner, a bundle page number?
A. Yes, it does.
MR JUSTICE FRASER: If you turn to {D3/1/245}, is that the
page you want?
A. Right, okay, these would have become visible to
subpostmasters without the subpostmaster knowing were
there. We are there.
Q. So the genesis of this is that at 1119 you explain that
you have established above that these would become
visible to subpostmasters and rejected Mr Roll's
account. That's what you did at 1119. You can see
that?
A. That is right, that is correct.
Q. You have accepted that the approach and the conclusion
were both wrong?
A. I was overstrong.
Q. That the approach was wrong?
A. Yes, I have accepted that.
Q. And that the conclusion was wrong?
A. I have accepted that.
Q. And you also accepted that it is clear from the evidence
that the conclusion was wrong --
A. I accepted that I was wrong in accepting Mr Godeseth's
evidence over Mr Roll's evidence in my first report.
Q. But we now know, I put to you, that those changes can be
made without the knowledge of the subpostmasters?
A. Well, it is without the knowledge of the subpostmaster
that's the nub of the issue.
Q. So you write paragraph 1119, then we get Mr Roll's
second witness statement, and Mr Roll -- we won't go
there now but it is {E1/10/6}. Then we get Mr Parker,
his second witness statement?
A. Which doesn't refer to knowledge of the subpostmaster in
the paragraph you took me to.
Q. Okay. Let's go back to Mr Roll at {E1/10/6}.
Now, the passage that you have now accepted between
114 and 119 was perhaps not as it should have been?
A. Yes.
Q. Is the very passage to which Mr Roll is referring in
paragraph 20 on this page, isn't it?
A. Can I read this long paragraph, 20?
Q. Please do. (Pause)
A. Okay, I have absorbed that paragraph now.
Q. And he specifically explains two-thirds of the way down
in that paragraph "therefore". Do you see "and
therefore"?
A. "... and therefore not a number greater than 32,"
absolutely.
Q. "... and not in a way which would distinguish it in any
logs as having been inserted by Fujitsu rather than by
the subpostmaster or an assistant."
A. Yes.
Q. Then at the bottom of the page, five lines up on the
right-hand side, "A transaction inserted". Do you see
that?
A. Yes.
Q. "A transaction inserted in this way would appear to the
subpostmaster as though it had been carried out through
the counter in branch."
A. Yes.
Q. Couldn't be clearer, could it, what he is saying?
A. It is pretty clear but --
Q. "I therefore disagree with Dr. Worden's conclusions that
these transactions would always have been visible to
subpostmasters ..."
And he references the specific paragraph he is
looking at:
" ... if he means to say that they would be shown to
subpostmasters as transactions inserted by Fujitsu,
rather than as transactions which appeared to have been
created in branch."
A. Yes.
Q. Right. And that is what, when we go to Mr Parker's
second witness statement -- let's look at that again.
It is {E2/12/9}. Mr Parker, paragraph 27, we have
already looked at it.
A. Yes.
Q. Mr Parker is referring to paragraph 20. That's the
paragraph we have just looked at, isn't it, Dr Worden?
A. That is correct.
Q. "... Mr Roll describes a process by which transactions
could be inserted via individual branch counters by
using the correspondence server to piggy back through
the gateway. He has not previously made this point
clear. Now that he has, following a discussion with
colleagues who performed such actions I can confirm that
this was possible."
A. Yes.
Q. So the conclusion that it is not possible to do so, we
can see at this point has been reached by Mr Parker as
well as --
A. The question is what do we mean by "it"?
Q. Dr Worden, you knew that a central issue, not only
a central issue legally but a very high-profile issue in
the case, was the extent to which Post Office had remote
access to the counters, didn't you? You knew that?
A. Yes, and what I'm talking about, what I was talking
about was the extent to which this could happen without
the knowledge of the subpostmaster.
Q. And that's the --
A. And we agreed in the joint statement that more or less
Fujitsu or Post Office could do anything.
Q. Let's just take it in stages. That turns out to be the
case, but we will come back to that separately.
At the moment what I'm asking about is the approach
that you have taken to disputed issues of fact in your
reports. And this is one of them. We have identified
what went wrong with paragraphs 1114 to 1119. I'm going
to ask you now to look again; I have shown you Mr Roll's
second statement, I have shown you Mr Parker's second
statement. Look now, please, again. Go back to where
we were before, {D3/6/20}.
This is your second expert report?
A. Sorry, I'm not there yet.
Q. Don't worry. Do you see at the top --
MR JUSTICE FRASER: We are not there yet.
MR GREEN: Sorry. {D3/6/20}, "ER2 Worden". Do you see
that?
A. Yes.
Q. "Charteris" at the top right-hand corner. This is your
second report. We are going to look, please, at
paragraph 82 again.
A. Right.
Q. "In his paragraph 20, Mr Roll addresses a factual point
..."
And you accepted earlier that that was a factual
point that you knew to be in dispute?
A. There are two factual points in dispute. There is
whether SSC could inject transactions and whether they
could do it unknown to postmasters.
Q. Quite. And the witness statements we have been looking
at specifically deal with whether they could do it
unknown to subpostmasters.
A. Well, what they specifically deal with is a particular
mechanism for counter 32 by which subpostmasters might
find out. And that, in my opinion, is probably not the
only mechanism.
Q. Let's take it in stages. The answer you just gave was
very interesting. It was in two parts. You said
that -- give me one second -- they were dealing with
counter 32, yes? Those witness statements, the
counter 32?
A. Dealing with counter 32. What do you mean by that?
Q. I will give you your exact words in a minute when my
realtime starts working again, but the point you were
making is that the witness statements were referring to
this counter 32 point that had been raised by
Mr Godeseth.
A. Yes.
Q. And the way in which they were dealing with them,
Dr Worden, is that they were saying: actually, as well
as counter 32, which would be capable of showing
a subpostmaster that it had been injected by Fujitsu,
there was also a separate way, which would not. That's
what those witness statements were addressing, isn't it?
A. There was a separate way which would not do the
counter 32 thing, but then the question is whether that
separate way is unbeknownst to the postmaster.
Q. Well, Mr Roll specifically says it is unbeknownst to the
postmaster --
A. Where does he say that? Can we go back to that
statement?
Q. We can do it one last time. I think I am going to have
to move on. Mr Roll, paragraph 20, {E1/6/1}.
A. Before we move on, I would like a short opportunity to
try and explain what I was doing here.
Q. Let's go back to your second witness statement. If you
are going to explain something, can you explain why it
is in your second witness statement that you did what
you do.
Let's look at D3 --
A. I would really like to explain that.
Q. We will do it in stages. I will ask the questions and
you give the answers and explain what you need to
explain to his Lordship about how you have adopted this
approach as an expert.
MR JUSTICE FRASER: Can you each try not to talk over one
another. It is not just you; it is Mr Green as well.
MR GREEN: I apologise.
MR JUSTICE FRASER: The transcribers need to get everything
down, so one at a time.
So where are we going?
MR GREEN: {D3/6/20}.
A. This is a long paragraph, or is it?
Q. No, this is in your second witness statement --
A. Report.
Q. Sorry, your second expert report. I do apologise. And
I am trying to take you to paragraphs 82 to 85. We
never get past 82. So let's just see what paragraph 82
says.
A. Yes. It summarises apart from the end of Mr Roll's
paragraph 20.
Q. If we go over --
A. Could I get to my supplemental report so I can get over
the page?
Q. Of course. Please get any documents you wish,
Dr Worden. I don't want to stop you.
A. Where are we? Can somebody tell me the chapter?
Q. It is your second report. Your second report is -- you
should be able to find it -- {D3/6/20}.
A. 20?
Q. Page 20.
A. Right, okay.
Q. Do you have that?
A. Got it nearly. Sorry, I'm not there yet. Not there
yet.
Right, okay. Got it.
Q. You got that page?
A. Yes.
Q. Okay. And I was asking you about paragraph 82 at the
foot of that page.
A. Yes.
Q. I will just wait until you finish pouring the water?
A. Sorry.
Q. Don't apologise. I just don't want to -- if you want to
try and turn the pages. Now, we were looking at the
foot of page 20, weren't we?
A. Yes.
Q. "Mr Roll addresses a factual point about injection of
transactions. He says: 'sometimes we had to ask for
a specific person to log in to the counter before
injecting transactions so that the software would not
detect any discrepancies. A transaction inserted in
this way would appear to the subpostmaster as though it
had been carried out through the counter in branch."
{D3/6/21}
A. I think his actual 20 has a kind of "if" after it.
Q. Let's look at what you are citing.
A. Let's look at what I say, let's do that.
Q. You say:
"He then goes on to disagree with my
paragraph 1119."
Which is the one we have already explored in --
A. And that's where I said I didn't think it could be done
without the knowledge of the --
Q. That's where you say you disagree it could be done
without the knowledge of the subpostmaster.
A. Yes.
Q. You knew that a factual witness who had actually worked
at Fujitsu, he is not a claimant, he has volunteered to
come and give independent evidence, was saying that this
was possible?
A. He was saying it was possible without the knowledge of
the subpostmaster. My opinion is --
Q. Can we take it in stages, please, because you knew that
he was saying that and you also knew by the date of this
report that Mr Parker had agreed, didn't you?
A. No. Mr Parker had agreed it could be done. He did not
agree that it could be done without the knowledge of the
subpostmaster.
Q. Okay.
MR JUSTICE FRASER: Can I just ask for a point of
clarification because it has not come up on the
LiveNote.
I think in your answer you just said to Mr Green "He
was saying it was possible without the knowledge of the
subpostmaster". By "he", you mean Mr Roll?
A. Yes.
MR JUSTICE FRASER: Then you said "my opinion is" and
I thought you said different to that, but it didn't go
on the -- is that right or did you --
A. Yes, my opinion is different to that.
MR JUSTICE FRASER: Your opinion is different to what
Mr Roll said?
A. I'm not sure it contradicts what Mr Roll actually said
because Mr Roll said there was this counter 32 mechanism
which made it clear. My opinion is yes, that would have
made it clear, but there may be other mechanisms like
actually seeing the guy doing it.
MR JUSTICE FRASER: Okay.
MR GREEN: If we read down through this section of your
witness statement, you say:
"It seems to me that I require further factual
information before I can comment on this evidence.
Which 'specific person'? Under what circumstances? How
frequently? Until I have that information, it remains
possible in my view that any transaction which 'would
appear to the subpostmaster as though it had been
carried out through the counter in branch' might only be
a transaction that he had given his consent for, as the
'specific person' - and which had in effect been made on
his behalf."
So what you are saying is you are reading Mr Roll as
saying that sometimes they asked for somebody in the
branch to be logged on, yes?
A. Well, I'm not reading Mr Roll, I'm saying I want to know
more about what Mr Roll says.
Q. And if we go to paragraph 85, you say that:
"In his paragraphs 27 - 34, Mr Parker provides
detailed and specific commentary on Mr Roll's
paragraph 20, using his knowledge and the appropriate
contemporary documents, where they have been found.
Here he acknowledges that Fujitsu could insert
transactions into branches by a piggy back process. I
am not yet able to comment on Mr Parker's evidence or
the documents he cites."
Why were you able to comment on Mr Godeseth's
evidence and apparently establish a disputed fact, but,
when Mr Parker gave evidence favourable to the
claimants, unable to comment on his evidence?
A. It was a time issue. You know, did I have other
evidence to drill down, did I have time to drill down?
But Mr Parker's evidence went to the point of whether it
was possible and subsequently the experts have agreed
that it was very difficult to say anything was
impossible. Mr Parker's evidence goes to that. The
issue that we are discussing is whether it is possible
without the knowledge of the subpostmaster.
Q. What evidence would you have needed in front of your
eyes, beyond Mr Parker agreeing with Mr Roll, to be able
to say: on the basis of that, I accept that it was
possible to inject a transaction without the
subpostmaster knowing?
A. Well, that's a bit hypothetical. I mean, I can
conjecture that if Mr Parker had said "I agree piggy
backing was possible" and "I agree that the
subpostmaster would have known nothing about it", then
that would have done it.
Q. Pausing there, Dr Worden. Do you accept that by
refusing to comment on Mr Parker's evidence when he
agrees with Mr Roll, that contrasts very strikingly with
accepting Mr Godeseth's evidence as establishing a fact?
A. I was not refusing to comment on Mr Parker's evidence.
What I was saying was that Mr Parker and Mr Roll seemed
to agree that something was possible, so I'm not going
further on that. What I don't know about is whether
something was possible without the subpostmaster
knowing.
Q. Well, I suggest to you, Dr Worden, that it is bizarre to
deal with that evidence in that way.
A. I don't agree.
Q. Let's look at the joint statement at {D1/5/6}. Now,
that, 10.6, is where the subtopic of remote access is
being dealt with. Yes?
A. Yes.
Q. And Mr Coyne is dealing with it. Where do you say to
the court that you have reflected your view in the light
of Mr Roll's further evidence and Mr Parker's further
evidence in this agreed statement? Or didn't you?
A. I would need to -- let me pore through that joint
statement, but basically my recollection of the joint
statement is -- which one is it?
Q. Dr Worden, how about this: is that something you would
like to have a glance at over the luncheon adjournment?
Just in case there's any particular parts that come to
mind?
A. It might be useful, yes. But as I say, my opinion in my
second report was that I wasn't clear from the evidence
I had seen, so whether this could be done without the
subpostmaster's knowledge, and I would have required
more evidence from either Mr Roll or Mr Parker to make
myself clear on that. In the joint statement I do
recall that the experts agreed that you cannot say
anything was impossible, that, you know, it is difficult
for the experts to say at this distance that certain
things were not possible. However, that was a separate
issue from whether a thing could be done without the
subpostmaster's knowledge.
Q. Okay, let's just go back one last time to {D3/1/245}.
So even though you say that the experts were agreed that
they couldn't say things were impossible, it is the
precise effect of what you conclude at 1119, isn't it?
A. 1119 --
Q. You say:
"... Mr Roll could not have made these changes to
the branch accounts ..."?
A. Can we look at the agreement on the joint statement to
see how that meshes with 1119? I can't remember where
it is in the joint statement.
Q. Go back to {D1/5/6}. You have referred to
paragraph 1114 --
A. Sorry, I'm not there yet.
Q. If you look at the screen. Do you see 10.6?
A. That's Mr Coyne's statement.
Q. Yes, Mr Coyne said that. And there's a reference there
in the right-hand corner of that box to
paragraph 1114 --
A. Sorry, it is not clear to me who put in the reference.
If Mr Coyne made a statement I think probably what
happened is he put in the references.
Q. Okay. What I was keener on was perhaps you can have
a glance at the joint statement over lunch?
A. I will do, yes.
Q. And if there was somewhere where you thought the court
should fairly look to see any revision of the way your
views were expressed as at 4th March, then perhaps you
can direct the court to that.
A. Yes, sure.
Q. Now, when you are doing that, could we just go back,
please, to your second witness statement --
MR JUSTICE FRASER: Report.
MR GREEN: I'm so sorry. I apologise.
Dr Worden, your second report, second expert's
report, at {D3/6/21}.
A. Yes.
Q. Having said that the experts couldn't say anything was
impossible --
A. Well, I really would like to see the wording of that
joint statement.
Q. No, no, but I'm asking you about what you put in your
second report before the joint statement.
A. Yes.
Q. Okay. And in the your second report in paragraph 84 you
have actually -- this appears to be the thing -- one
thing that you are able to say is impossible. So you
say:
"Therefore --
A. I say:
"Mr Roll's ... evidence does not cause me to alter
the opinion ..."
Q. Yes, the one that we agreed was incorrect in approach
and conclusion.
A. We agree now, but at the time I wrote the second report,
on the basis of Mr Roll's evidence and Mr Parker's
evidence, I didn't see reason to change that opinion.
Q. I would suggest to you that you were inexplicably
reticent to accept something that was contrary to
Post Office's interests.
A. No, I was reticent. Not inexplicable, I was reticent
because I had not seen sufficient evidence to convince
me that these things could be done without the knowledge
of the subpostmaster.
Q. But Dr Worden, would you accept that the approach you
have taken there contrasts very, very strikingly in how
you approach Mr Roll's evidence with the approach you
took at 1119 in your first report when you accepted,
effectively, Mr Godeseth's evidence?
A. I have accepted that my approach at 1119, that use of
the word "established" was wrong and my approach was
wrong, and we have established that -- you know we have
done that before the interval --
Q. My question is: do you accept the contrast is very
striking?
A. I think the court will have to -- I accept that my
attempt to make my position clear that I'm not trying to
find findings of fact, I'm not trying to find one
witness or the other, I accept that on this occasion
I fell short of that.
Q. Do you mean only in your first report or also in your
second report?
A. Not in the second report. I believe the second report
was -- you know, I believe that the evidence from
Mr Parker and Mr Roll, and if we look again at Mr Roll's
paragraph 20, I believe that the question that I asked
in paragraph 83 that I want further evidence, I believe
that was a valid approach.
Q. When you say in paragraph 84 -- because we have dealt
with what you accept about your first report at
paragraph 1119 several times, haven't we?
A. Yes.
Q. So in your second report, you say, notwithstanding what
you have now accepted about 1119, in your second report
that you were right not to change that view.
A. I hadn't seen sufficient evidence to change that view.
Q. But do you say you were right not to change that view
when you revisited the entire piece in your second --
A. At the time I believe it was right not to change that
view because I hadn't seen evidence that convinced me
that this change could be made without the knowledge of
the subpostmaster. That's where I am.
Q. Dr Worden, do you accept that that betrays a complete
failure to appreciate the need to consider the situation
both on the basis of whether the claimants' evidence is
right as well as on the basis that the defendant's
evidence is right?
A. No, I don't accept that.
Q. Do you accept that you have failed to make obvious
observations on the basis of evidence of the claimants
with which one of the defendant's evidence has in fact
agreed?
A. I accepted what I accepted about the first report,
1119 --
Q. No, on the second report.
A. On the second report, no, I don't accept that.
I believe that Mr Roll's evidence about counter 32 and
sometimes we had to get somebody's permission and so on,
that left me in a position of doubt where I was not able
to change my opinion.
Q. How did you pursue any genuine doubts that prevented you
from saying anything helpful to the claimants about
that?
A. What do you mean pursue?
Q. Did you ask anyone at Fujitsu?
A. I mean, I had to write this report and this was the
position at the time of this report. I was not
expecting a third witness statement from Mr Roll
addressing that there might have been that, for
instance. I just wrote down what I felt was appropriate
to write down in the second report.
Q. Let's pause there. You identify your sort of, as it
were, standing on the island of not having enough
information to say anything, and the question is, what
did you do about it because it is obviously an important
issue? Did you ask for any further information
yourself?
A. I do not think I did, no.
Q. Because not only are you entitled to ask for further
information from those instructing you, but the courts
made a specific direction that you can apply for
directions, and you also have the ability to request
information formally, don't you?
A. Absolutely. But let me try and explain the sources of
my doubt. Counter 32 is clear cut. If counter 32 is
there, the subpostmaster sees it. But in my opinion,
there may be many other mechanisms whereby the
subpostmaster may observe that something has been done,
and that includes simply observing that somebody had
been in his branch and he had given his permission to do
something, and that was the area of doubt which was
raised by Mr Roll's statement.
Q. So pausing there, Dr Worden. Did you understand that
what was being agreed was that a Fujitsu person would
actually go into the branch? Is that what you think
this is all about? It is, isn't it?
A. This was my doubt, right, which person, what
circumstances. And if we look at Mr Roll's
paragraph 22, you know, we had to ask for a specific
person to log in and that means somebody at the counter
in the branch doing something specific.
Q. Because the position is that they can't use that unless
someone is logged in at the branch. And my question to
you is slightly different. Did you understand this
entire exchange as being about someone from Fujitsu
turning up at the branch and saying "Jeff, can you log
in, please"?
A. No.
Q. In the branch?
A. I didn't understand the entire exchange in full. It
left doubt in my mind that something had to be done at
the branch, somebody had to log in at the branch to make
this thing happen. I didn't understand the full detail
of that so I was left in doubt.
Q. Because you made references in your answers to seeing
someone actually make a change at the branch?
A. Well, the whole question I was asking consequent to
Mr Roll's witness statement is: what does he mean?
Q. But the entire issue is about remote access, not at the
counter, isn't it?
A. Remote access means not at the counter, but this seems
to have been facilitated by something at the counter.
Q. I understand, but it sounded as if you were saying you
understood a Fujitsu person would go into the branch and
be seen to make an alteration?
A. No, I didn't mean that.
Q. So if you have referred to that in the transcript we
should disregard that; that's not what you had in mind?
A. What I have referred to is that Mr Roll's paragraph 10
to 20 leaves me in doubt as to what actually happened.
Q. Let me go back to the original question from which we
have gone round a little bit. What steps did you take
to clarify the ambiguity which bore directly on one of
the identified Horizon Issues? Did you take any steps
at all?
A. I didn't take any further steps beyond this report.
Q. Could you tell his Lordship why not?
A. Priorities and, you know, whether I felt that further
investigation would get me further. It was kind of my
feeling that it was a difficult area and that -- put it
this way, sorry, it is very hard to explain this, but
there are levels of depth and complexity in the way
Horizon actually works which the experts have not been
able to plumb, if you like, and there is a whole lot of
detail about how a transaction might have been
identified. For instance, there is a PEAK that talks
about counter 11 and 12. That's not counter 32 but that
will maybe give an indication.
To my mind there was a kind of swamp of difficult
questions there and I was not going to -- I felt,
rightly or wrongly, going to make progress in that area.
Q. Even in circumstances where in your first report you had
concluded that Mr Roll was factually wrong?
A. I was taking the situation as at my second report, and
as I say, I wrote down what I believed I could conclude
at the time and I did not have much expectation that
I would be able to conclude more.
Q. Dr Worden, what is weird is you conclude very forcefully
one way in your first report and then you are not
prepared to reach a conclusion which would have been
correct the other way in your second?
A. I do not think that is weird. I'm trying to go with the
evidence and provide what I know to the court.
Q. But you are going only with the defendant's evidence in
report 1 and then when the defendant's evidence agrees
with the claimants', you stop going with anyone's
evidence?
A. No, I have to go where I think the evidence takes me and
I have to try and be balanced and neutral about it. And
sometimes I may fail to do that, I'm sorry, but I'm
trying to actually assess the evidence. And to my mind,
in my second report, what Mr Parker said and what
Mr Roll said did not sufficiently convince me that
I could write down, yes, it could be done without the
knowledge of the subpostmaster.
Q. What threshold of proof were you applying when you are
faced with the defendant's evidence, agreeing with the
claimants' and you are still unconvinced? Beyond
reasonable doubt? 99%?
A. It is not a question of threshold of proof, it is
a question of what opinions I can usefully offer the
court.
Q. But you had offered one and you are refusing to change
it in your second expert report. It is not that you
haven't offered one at all. Dr Worden, if we were there
we would all understand. You say too complicated, it is
absolutely impossible to plumb the depths of this
system, as you mentioned, I can't give an opinion one
way or the other. That is not what happened. You give
a very clear opinion which we've analysed with some
care --
A. And I've agreed it was overkill.
Q. Right. And then you refuse to change the overkill
opinion in any way because you say there's some doubt.
A. Well, we established today that my approach in my first
report was wrong.
Q. Yes. And you stick to it in your second report, that
the point --
A. No, I wrote in my second report -- in my second report
I wrote what the evidence persuaded me and convinced me
about, and that was I hadn't seen sufficient evidence
that this thing could be done without the knowledge of
the subpostmaster. I was trying to be balanced and fair
and that was what I was doing in my second report.
Q. So this exchange that has gone on a bit longer than
I expected, but this is helpful to the court in
understanding your approach to being balanced and fair,
what you have done here is a fair example, is it, of you
being balanced and fair throughout your reports? That
is right?
A. I don't see that I can comment to how this example
relates to the whole of my reports.
Q. Do you think you have been especially fairer elsewhere
in your reports? You have given this, you have
described this approach as balanced and fair --
A. I'm trying not to go beyond the evidence.
Q. Just focusing on the balance and fair point, let's not
go back to the evidence. You say that this is
an example of you being balanced and fair?
A. Yes.
Q. You are a statistician, yes? You understand what
a representative example means, don't you?
A. Yes.
Q. Is this a fairly representative example of you being
balanced and fair in your approach in your reports? It
is, isn't it?
A. I think this is a fair example of me trying to assess
the evidence in front of me and trying to draw
conclusions as far as I can and not trying to go beyond
the evidence, and that is fair, I believe.
Furthermore, if you look at the balance of my
reports, there are many, many occasions where I have
tried not only to be fair on the claimants but to be
biased in my numerical estimates in favour of the
claimants.
Q. We will come to those.
A. Absolutely.
Q. So, one final question and we will move on. You would
have no problem with the court using this as an example
of your approach overall?
A. What's written in the supplemental report, I have no
problem with the court using it as an example of my
approach.
Q. Even if it were a totemic example of your approach --
A. Totemic is not a word that occurs --
Q. A very good example of how you approached the evidence?
A. I would hope the court will take into account all sorts
of examples of my approach in assessing whether I've
been balanced or not.
Q. I understand. Let's move on.
You have given your evidence that you didn't follow
up your doubts on this point. More widely, have you
spoken to any of the Fujitsu witnesses in the course of
preparing your reports?
A. There was one early conversation.
Q. Who was that with?
A. With Gareth Jenkins, and it was a phone call with
lawyers present and I was trying to clarify --
Q. Well, if lawyers were present I'm not going to ask you
about the content, if it was in any way privileged. So
my learned friend will indicate if he has got any
problem with that?
MR JUSTICE FRASER: Was the conversation with Gareth Jenkins
on the phone as well?
A. It was on the phone.
MR GREEN: Okay. What about anyone else from Fujitsu?
Because there are lots of witnesses have been here, you
have all got a room, everyone has been going in and out,
you have had the opportunity to speak to people from
Fujitsu, haven't you?
A. Well, the defendant's lawyers set down a set of ground
rules for what --
Q. I don't want to ask you about what they told you, I just
want to ask you have you spoken to anyone from Fujitsu?
Mr Godeseth? Mr Parker?
A. We talk about the weather, yes. I mean --
Q. No, about the case?
A. About the case, no.
MR DE GARR ROBINSON: My Lord, I'm perfectly happy for my
learned friend to ask about the ground rules that the
defendant's lawyers set down if my learned friend
wishes to.
MR GREEN: And has anyone from Fujitsu provided any written
comments or observations on your report, because we
heard Mr Godeseth had had some comments which originated
from Mr Jenkins on one of his statements?
A. It is obvious that when I did analysis of various KELs,
Fujitsu did the same analysis and that came out in
a witness statement which I saw at the same time as
everybody else. That was the only instance.
Q. But they haven't actually provided any comments for you?
A. Not for me, that went into my reports.
Q. But you did know you were entitled to ask questions of
Fujitsu, didn't you?
A. Well, that always happened through the Post Office
lawyers.
Q. No, of course. And you have in fact asked questions
through Post Office lawyers of Fujitsu, haven't you?
A. I have done, yes.
Q. If we look, please, at {H/302/1}. It is a letter of
29th May, and if we go, please, to the second page
{H/302/2}, we can see halfway down:
"Question from Dr Worden."
A. Yes.
Q. And there are questions and Fujitsu responses?
A. Yes.
Q. You hadn't actually agreed those questions with
Mr Coyne, had you?
A. No.
Q. Was there a reason for that?
A. It was just my -- not curiosity, but there were various
questions in my mind that were unanswered to me from my
examination of the documents. I was not aware that
Mr Coyne was interested in the same questions.
Q. Okay. So you followed up things that you had some
uncertainties --
A. I found those documents puzzling in various places.
Q. So on this puzzling issue you did follow up by asking
questions?
A. Yes, because I felt answering the questions could be
productive, I felt there was a big chance that, you
know, I would get an answer which would actually resolve
that uncertainty.
Q. Okay. If we go back a page {H/302/1} to the first page
of this document, we can see that the date on the letter
was 29th May.
A. Yes.
Q. And can you remember roughly when you asked the
questions?
A. I can't, I'm afraid. I mean --
Q. Was it a few days -- because 29th May is the Wednesday
before we restarted the trial.
A. I think it was rather earlier than that. I mean, I have
read the low level design of the transaction correction
tool a considerable time earlier and I think my
questions about it had nothing to do with the trial
starting.
Q. Okay. On the face of it, let's look at {H/324/1}, on
3rd June in a reply to a letter from Freeths, it says:
"Dr Worden asked the questions which are set out in
our letter on 19 March 2019."
Does that sound about right?
A. Yes, that probably is ...
Q. If we just go back to {H/302/2}. Just under the heading
it says:
"Dr Worden has asked a number of questions ..."
A. Yes.
Q. It seems these are the ones you asked back in March,
yes? And it then goes on to say:
"We have taken instructions from Fujitsu ..."
Yes?
A. Yes.
Q. "... in respect of these questions and set out below
both the questions asked and responses given. These
responses are being disclosed to Dr Worden at the same
time as this letter is being sent to you."
So did you get the answers for the first time last
Wednesday?
A. I believe so.
Q. Okay.
A. Again, I'm not brilliant on these sort of dates and
chains of events, but I believe so. I mean --
Q. Dr Worden, can I ask you this. Mr Coyne made lots of
requests, didn't he?
A. Yes.
Q. Lots?
A. Quite a lot, yes.
Q. And although they were put in a form of a joint request
document, in fact he was making lots and lots of
requests for information which only he was making?
A. Yes.
Q. Was there a reason why you didn't support any of them?
A. Well, basically I think it became evident, and it is
evident now, that two experts took very different
approaches, and my approach was top down, understand the
architecture and work down through things like KEL. So
as far as I was concerned, I had plenty of information
to go on. And Mr Coyne's information requests didn't
strike me as things that, yes, I have got to really see
that because I was -- had a different priority. I was
trying to do a top down understanding of the
architecture and top down look at robustness, and so on
and so forth.
And so I had plenty of documents to look at,
basically. So I think it is the different approaches
taken by the two experts that led to lack of overlap.
Q. That's your explanation for not having supported any
requests?
A. I do not think I supported any requests --
Q. Sorry, not having supported the very many requests he
had to make?
A. Yes, I felt his interests were different from mine and
I really had plenty to do.
Q. And you initially only looked at KELs, didn't you, you
didn't actually look at PEAKs?
A. No, I looked at KELs and I looked at PEAKs where they
were relevant. I felt that KELs were a more distilled
form of information. I felt they were sufficient in
many ways, especially when you go and look at the PEAKs.
Q. You looked at one or two PEAKs referred to in the KELs
you looked at?
A. Yes. Look at some of them where you feel the KEL
doesn't tell you enough.
Q. And Mr Coyne was looking at PEAKs and KELs?
A. I think in the first report he was looking mainly at
KELs, like me, and in his second report he turned on
PEAKs.
Q. And --
A. For instance, in my report Mr Coyne had commented on
nine PEAKs in his first report and I commented on those
in my first report as well.
Q. I understand. You heard Mr Coyne was asked whether
anyone provided documents for him to look at?
A. Mm.
Q. Ie ones that he hadn't found or his assistants hadn't
found. Were you provided with any documents, particular
documents by anyone to look at, have a look at this?
A. Very early on there was a tranche of about 75
architecture documents that were put to me, but I think
they were in the disclosure and that was when I was
doing my high level early exploration. Other than that,
I think the Post Office lawyers have been really trying
to get a very level playing field.
Q. Other than that?
A. Mm.
Q. So when we look at {C5/9/2}, it is a document of 30 May.
Just to give you the chronology, 30 May is the Thursday
before we start the trial, 31 is the Friday. Sorry,
2018, I do apologise. This is 30 May 2018. I was
confusing it with another document. This is 30 May 2018
when there was a question about whether or not Mr Coyne
had had the same documents as you had in early 2018.
Yes?
And if we look on page {C5/9/2}, the answer is in
the final paragraph, three lines up from the bottom:
"Dr Worden commenced work in around February 2018
and has had broadly the same information and documents
as Mr Coyne."
Yes?
A. Yes.
Q. Mr Coyne had actually asked you in an email, hadn't he?
If we go to {F/1792.2/1}, do you remember Mr Coyne asked
you in an email had you:
"... had access to additional resources that [he]
had not had access to? ... any document repositories
other than those formally disclosed? Or access to PO or
Fujitsu staff who had imparted any knowledge?"
A. I'm afraid my knowledge of the precise chronology a year
ago is not very -- now we have got it --
MR JUSTICE FRASER: I do not think it can be F.
MR GREEN: It is {F/1792.2/1}. You might need to refresh.
MR JUSTICE FRASER: That's 1792.1.
MR GREEN: We will come back to it. If you had been sure
that Mr Coyne had had the same access that you had, you
would be able to answer that straightaway, wouldn't you?
A. I was just trying to get on with the job of drilling
into this stuff.
Q. Fair enough.
A. And I'm not a good witness on precise blow-by-blow who
said what.
Q. Dr Worden, don't worry. Let's move on. Let's look
at --
MR JUSTICE FRASER: On the basis it is 12.58 am and you have
asked the witness to look at something over the short
adjournment, I think we ought to stop now.
Mr de Garr Robinson made effectively an offer about
the ground rules which I don't intend to say anything,
other than if you are going to ask any questions about
it and there is a document, it is probably best to make
sure that there are copies rather than pose it as
a memory test. But it is completely in your ballpark.
MR GREEN: My Lord, yes.
MR JUSTICE FRASER: And do you just want to remind the
witness what it is you asked him to do over the short
adjournment?
MR GREEN: Yes.
Dr Worden, I think you were very kindly going to
look at the joint report.
A. JS4?
Q. JS4. Have a look there to look at where you feel you
have commented helpfully or as you felt appropriate in
relation to that remote access issue.
MR JUSTICE FRASER: Now, Dr Worden, just to be helpful to
you, do you have a hard copy there of the joint
statement?
A. Yes, I do.
MR JUSTICE FRASER: You have, all right. Because it is
probably asking too much of a witness to do it on
screen.
A. No.
MR JUSTICE FRASER: But if you have got a hard copy it is
much easier.
We will come back at 2 o'clock. Thank you all very
much.
(1.00 pm)
(The short adjournment)
(2.00 pm)
MR GREEN: My Lord, I have already shown my learned friend
this. This is Mr Coyne's mobile telephone on which
a missed call from Dr Worden is recorded at 1.40 pm.
Can I just show your Lordship so your Lordship has
seen it.
MR JUSTICE FRASER: Yes.
MR GREEN: Dr Worden, you understand you are not supposed to
talk to anyone --
A. Yes, I noticed -- this was a pocket call. I noticed my
phone was ringing somebody -- I didn't even notice what
the number was -- and I stopped it.
Q. I understand. You will understand why --
A. Absolutely, of course.
Q. Now --
MR JUSTICE FRASER: I think the modern expression for that
is a bum call; is that right?
A. I understand it is.
MR JUSTICE FRASER: I think pocket call is a much better way
of putting it.
MR GREEN: Dr Worden, the ground rules you were given you
mentioned before lunch, could you just explain what they
were?
A. Basically that conversations between me and Fujitsu
about the case had to happen with a lawyer present who
would intervene if anything crossed any boundary.
That's basically it.
Q. I understand. So you weren't able to call to Fujitsu
without a lawyer being there?
A. Yes.
Q. To see what was being discussed. And can we get up,
please, on the Opus system -- I think it is
{F/1792.3/1}. I think I may have thought it was 1792.2.
This was the email I was referring to. This is May 2018
halfway down.
A. Yes.
Q. You didn't respond.
A. I don't recall the email. I will accept that I didn't
respond.
Q. Okay. Can I ask you this: when you were given the
ground rules, were you given the ground rules before or
after you spoke to Mr Jenkins with Womble Bond Dickinson
on the line?
A. No, the ground rules were for the trial.
Q. For the whole trial?
A. Yes.
Q. Can I just ask you whether you can remember whether you
were given the ground rules before you spoke to
Mr Jenkins or after?
A. No, Mr Jenkins was a year ago.
Q. Yes. But the ground rules were for the trial only?
A. Yes.
Q. So they didn't apply before the trial started?
A. No, but conversation before the trial was Mr Jenkins'
phone call, nothing else.
Q. I understand. So there was no prohibition on you
talking directly to Fujitsu people prior to the ground
rules --
A. No, that's not correct. The position was I wanted
a clarification in my understanding of the receipts
payment mismatch, and a phone call was set up with WB
present. Very shortly after that, and that was
May 2018, Mr de Garr Robinson said we have to be
absolutely whiter than white about this, no direct
contact with Fujitsu at all, and so everything was
through Post Office lawyers --
Q. Thereafter.
A. Yes.
Q. I understand.
Can we now look quickly at the transcript, please,
today's transcript, at page {Day18/9:1}. Do you have
page 9 at the top?
A. Yes, I have it.
Q. And it might not be quite the right page. You said
earlier in your evidence that you had had plenty of
information to look at and you were taking a top down
approach. Do you remember that?
A. Yes, I remember that.
Q. And you had looked at the sort of high level documents
of architecture --
A. Looked at a lot of architecture documents, yes.
Q. Can I ask you to look, please, at {F/1611/1} and we will
look at the first page to see what it is. This is
a Post Office board agenda from 31 January 2017. Do you
see that?
A. Yes.
Q. If we go to page 100 of that document, please
{F/1611/100}. Now, this is in the context of looking at
the Horizon architecture, and can I take you back to
page {F/1611/87} first, please. Just look at the --
this is a "Technology Strategy Update"?
A. Yes.
Q. I just want to take all these documents reasonably
briefly. The first question is, is this the sort of
type of document that you would have had regard to when
compiling your report, or not?
A. Well, it is certainly not the type of document I was
interested in in the early stages of putting together
what I called the foundation sections of my report.
Q. What about later stages?
A. Well, I would like at these documents typically in
response to seeing references in Mr Coyne's report.
Q. If we look at the second line, at the context there?
A. Context?
Q. Do you see under "Context"?
A. "IT not fit for purpose", that is right.
Q. It is fair to say, isn't it, that where Post Office
themselves are recognising the limitations to their own
system, that is likely to be a fair and sensible view?
A. This is Post Office talking about the whole IT estate.
Q. Yes.
A. And they put their phrase "not fit for purpose" in there
and they were obviously discontent, yes.
Q. And if we look, if we may, please, at page {F/1611/98}
A. Sorry, could I check the year of this?
Q. 2017.
A. 2017, right. Okay.
Q. If we look at page 98 you will see paragraph 25 down at
the bottom.
A. "There are tensions in each contract." Yes. (Pause)
Q. You see Fujitsu:
"... a 6 year fixed contract signed with PO which
continues to invest in legacy and obsolescence where FJs
own strategy globally is to move to cloud."
Do you see that?
A. Yes.
Q. So that's their own perception of the system and the
situation they are in?
A. That's PO's perception of Fujitsu's strategy, yes.
Q. Yes. If we look at page {F/1611/100}, please. This has
got senior people like Tim Parker, the chairman, Angela
Van Den Bogard, also attending, various others. On
page 100 if we look at paragraph 35, for example, do you
see there "For Retail"?
A. Yes.
Q. They are specifically talking about HNG-X, which is the
Horizon system?
A. Right.
Q. So it is not the general IT environment, is it?
A. No, that is Horizon specific --
Q. It says:
"The Horizon (HNG-X) platform is at the end of its
life and needs replacing."
Yes?
A. Yes.
Q. That's fair comment, isn't it?
A. It is interesting, that. I mean, 20 years is a long
life and that's what they were saying, yes. Personally
I think it is doing quite well, but there we are.
Q. "Previous attempts to move away from HNG-X platform,
specifically with IBM, have been unsuccessful."
A. Yes, I didn't know about that.
Q. Did you see this document at all, can you remember?
A. I have seen reference to this end of life quotation at
some time, but I haven't read this document in detail.
Q. You didn't follow this one up?
A. No.
Q. Let's look, please, at {F/1603/5}. So that document was
31 January 2017, this is 17 January 2017.
MR JUSTICE FRASER: I think we are at 1603, but did you want
to be at 1603 --
MR GREEN: Yes, 1603, this is 17 January 2017. If you look
at "Credence", do you see that?
A. Yes. (reads to self)
Q. Do you see "Credence" underlined?
A. Yes, I'm just reading that paragraph (Pause).
Q. Do you see where it says Accenture picked up a difficult
pass?
A. Yes, and additional costs at the moment.
Q. Starting to see light at the end of the tunnel. Then
Fujitsu. Do you see that heading?
A. Yes. Not hold the power.
Q. "FJ see the contract as a cash cow, so need to persuade
them that working with POL to migrate to cloud
technology is to their benefit against a 'too good'
contract."
Did you see this document at all?
A. I didn't read these pieces -- this page.
Q. Looking at it now, does that chime with a sense of
anything you have seen?
A. Well, that Fujitsu paragraph, it sort of chimes because
I have always had the impression from the governance
structure and the documentation and so on that Fujitsu
were not short of budget, really.
Q. They weren't, okay. And at least not short of budget
coming from Post Office?
A. That is right.
Q. If we go forward, please, at {F/1586/3}. Perhaps you
should see the first page, November 28, 2016. Then if
we look at page 3 as it comes up?
A. That's interesting.
Q. Yes?
A. Okay.
Q. If we look at the first column:
"Horizon software was developed in 1996, originally
as a DSS IT project."
Yes?
A. Yes.
Q. There was a heavy part of it which was DSS, but also for
the Post Office is what happened, isn't it? You know
about that?
A. I hadn't gone into the detail of that prior project but
I know it existed.
Q. DSS wanted a big system and Post Office went to be
a joint procurer. Then the DSS dropped out and the
Post Office was left holding the contract?
A. I'm just aware of that pre-history but have not gone
into it.
Q. "At the time Horizon was one of the first electronic
Points of Sale.
"Horizon was created before the internet had any
real effect on Retail or Banking.
"It was built as a 'closed' system & designed based
on paper processes, is clumsy & operator unfriendly."
Now, were you aware of that recognition, either in
this document or anywhere else, that Post Office
internally recognised that it was operator unfriendly?
A. Well, this is a document in 2016 talking about 2001.
Q. No, I'm just asking you a clear question, Dr Worden. It
says it is clumsy and operator unfriendly. Were you
aware from this document or from any other document that
that was Post Office's internal view of Horizon?
A. I was not aware from more contemporary documents and
I had not read this one.
Q. Thank you. Did you say from more contemporary
documents?
A. No, I mean I do not think I have seen a document before
2000 -- I mean, the case for moving the HNG-X included
"Let's get rid of Escher", and so on. It didn't include
"God, the interface is clumsy".
Q. Okay, so you weren't aware of that?
A. I wasn't aware of that --
Q. -- in forming your views you have expressed?
A. No.
Q. Let's look at the third column, if we may, "IBM deal
ends". And the only bit I would like you to look at
really there is halfway down, third bullet point:
"However, whilst modernising the 'front end' is
relatively straightforward, as the project developed we
realised that shifting the 'back-end' is extremely
difficult."
A. Yes.
Q. So they reverted back to Fujitsu.
A. Yes.
Q. And then if we look on the right, "HNGA":
"Whilst HNGA runs on updated Windows software,
fundamentally its architecture is the same as HNG-X."
That is fair, isn't it?
A. Yes.
Q. "This means that it remains a 'closed' and inflexible
platform that retains the complexity of transaction
journeys and operational requirements - not Simpler to
Run!"
A. Yes.
Q. I think it appears to follow from not having read this
document that you may not have appreciated that when you
were putting your reports together?
A. No. But if I may comment briefly on the shifting the
back end extremely difficult, that seems to me to chime
with my understanding that the bulk of the investment
actually was in the back end and that's where the
complexity was.
Q. And the front end was the bit the subpostmaster was
using?
A. Yes, and I would guess, and it is a bit of a guess,
that, you know, in terms of lines of code, complexity
and so on and so forth, the back end is more than the
front --
Q. So more had been invested in the back end than in the
front end that the subpostmasters had been using?
A. I feel broadly the back end is more complex.
Q. {F/1557/1}, please. This is a month earlier than the
previous one, 22 October 2016.
MR JUSTICE FRASER: Just before you go there, do you know
what the expression "thin client" -- how would you thin
client --
A. Thin client generally nowadays corresponds to a client
that's just a web browser and it goes through to some
internet site. I think that's pretty much the meaning
now.
MR JUSTICE FRASER: That's perfect.
Mr Green.
MR GREEN: This is a document from October 2016, and on
page 1 under "Context" would you look at paragraph 3:
"Our back office"?
A. Struggles, yes.
Q. "Our back office also struggles with the complications
of dealing differently with each of our many clients,
heavily manual processes ..."
Do you see that?
A. Yes.
Q. "... reconciling disparate sources of data,
retrospectively financial controls and a lack of
flexibility."
A. Yes, and this is referring to the Post Office bank.
Q. Yes. And it says:
"This backlog of challenges, poor support contracts
and a lack of skills have led to a prohibitive cost of
change preventing the improvements that should occur as
a part of business as usual."
Do you see that?
A. Yes.
Q. Now, first of all, did you broadly pick up that
impression from anything else or not?
A. There were various other documents, like finance roadmap
and so on, which conveyed a general impression to me
that the Post Office back office in its IT and its use
of POLSAP and various different SAPs and so on, and the
business processes around it, that they had to interface
with the clients, but that back office process was more
complex than the Horizon back office.
Q. I understand the complexity point. I'm talking about
the satisfactoriness point. Did you get that
impression?
A. Well, obviously from these documents Post Office was
unsatisfied about something.
Q. You hadn't picked that up yourself?
A. Well, I picked it up when I started looking at the
documents cited by Mr Coyne.
Q. Right. Did you expressly deal with that anywhere?
A. Well, it seemed to me not really directly part of
Horizon.
Q. Okay. I do not think you have dealt with these other
documents we have been to already, have you?
A. None are very familiar.
Q. Okay.
A. But as I say, I did start looking at these Post Office
IT strategy documents when Mr Coyne cited them.
Q. Let's go a bit earlier, 29th August 2016. {F/1522/1},
please. Let's look at paragraph 1. Can you look at the
right-hand side three lines down:
"The Back Office process and applications remain
complex, unreliable, expensive to maintain and not
suitable for today's business."
A. Yes.
Q. Did you look at that document?
A. Again, I can't recall looking directly at this document.
I'm not brilliant on looking at this --
Q. I understand.
A. -- etc, but it doesn't surprise me.
Q. And it doesn't surprise you that they mention that it is
unreliable?
A. This is the Post Office back office --
Q. Yes.
A. -- things like POLSAP and all sorts of stuff --
Q. All things they --
A. Much more than that, the whole of Post Office business.
Q. Yes, the whole thing, and that didn't surprise you that
it was unreliable?
A. They say unreliable. It didn't surprise me it was
unwieldy.
Q. You said this didn't surprise you a moment ago. You are
just rowing back on that a bit. Bit worried it is
damaging to Post Office to say so?
A. Unreliable, good question. Well, they say that and that
seems to be part of the package, yes.
Q. You said it is a good question, the reason why I'm
asking is reliability of the overall process at least on
one view is a question which may be of relevance in this
trial, isn't it? Including reliability of the back
office aspects. Is that fair?
A. Well, there has been a bit of prohibition about looking
at Post Office business processes.
Q. Who prohibited you?
A. Well, people said that's out of -- I got a flavour that
Post Office business processes have been out of scope.
For instance in TCs, the process for creating TCs has
not been a thing the experts have looked at.
Q. You sort of said people said and then you said you got
a flavour?
A. Yes.
Q. Can I ask you, did you form a view yourself about
whether that was out of scope or were you told not to
look at it?
A. I was told that it was out of scope, and I found some
difficulty with that because in a sense things like
robustness of Horizon actually depends on all sorts of
things and one has to try and assess the consequences of
certain things, where the causes, how business processes
work and so on, are out of scope. And it is a bit of
a blurry boundary.
Q. We will come back to it later, but you refer to
transaction corrections and transaction acknowledgements
and things as very important countermeasures in your
report?
A. That is right, yes.
Q. So they were central to your analysis?
A. They are essential to my analysis.
Q. And they rely on Post Office back office systems that
we've been talking about?
A. The approach I took to that is transaction corrections
are a corrective measure, they are a countermeasure. It
is obvious they work a very large proportion of the
time, and I tried to work out numerically the small
proportion when they didn't work.
Q. We're going to get to that in some detail, but for
today's purposes I'm only asking about relevance. So
where we are is you were told it was out of scope. You
had difficulties with that because you felt that that
was an important countermeasure --
A. Well, I think I reconciled it to myself in that certain
causes were out of scope but the effects were in scope.
Q. I understand.
A. Does that make sense?
Q. We will explore whether it makes sense or not. Against
that background you hadn't had any particular regard to
a document we see at 1522?
A. Well, the view I have taken is that the Post Office
whole IT estate is much bigger than Horizon.
Q. I understand. But it is right, just to be clear, that
the Post Office back office estate, yes -- back office
systems that we have been talking about just now --
A. Yes.
Q. -- are relevant to the issue of data integrity,
transaction corrections, transaction acknowledgements
and those things?
A. Indeed they are. As, you could say the same thing,
certain client IT systems are relevant.
Q. Absolutely. So if you had concerns about data integrity
coming in from Camelot on National Lottery, that would
be relevant as well?
A. Yes.
Q. Okay. Let's look at {F/555/1}, please. Let's go a bit
earlier. Let's go before 2010. My screen has gone
completely --
MR JUSTICE FRASER: It is because the F folder is so big.
Give it a second, I daresay it will appear in a moment.
MR GREEN: Thank you very much.
So this is a document from December 2009,
7th December 2009, and if we could go to page
{F/555/10}, please. This is the 2009 appraisal,
internally:
"Horizon - Current State."
It says, as at 2009:
"13 year old design and technology to satisfy
a different business."
A. Yes.
Q. "Evolved rather than designed"?
A. Where does it say that?
Q. Third bullet point.
A. I see. I was looking at the footnotes, sorry.
Q. Just look at the big points, the ones they thought were
really important.
A. (Reads to self) Yes.
Q. And "slow and expensive to use". Do you see that in the
middle?
A. Which year was this assessment?
Q. This was 2009.
A. And that's what they are saying there, yes.
Q. Had you looked at this document?
A. Again, I do not think I have seen this actual document.
Q. Well, let's look at the bottom four lines, please:
"Horizon is also a system that is wrapped up in
'barbed wire' - making changes difficult and costly -
test everything!"
Do you see that?
A. Sorry, where is "test everything"?
Q. If you look at the bottom four lines of the text, at the
bottom of the page, four lines up from the bottom --
A. (Reads to self) Yes, I must admit this is different
from my own understanding of Horizon.
Q. This is different from your own understanding?
A. Yes.
Q. Okay. And:
"Design was optimised at the time to minimise costs
..."
Do you see that, at the bottom?
A. "... ([especially] network) - offline -- what's their
reference to offline working?
Q. ?i think it's probably because internet access was more
expensive in those days.
A. The question then is what was it at the time: was it
original Horizon or HNG-X? This is before HNG-X, so
this is the original design. Right, okay.
Q. Original design. But you fairly accepted that that's
different to the view of it you formed in your reports?
A. Yes.
Q. And you probably hadn't seen this document?
A. I hadn't, no.
Q. Can we look now at two individual documents relating to
individual subpostmasters which we know about because
they were lead claimants in the November trial.
Let's look, please, at {F/68/1}. You will see that
this is an email and I would like to take you over the
page to page {F/68/2} where the chain begins, if I may,
please. You will see there that it is from
Frank Manning. Do you see halfway down?
A. From Frank Manning, yes.
Q. See that? And we look up two lines, it is to Sue Lock?
A. Yes.
Q. The subject is "Horizon matters --
A. Yes.
Q. -- Barkham SPSO"?
A. Yes.
Q. And it explains in the third paragraph:
"The balances are a mess ..."
Do you see that?
A. "Balances are a mess", yes.
Q. "... (in pre-Horizon times - the Postmistress virtually
achieved a clean balance every week) ..."
Do you see that?
A. Yes.
Q. Do you see "but" in bold?
"... I worry that something like 25 re-boots in one
day is having an effect overall."
A. Yes.
Q. And if you look at the second paragraph on that email
you will see he says:
"I visited there today & was too scared to accept a
cup of tea in case the Horizon system crashed cos the
electricity supply is still a live (excuse the pun)
issue."
A. Yes.
Q. Now look at the bottom big paragraph, before "help
please", the one that starts "Need your best offices".
Do you see that?
A. "Need your best offices", yes.
Q. "Need your best offices to get this case to a proper
solution - she keeps getting promises of attention - but
nothing is actually being done now to clear up the
problem. It is Horizon related - the problems have only
arisen since install & the postmistress is now barking &
rightly so in my view."
A. Yes.
Q. "Help please."
Now, it is fair to assume, isn't it, that where
Post Office internally recognised that Horizon was
likely to blame, or in this case positively asserts that
it definitely is, it is quite likely to be right?
A. Well, I don't know who Frank Manning was or what kind of
evaluation he had done. I mean, it seems he has made
a visit to the site.
Q. Indeed. Do you think --
A. And on that basis he had seen that apparently there are
lots of power cuts.
Q. 25 reboots in one day. Do you think that would have
an effect to --
A. Well, it seems -- I'm just guessing because, you know,
I haven't studied this evidence in detail, but it seems
to me electricity supply is something to do with the
reboot.
Q. Yes. But do you accept it is quite possible that
a problem with the electricity supply causing 25 reboots
might well be the source of problems in her accounts and
balances on the basis of this document, or have you got
some other explanation that you would like to --
A. "Balances are a mess ..."
Q. You can't diagnose it at this distance in time?
A. No.
Q. I'm only putting to you do you think it may be fair that
that may well be right?
A. It is certainly possible.
Q. But is it fair to accept that it may well be right on
the face of the document?
A. I accept it may well be right.
Q. Let's go back to page {F/68/1}, please. This is to
Kevin Cox from Sue Lock. It says:
"Frank came to see me about this office and we
discussed it with Sanjay and said that she needed to
prove that it was Horizon that was causing all these
power failures in the office."
A. Yes.
Q. "Can you tell us please how we can now get this resolved
as it appears now it is a direct consequence of the
installation and not anything that has happened in
steady state."
A. Yes, that what it says.
Q. It is fair to assume that may well be right?
A. Well, it is a puzzling paragraph because Horizon causing
power failures --
Q. I'm not suggesting Horizon is causing the power failure.
A. That's what this paragraph says.
Q. I'm suggesting that power failures and lots of
recoveries are quite likely to cause balances to be in
a mess at this branch, aren't they?
A. All I'm saying is that the phrase "Horizon was causing
these power failures" indicates to me some possible
confusion about cause and effect.
Q. You don't know this, Dr Worden, but what in fact
happened is this had been all installed as the system.
A. My general view would be that Horizon has been built to
be robust against power failures, and therefore the
prior expectation is that there can be a number of power
failures which do not cause discrepancies in accounts.
Q. So you would not have expected that outcome on the basis
of your understanding of the Horizon system --
A. My understanding is that power cuts happen approximately
once per branch per month, I believe, or was it once per
branch per year. I'm sorry about that, but they do
happen and Horizon has to be robust against them.
Q. And what we have seen is not consistent with your
understanding of its robustness?
A. That is right.
Q. If we look, please, at another document, last example,
{F/99.1/1}. Now, this is an audit in 2001 of Mr Bates'
branch. You probably read his name. He is the lead
claimant.
A. I think I know what he looks like.
Q. Bates v Post Office?
A. I think he is here, actually.
Q. Yes. This is his branch. If we could go forward,
please, to page {F/99.1/4}, if you go down to "Cash
Management" heading, do you see that? Halfway down?
A. Yes.
Q. Go down to the next heading "Control Gaps - High Risk".
A. Yes.
Q. The cash holdings were 25% higher than the ideal target
holding.
A. Yes.
Q. And then underneath "Comments":
"A correct assessment of cash holdings could not be
made because the Horizon system intermittently adds the
previous days cash holdings to the daily declaration."
A. So that is what the auditor is saying?
Q. Yes. The auditors are very familiar with the system,
aren't they?
A. I am sure they are.
Q. And that is also not consistent with your understanding
of how Horizon works, or should work?
A. I haven't seen, I think, other evidence about adding
previous day's cash --
Q. I'm not asking whether you have seen other evidence, I'm
asking is that consistent or inconsistent with your
impression of Horizon's robustness or reliability?
A. That is inconsistent.
Q. Can we now move on, please, to documents recently
disclosed, so new. So the Friday before the hearing was
31st May, Thursday was the 30th and the Wednesday was
the 29th.
A. Fine.
Q. So let's start with documents closed on Friday, 31st May
or uploaded to the bundle on the 31st. I will just show
you.
A. It is uploaded to the bundle we are talking about.
Q. It is in the bundle. {F/1834.14/1}
A. I have tried to look at some of the things recently in
the bundle. I haven't done it all.
Q. Dr Worden, we just want to know what you have looked at,
what you haven't and what you made of it, that's all.
So {F/1834.14/1}.
A. Yes.
Q. This document as a document was first disclosed on
31st May, but it contains information from the MSC
spreadsheets which have been disclosed a few months
prior.
A. Right, okay. Yes.
Q. Now, have you seen this document before?
A. No, I haven't.
Q. Do you have any idea why it has been uploaded?
A. No, I don't.
Q. Can I show you just how it works. If you look at the
top of this one, you have the MSC number in the middle?
A. Yes, middle of the page.
Q. There is the heading, 043J -- I probably should have
said -- 043J0348236. This then explains:
"Below is the extracted text relating to," this MSC,
"taken from Rows 284594 to 284654," in one of the three
documents to which one had to put together the MSC logs?
A. Yes.
Q. Did you find the MSC logs in their original form easy to
use?
A. I didn't find them too problematic. They were clumsy.
They are monster spreadsheets, but what I found is one
of them, I think it is this one that the extract is
from, was fairly narrative and you could go to
a specific MSC just using an Excel search and you could
read the narrative. And, indeed, I actually built
a JAVA tool to filter the MSCs and get a smaller
spreadsheets of things I wanted to --
Q. Okay. Let's look at this one. We can see what's being
done because the third paragraph says:
"The proposed change is to insert 1 dummy
Transaction Acknowledgement (TA) row for Branch 74005
into the Branch Database (via an SQL INSERT statement)
with transaction amount value zero and a quantity of 1."
Yes?
A. Sorry, I'm not there at the moment.
Q. It is the third paragraph at the top.
A. I see --
Q. Third paragraph --
A. (Reads to self)
Q. Okay?
A. Yes.
Q. Then do you see "Justification for the change and
urgency"?
A. Yes.
Q. "POL have requested this issue be resolved," and there
is a POL incident reference, Q17628223, yes?
A. Yes.
Q. And from there we can see the text about it, and if we
go over the page {F/1834.14/2}, at the top of the page:
"POL approval required (Y or N) ..."
Do you see that?
A. Yes.
Q. And it says:
"No," so POL approval isn't required, "- POL aware
..."
And gives the same reference that we have seen on
the previous page?
A. Yes.
Q. So what we learn from this MSC on a fair reading is that
POL have asked for something to be sorted out, this was
how it was going to be done, POL was aware, Post Office
was aware and so they didn't need any approval because
they were already aware, they had asked them to do it.
Yes?
And we can see the other questions that follow on
that page. Go forward one page, please, {F/1834.14/3}.
You see:
"SSC (managed by Steve Parker)."
At the top?
A. Yes.
Q. We go over the page again {F/1834.14/4}, you have to
note here at the top it says:
"Below is the extracted text relating to MSC
043J0348236 taken from Rows 121273 ..."
A. Yes, this is another one of the spreadsheets.
Q. Yes. So what they have done is they have put the rest
of the information about the MSC on --
A. On the same MSC number.
Q. On the same MSC number. So they are right to do that,
aren't they?
A. Yes, absolutely. Absolutely.
Q. That is the only right way to do it?
A. That is the correct way to do it.
Q. And you can see expected impact, testing, security
implications:
"Do users need to be informed of the change?
"No."
Etc.
Do you see that?
A. Yes. And this was inserting a TA?
Q. Yes.
Then let's look on page {F/1834.14/5}, please. It
says:
"The information relating to [this MSC] cannot be
easily extracted from," the complete data spreadsheet?
A. Well, there are three of them. Are we talking about
different ones of the three now?
Q. Yes. So they have two in there, but not the third one
because it is too difficult to get out.
Now, if we now compare that one with one of the
others also that was provided on Friday, 31st May in
a similar format, {F/1834.12/1}, the same principle. It
explains the text below, yes?
A. Yes.
Q. "Below is the extracted text". Top line:
"Below is the extracted text ..."
Do you see that?
A. Yes, that is the one that I have looked at --
Q. So you looked at this one, have you?
A. No, not this particular MSC, but I'm saying the source
of my information on MSCs is mainly the narrative one.
Q. Let's look at what the non-technical overview says. It
says on the right-hand side:
"This causes discrepancies in Stock declarations
report."
Do you see that?
A. Yes.
Q. So:
"Older unused branch declarations are being picked
up in Branch Declarations Report in the same trading
period ... [or] balance period ... (BP) from previous
year occurs again."
A. Yes.
Q. So it is picking those up and causing discrepancies in
the stock declarations report?
A. Yes.
Q. At the bottom under the second line of asterisks you
will see:
"Steps to carry out the fix."
A. Yes.
Q. And it is:
"Logon to BDB Host Node 1 as 'oracle' user."
A. Yes.
Q. Then there and over the page we see the scripts being
used {F/1834.12/2}.
A. Sorry, can I just read that? (Pause)
Q. I'm not going to ask you about the scripts, I just want
to show you aspects of the document.
A. Yes.
Q. If you look at the bottom --
MR JUSTICE FRASER: Of? Bottom of page 2.
MR GREEN: On the bottom of page 2. There is a mandatory
field:
"POL approval required (Yes or no)."
A. Yes.
Q. We see there "no" on the next page {F/1834.12/3}.
A. I see, on the next page, right.
Q. So it looks as if there POL approval was said not to be
needed?
A. Yes.
Q. And if we look further down we can see there is
a "delete SQL" being applied?
A. Yes.
Q. And this is effectively to try and remedy the stock
discrepancies problems, isn't it?
A. It is remedying a problem in BRDB, yes.
Q. And we go over the page to page {F/1834.12/4}, please.
And it says:
"Who will action this change ..." at the bottom.
Then go over the page to page {F/1834.12/5}, "RMGA
UNIX"?
A. Yes.
Q. So that is the end of that one. So we don't have there
the data from the other spreadsheet that we did have on
the previous one we looked at. Do you remember the
first one had two spreadsheets output?
A. Yes.
Q. Now, what we now need to do is we need to look in the
spreadsheet itself at {F/1843/1}.
A. Can I get it on this screen?
MR GREEN: It will come up.
MR JUSTICE FRASER: It might take a moment, but it will
appear currently.
A. Previously I didn't get spreadsheets on those screens at
the back there for some reason.
MR GREEN: What we are going to do, Dr Worden, is we are
going to look on the second spreadsheet that we haven't
got reproduced within the Womble Bond Dickinson document
and just see what's in there.
Because what we are looking for is on row 90474 --
don't scroll, please. Could you go into the A1 thing
and type "A90474"?
A. There we go.
Q. There we go. And scroll down a tiny bit, if you would,
so we can look below that?
A. I see, yes. So this is all the impact stuff?
Q. This is all the impact stuff which is not included in
the document prepared by Womble Dickinson.
You see there 1:
"What is the expected impact on the live/production
service of implementing this change?
"None.
"2. What level of testing ...
"N/A.
"3. What are the security implications ...?
"None.
"4. Do users need to be informed ...?
"No."
Regression path, and then:
"8. Has the change been implemented before? If yes
give details of the previous implementation impact and
success criteria:
"N/A.
"9. Taking into account all of the above what are
the worst case risks implementing the change?"
You see that?
A. I have read through. I am not quite in sync with you
but I have read through that case.
Q. Number 4 is:
"Do users need to be informed of the change," and
the answer is "no".
A. Yes.
Q. Now, if we go, please -- and close that spreadsheet --
to {F/812.1/1}.
Now, Dr Worden, this is a PEAK when it comes up. It
is {F/812.1/1}.
A. So this is 2011. Is it referring to the same MSC?
Q. It is referring to the MSC, but I'm going to -- because
it is just about to be 2.50 and it is a new document,
I'm going to ask his Lordship if that's a convenient
moment to have a break and perhaps you can have a glance
at it if you want to.
A. I mean, his Lordship said something about paper copies.
Now, would it be possible to have paper copies of PEAKs
or is that a bit difficult?
MR JUSTICE FRASER: Even my paper copies of PEAKs have
disappeared, although I'm expecting them back at some
point. But I am sure --
MR GREEN: My Lord, I do not think it is going to be easy
for him to look at various pages.
MR JUSTICE FRASER: How many pages is this PEAK?
MR GREEN: It is reasonably long. It is eight pages long.
MR JUSTICE FRASER: Does anyone on your side of the court
have an unmarked copy of the eight pages?
MR GREEN: No, I don't think so.
MR JUSTICE FRASER: So what is it you would like the witness
to do?
MR GREEN: If he could look at page 2 and just read that,
that would be helpful.
MR JUSTICE FRASER: Over the 10 minutes or now?
MR GREEN: Over the 10 minutes if that's not --
MR JUSTICE FRASER: Would you prefer for us to break for
10 minutes to that you can look at it carefully, or
would you -- it really is not an issue one way or the
other, it is whichever you find easier to do.
A. If it is convenient for a break now, it is nice to look
at one. But I suspect there will be loads of PEAKs
where I won't get a break.
MR JUSTICE FRASER: At the moment I'm unclear whether that
is the case or not.
MR GREEN: Shall we break now, my Lord?
MR JUSTICE FRASER: No, we will. I'm just looking at
logistics if this is going to be replicated tomorrow or
the day after.
MR GREEN: My Lord, yes.
MR JUSTICE FRASER: It is just some people are used to
working on screen, some people are used to working on --
MR GREEN: I understand.
A. And I have a particular way of looking at PEAKs.
MR JUSTICE FRASER: That is entirely understood, Dr Worden.
So after today, counsel is going to make sure that he
has got a hard copy for you.
A. Excellent, thank you.
MR JUSTICE FRASER: Who knows, by then I might even have my
hard copy file back, so I might be able to do it both on
screen and hard copy.
For the moment we will break until 3 o'clock for the
shorthand writers and for you also to have a look at it.
If, when you have read that page, you want to look
through subsequent pages, just ask either of the counsel
who will be standing there on the claimants' side of the
court and they will ask the Opus people and they will
flick it forward on the screen.
Mr de Garr Robinson, does this seem like a sensible
way forward?
MR DE GARR ROBINSON: My Lord, yes.
MR JUSTICE FRASER: All right, I will come back in at
3 o'clock.
(2.51 pm)
(A short break)
(3.00 pm)
MR GREEN: Dr Worden, so if we can just start on the first
page of that PEAK, please, if you go back a page
{F/812.1}.
A. First page, okay.
Q. We have the PEAK number at the top, which is PC0211010?
A. Yes.
Q. And at the bottom of the page under the "Impact
Statement", it says user "unknown", 23rd June 2011. Do
you see that?
A. 23rd June 2011. Is that right?
Q. User unknown, date 23rd June 2011. Then the problem is
explained:
"Branches will be forced to declare stock when they
don't want to. Apparent reappearance of withdrawn stock
may cause spurious discrepancies."
Now, Dr Worden, to give you context we have only
found this document by trying to trace through what was
happening on that MSC.
A. Yes. Could I ask about the linkage to the MSC?
Q. Yes, I will show you. It is page {F/812.1/3}.
A. If we go back to the MSC and the beginning of the MSC,
I think --
Q. Can I take it in stages and just show you --
A. Well --
Q. Page 3 at the bottom, do you see 24th June 2011?
A. Yes.
Q. Last yellow box?
A. "Reference Added", that is the last page of the PEAK.
Q. That is the last page of the PEAK.
A. And it refers to that MSC.
Q. It refers to that MSC.
A. Fine.
MR JUSTICE FRASER: I think it is on the first page as well,
isn't it?
MR GREEN: I think it is also.
MR JUSTICE FRASER: At the top.
MR GREEN: Yes, it is MSC on the first line under the
references value. Do you see that?
A. Yes, and that is the MSC we just looked at?
Q. That is the MSC we just looked at, okay?
A. That's my puzzlement, actually.
Q. But I'm just asking you --
A. We will come to my puzzlement when we do.
Q. Okay. For the moment you can see that the MSC refers to
this PEAK?
A. The PEAK refers to the MSC, I will check --
Q. Sorry, you are quite right. The PEAK refers to the MSC.
A. Yes.
Q. And we have seen what the problem is: spurious
discrepancies at the bottom?
A. That's the PEAK problem, yes.
Q. Now, pausing there, had you identified this document in
your researches?
A. Well, I can't say I have seen this PEAK before, but when
you go to page 2 in the Anne Chambers piece on page 2,
it is one of these recurring every year when the same
trading period comes around effects, which is just like
the suspense account book.
Q. This is a slightly different one, but --
A. It is similar.
Q. But pausing there, I was asking you had you seen this
PEAK.
A. I do not think I had, no.
Q. Right. If we go over the page to page {F/812.1/2}, you
will see in the yellow box --
A. I don't have colours.
Q. Second box down:
"PC0208335," which we won't look at but for his
Lordship's note is {F/773/1}, "addresses the need to
remove old branch declarations from the BRDB so they are
not picked up and reused when a stock unit reaches the
same TP/BP a year on."
Do you see that?
A. Yes.
Q. So there had been a previous PEAK about it?
A. Yes.
Q. Then she says:
"I have now found that there are some old
declarations which were created but never updated, which
are not removed by the SQL used for both the MSC and the
code fix."
Do you see that? We will look at what that means
and what it is referring to in a moment:
"In particular there are 36 stock declarations which
will almost certainly give branches problems when
balancing between now and the end of 2011."
A. Yes.
Q. And you would agree, wouldn't you, that's not how
Horizon is supposed to work?
A. Yes.
Q. The options discussed are:
"Can we either:
"a) change the fix already in the pipeline ...
(I know this is probably not possible) ...
"b) schedule and produce a fix which also targets
these extra declarations ..."
Do you see that?
Then the next paragraph, bottom of that box:
"If the MSC can be run by 1st July, it should not be
necessary to contact POL or any branches to get them to
take action to remove the declarations themselves (since
we confirmed to POL earlier this year that we had
already taken steps to prevent further instances, this
would be a good idea)."
A. Yes.
Q. So what they are trying to do is get it done quickly so
that the fact that the previous fix didn't fully work
doesn't show up?
A. I think that is right.
Q. That's fair, isn't it?
A. Yes. I think that the reference to the previous MSC
wasn't done quite right and it left some bits over.
Q. Yes. So the previous fix left some bits outstanding?
A. Yes.
Q. Because it wasn't done quite right. Post Office Limited
thought it had been fixed and we know that the problem
first occurred, if we go back to page 1, in
February 2011?
A. Right, okay.
Q. Look at the bottom of the page, first occurred 2011?
A. Yes.
Q. Go forward again {F/812.1/2} and "updated" at the bottom
of that page, Anne Chambers says:
"Apparent reappearance of withdrawn stock may cause
spurious discrepancies."
That is the phrase we see throughout.
Then if we go over the page, please, to page
{F/812.1/3}, it says there at the second box down:
15:07:16 20 "23rd June 2011 at 17:06:26 User: John Simpkins
"POL are aware of the declarations issue which was
not completely fixed by the previous release, I do not
yet think that they realise that fix was not complete so
a quick resolution to this would be good."
A. Yes.
Q. So what he is saying is: they think we fixed it last
time, if we fix it quickly they may not realise it?
A. That's pretty much what was on the previous page as
well, I think.
Q. Yes, and that tends to confirm it?
A. Yes.
Q. So if we go to page {F/812.1/6}, please, and we look at
the top box:
"RISKS (of releasing or not releasing proposed fix):
"What live problems will there be if we do not issue
this fix?
"Incorrect branch declarations belonging to a
previous year will be picked up and cause stock account
discrepancies to a live trading branch."
Do you see that?
A. Yes.
Q. So the position we have now is that Post Office doesn't
know the previous fix hasn't worked?
A. Yes, Fujitsu had told Post Office we fixed it.
Q. We fixed it --
A. And they weren't quite right, so they are going to tidy
up before it happens.
Q. So the NBSC haven't been told the fix didn't work?
A. Whether it is Post Office or --NBSC would be part of the
chain somehow, T.hey won't know.
Q. And they are going to try and fix it before Post Office
realised it?
A. They are going to try and head it off at the pass before
these branches run into it.
Q. So neither the users nor Post Office were informed about
this fix being done on the face of these documents?
A. That is right. What they are saying is we haven't
completely sorted it, we can sort it before any damage
is done.
Q. That's what they hope to achieve.
A. Yes, they are hoping to do that.
Can I mention my puzzlement about the MSC, or is
that --
Q. By all means, if it is helpful to the court?
A. The MSC, when I read it I thought it was to do with
a transacting acknowledgement for zero pounds, and that
seems to me to be a different view from this.
Q. No, there were two separate MSCs. The first one
I showed you that was very fulsome explaining PO
approval not needed, yes? The fulsome one was for TA
zero pounds?
A. I was just confused.
Q. I understand. The second one didn't have all the PO
approval detail from the second spreadsheet.
A. Right, okay.
Q. So we had to go and find it in row whatever 100,000, dig
it out, then search for the PEAK and, through that, find
out what actually happened, which I have just put to you
and you agreed with.
A. So the second MSC didn't actually say what was going to
be done. I was confused between the two.
Q. I understand. Don't worry.
A. Okay.
Q. Dr Worden, can we move now, please, to the approach you
have taken to Horizon Issue 1. Can I ask you, please,
to look at page {C1/1/1}. It will come up on the
screen. There's the top of the page. There is
a definition of the Horizon system. Yes?
A. System? Yes, right.
Q. "The Horizon system ..."
A. Okay.
Q. "... shall for the purposes of this list of issues mean
the Horizon computer system hardware and software,
communications equipment in branch and central data
centres where records of transactions made in branch
were processed," as defined in the pleadings, yes?
A. Yes.
Q. And that doesn't include the manual processes of
deciding whether to issue TCs or not, does it?
A. Right. That is correct.
Q. When we look at Issue 1, the question that's posed is:
"To what extent was it possible or likely for bugs,
errors or defects," any of those, "of the nature
alleged" and so forth "to have the potential to (a)
cause apparent or alleged discrepancies or
shortfalls ..."
A. Yes.
Q. And it doesn't say "to cause" without being first
qualified by "to have the potential", does it? The
words "to have the potential" are important, aren't
they?
A. What's the distinction between having the potential and
actually doing? I mean, if it doesn't actually do, then
in some sense it didn't have the potential.
Q. Well, let's put it this way: something might have the
potential to do something which you can identify, but
you can't identify whether it has actually done it;
that's fair?
A. Yes, that's fair.
Q. So whether it has actually done it and whether it has
the potential to do it are two different things in that
respect?
A. In a sense the word "potential" reflects a lack of
knowledge. If I knew more I would know whether it is
going to happen or not, but I know less so it is
potential.
Q. Yes. As you fairly accepted earlier, that the Horizon
system by its nature has been quite difficult to plumb
for the experts?
A. To plumb in that sense, yes.
Q. Indeed, you refer to a swamp and so forth. I mean, it's
not the most transparent task you've been engaged in?
A. It is a very big system to understand in a year, yes.
Q. So just coming back to this, what we can say with
confidence, if something actually does cause
a discrepancy in a branch --
A. Then it did had the potential.
Q. -- then it plainly had the potential?
A. Yes.
Q. So potential is necessary but not sufficient to have
an actual impact?
A. Yes, I agree with that.
Q. So it says to have the potential, and then there's two
parts to it: cause apparent or alleged discrepancies or
shortfalls relating to subpostmaster's branch accounts
or transactions?
A. Yes.
Q. Not necessarily in their branch accounts, but relating
to --
A. Ah --
Q. -- branch accounts, bare branch accounts or
transactions?
A. Discrepancies relating to branch accounts. That could
be interpreted very broadly.
MR JUSTICE FRASER: You are both talking at the same time,
I'm afraid.
A. Sorry.
MR JUSTICE FRASER: Go on, Dr Worden.
A. I'm saying discrepancies relating to branch accounts
could be interpreted very broadly and it might be taken
to include all back end discrepancies because they
relate to accounts.
MR GREEN: Okay. (b):
"... undermine the reliability of Horizon accurately
to process and record transactions ..."
Just an example of (b), if a scanner doesn't
correctly scan in a pouch that undermines the
reliability of Horizon accurately to process and record
the transaction, doesn't it?
A. It does, and for better or worse I took my role to be
concerned with lasting effects on branch accounts rather
than transient ones.
Q. Well, you did two things, didn't you? You focused on
actual impact rather than potential?
A. I tried to derive an upper limit on actual impact.
Q. And you added in lasting?
A. I added in lasting.
Q. And in fact we see where you define the issue in your
report, you also focussed on the impact on claimants'
accounts?
A. Yes.
Q. That was your analysis?
A. That was what I felt would be useful to calculate.
Q. But you saw in Issue 1 that it said subpostmaster's
branch accounts?
A. Yes, I agree.
Q. So you took your own course in your own sense?
A. Well, yes, it is in two stages really. The first stage
is I felt extent in Issue 1 required numbers, and for
a second stage I needed some calibration of those
numbers, and those are two steps we can look at in
return, if you like.
Q. In all of this you were still clear that in the Horizon
system, transaction corrections, the manual process of
reconciliation and issuing TCs was not in?
A. The manual process, the cause was not in, yes, but the
effect was.
Q. I see. The helpline was not included in the system,
was it?
A. First line support was not included but the experts did
look at evidence about it.
Q. To have understood a bit about -- we have touched on
this before -- about the processes and the scope for
errors, you need to know a bit about the helpline if you
are going to look at whether people get through and how
often?
A. That is correct.
Q. Did you know whether people on the helpline were reading
out scripts when people rang up about things?
A. Which helpline are we referring to?
Q. NBSC helpline.
A. That is the business helpline.
Q. The first port of call, as it were. Did you know if
they were reading out scripts --
A. I'm sure they had prompts of various kinds.
Q. But you didn't know, was the answer?
A. I didn't know because I hadn't been asked to look in
detail at the MSC.
Q. You heard the helpline referred to as the "hell line"
at all?
A. I didn't see that phrase.
Q. You didn't. Can we look at {F/1257.1/1}, please. This
is a forum post in which Mr Tank had made a post. If we
look at page {F/1257.1/6}, please?
A. Can I ask what forum?
Q. It was a subpostmaster's forum.
A. Internal Post Office thing.
Q. No, a subpostmaster's forum that they used to use.
A. Right.
Q. Okay. {F/1257.1/6}, if you look at the second box down.
A. It is very dim on my screen.
Q. The second box down says:
"When I balanced on 17th September I had
an unexplained loss of £176.74. I paid it in rather
than go through the stress of the hell line. It has not
come to light."
A. Okay.
Q. It is only a yes or no answer to identify what
information you have had regard to --
A. Right, I haven't read this before.
Q. You have not seen this before?
A. No.
Q. Had you seen any other factual evidence about people
deciding whether to pursue things through the helpline
or not of this sort?
A. Well, what I have particularly looked at is the helpline
logs for one or two weeks. Now, that's not directly
about the decisions made by the postmaster.
Q. If we look at {D3/1/201}, which is paragraph 904
Just look at 904 at the bottom. The last two lines,
couple of lines, you say:
" ... although I have no knowledge of the business
processes for creating TCs used by the central
Post Office departments."
A. Yes.
Q. And that is correct isn't it, you don't?
MR JUSTICE FRASER: Is that the correct word?
A. I said yes.
MR JUSTICE FRASER: I'm sorry, it didn't come up.
MR GREEN: Now, we mentioned that you placed quite a lot of
emphasis on transaction corrections.
A. Yes.
Q. And essentially what you say is where there are
discrepancies in the branch accounts that have to be
corrected, the transaction corrections process is what
then makes that happen?
A. It is a very important part of the correction, yes.
Q. And it is very important for your analysis as well,
isn't it?
A. It goes to my analysis of robustness. It is an example
of user error correction, which I think is quite
an important one.
Q. And because you placed a lot of emphasis on it for
robustness, if the Horizon system -- it is a question
for the judge, but if the Horizon system was as it
appears in the definition at the beginning and didn't
include the transaction correction process itself, you
would have to answer the question, well, for every time
a TC is issued, that may be because a discrepancy has
arisen in branch accounts that needs to be corrected and
that may be due to the Horizon system?
A. Sorry, I haven't quite understood the question.
Q. Your answer to the robustness of Horizon is heavily
dependent upon TCs and, you say, countermeasures which
effectively reflect that process, and you effectively
say that those are important because that's how
discrepancies are corrected?
A. Yes. I'm trying to analyse robustness and the
correctness of errors in asking how effective are the
countermeasures in countering the effects of errors
often when I don't know the causes. It might be
something out in the client's system, or whatever.
Q. And it is right, isn't it, that the answer to the
robustness question on your analysis would be very
different depending on whether you included the
corrective impact of TCs or not?
A. Absolutely. Yes.
Q. Fundamentally different, in fact?
A. Yes, very different.
Q. It is right, isn't it, that you and Mr Coyne took
slightly different approaches to the Horizon Issue 1 and
what it meant, because we see, if we look at joint
report 2 at {D1/2/28} at 1.9, it says:
"The experts have differing views on 'branch
impact'. Mr Coyne refers to any discrepancy that caused
a loss (or gain) within branch accounts that needed
corrective action as an 'impact to branch accounts'.
Dr Worden only considers an effect or impact on branch
accounts where a discrepancy loss (or gain) was not
rectified by a correction such as a Transaction
Correction."
A. Yes.
Q. There was a fundamental difference about that and then
there was also a difference about transient and lasting
inaccuracies in branch accounts, wasn't there?
A. Well, I think that is the same difference, really. "Not
rectified" means lasting, whereas "rectified" means
transient.
Q. Well, we will come back to that point later, Dr Worden,
but that rather depends how long the TC process takes,
doesn't it?
A. It does, absolutely.
Q. So if the TC process takes a long time, then you might
say something was lasting, even notwithstanding years
later it might be corrected?
A. That is not how I understood lasting --
Q. How did you understand lasting?
A. There were delays in the TC process which might be due
to client organisations or might be due to all sorts of
things, and they could be at the outside, I believe,
several months.
My definition of lasting did not depend on TCs
coming in within a certain timeframe. If TC never came
in, that would be lasting, but if TC took several months
to come in that is not what I would call a lasting
effect. A lasting effect is permanent; it is at the end
of the day, you know, he has lost money forever.
Q. So in fact, 1.9 and 1.10 in the report {D1/2/29} are
really directed at the same thing.
At 1.10 what you say -- this is your comment:
"... transient inaccuracies in branch accounts which
needed some form of correction, have arisen so
frequently and from so many causes that to list them is
not useful; and that evidence of each correction being
carried out is unlikely to persist to this day."
That is your position?
A. Yes, that's my view.
Q. Now --
MR JUSTICE FRASER: Just so I'm clear, the example you just
gave to counsel of a transaction that took several
months --
A. A TC.
MR JUSTICE FRASER: I beg your pardon, a TC, a transaction
correction, would that be a transient inaccuracy?
A. That would be a transient inaccuracy, yes.
MR JUSTICE FRASER: That's very useful.
A. The branch accounts would look wrong for that period.
MR GREEN: So now that was one big tectonic difference of
approach between you and Mr Coyne.
A. Yes.
Q. The other one was the identification of particular bugs
and their impacts?
A. We ended up in different places on what we thought were
actual bugs, yes.
Q. I understand. It is right, isn't it, that in seeking to
reach some compromise given the differences of approach,
you managed to agree what we find at 1.15?
A. Yes.
Q. And that was the context in which that sort of
compromise agreement was reached?
A. Yes.
Q. Now, can we look please at page 9 of your first
statement at {D3/1/9}. You describe section 28, at
paragraph 29 of that page. Yes, you describe what
section 8 in your report is about.
A. Yes.
Q. And you describe section 8 as addressing Horizon 1, yes?
A. Yes.
Q. "... the extent to which bugs in Horizon may have
affected the Claimants' branch accounts"?
A. Yes, I've stuck in the word "claimants" there.
Q. You stuck the word claimants in because that's what you
in fact do in section 8?
A. It is mainly what I do in section 8, but I could come to
that.
Q. Now, can we look at how the reasoning in section 8 works
so we understand it. You'll appreciate, Dr Worden, that
there's a difference between the parties on what the
right approach might be?
A. Absolutely.
Q. But I am not going to proceed from here on to analyse
your approach on the basis of how you did it. I'm not
suggesting I agree with your approach, but I'm trying to
understand how you have done it in case you turn out to
be right. Do you understand?
A. I might do.
Q. And it is really to focus on the methodology you have
adopted.
A. Yes, right.
Q. You have made a number of assumptions in section 8 of
your first report, haven't you?
A. Yes, I have, to arrive at a number I made some
assumptions.
Q. It is meant to be an easy question. You are meant to go
yes and then I ask you the next bit.
A. Yes.
Q. Let's look at a couple of them, please. {D3/1/176},
paragraph 769.
A. 769?
Q. Yes.
A. "... would affect all branches ..."
Yes.
Q. Yes. So one of the assumptions you have used for your
model is that all branches, whether claimants or
non-claimants, are effectively --
A. Yes.
Q. There is another assumption if we look at {D3/1/148} at
paragraphs 621 to 622, and I'm going to suggest to you
it is basically the same point underlying it because on
the premise that it affects, there is an even
distribution of bugs, you have then considered a scaling
factor, haven't you?
A. Yes.
Q. And you specifically say at 622:
"It seems implausible to me that there is some
special factor about Claimants' branches, which makes
them more prone to --"
A. Much more prone.
Q. Quite right:
" ... much more prone to bugs which one would expect
to strike any branch at random."
Then you give quite careful consideration to this,
don't you, because it is important, in your report?
A. I have thought about it. It is in the appendix,
I think.
Q. Absolutely, so let's trace it through:
"Nevertheless, I have considered the possibility
carefully in Appendix F."
You say {D3/1/149}:
"I have shown that that there is no significant
difference between Claimants' branches and other
branches, in proneness to bugs in Horizon."
Pausing there. Bearing in mind what I asked you
about this morning, when you say "I have shown there is
no significant difference between Claimants' branches
and other branches", would it be fair to say that when
you say "shown" you mean thought about it, can't think
of a reason and therefore concluded that was the case?
A. I think that's fair, yes.
Q. That's what we see in appendix F.
If we look at {D3/2/207}, paragraph 430. Let's look
at the start of 429, if we may, on page 207. In your
appendix F, 429, you say:
"I have tried to think of possible differences of
this sort, and I have only been able to find one
candidate difference."
So you wracked your brains and the only thing you
could think of is then in paragraph 430?
A. That is right.
Q. And what you come up with, you say:
"It might be said that Claimants tend to make more
errors than non-Claimants ..."
Yes?
A. Yes.
Q. " ... and that these human errors particularly trigger
bugs in Horizon. Or it might be said that certain bugs
in Horizon are successfully handled by non-Claimants,
but tend to cause Claimants to make errors, which cause
losses."
Then you discount both of those?
A. Yes, I say -- the argument that follows is that they are
both second order effects.
Q. I understand.
Now, you heard Mrs Van Den Bogard give evidence, or
you may have heard her, where she accepted that
Post Office may suffer from UEB or user error bias
sometimes?
A. I didn't hear that evidence, but --
Q. Do you think that might be a fair suggestion against you
there?
A. I don't see how that's relevant to this consideration.
Q. You don't think -- you have approached this on the basis
that the claimants are probably making more mistakes?
A. I put that up as a possible hypothesis --
Q. As the only thing you could think of?
A. As the only thing I could think of.
Q. I understand. You then, if we go back very kindly to
your first report at {D3/1/149}, you have proceeded on
the basis of paragraph 623 that claimants' branches are
smaller and have fewer transactions?
A. Yes.
Q. And therefore they are less likely to be hit by
a Horizon bug in a given month?
A. That's the assumption.
Q. Have I fairly summarised those key planks of your
reasoning?
A. I think that's all fair.
Q. Your calculations then follow from those assumptions?
A. Yes.
Q. You reach conclusions which are likely if not calculated
to suggest that the claimants' claims are so implausible
as to be wholly unrealistic, don't you?
A. They are not calculated to suggest -- I did the numbers
as what comes out.
Q. What you conclude is that the claimants' claims are so
implausible as to be wholly unrealistic --
A. I conclude that the maximum amount of shortfall in
claimants' branches which could arise from bugs in
Horizon is approximately 0.5% of their claimed
shortfalls. That is one way of stating the number
I come up with.
Q. So through the prism of looking at the claimants
specifically rather than subpostmasters in general, you
have effectively reached an opinion that it is absurdly
unlikely that their claims can be right; is that fair?
A. Yes. Could I add a little bit? And that is I have
taken two approaches to Horizon Issue 1, one of which is
that I think numbers are important to give meaning to
phrases like "likely" and "risk" and so forth, and you
have to define the scope of those numbers. Because if
you say it is likely in the whole lifetime --
Q. I understand what you have done. We will come to the
detail.
A. But I'm coming to the second stage. The second stage,
I have said I think it is useful for the court to
calibrate those numbers with respect to claimants, but
I may be wrong in that. And if I'm wrong in that, then
the court can convert my numbers from a claimant basis
to an old postmaster basis. I have given the conversion
factors in my report, so if I'm wrong about emphasis on
claimants the court can convert the numbers to something
else.
Q. But Dr Worden, I'm going to follow through what you have
done.
A. Yes.
Q. And ask you at each stage whether you think it is right.
A. Yes.
Q. Or wrong.
A. Fine.
Q. Now, I would like, if you would just bear with me while
we go through a relatively simply statistical exercise,
if that is all right?
A. Yes, fine.
Q. And I would like you first to consider how one would
assess the chances of meeting somebody called Penny
Black in the UK. If you bear with me.
A. Yes.
Q. There are 60 million people in the UK of whom, as far as
we can tell, roughly 24,000 have the surname Black?
A. Yes.
Q. And of those, assume half of them are men and half of
them are women, so that is 12,000. Let's assume 1% of
all women with the surname Black are called Penny. So
1%, 120 people, are called Penny Black in the UK, let's
assume.
That is a consequence of that --
A. Sorry, we started with 60 million? Let me just track
the numbers.
Q. 24,000 are called Black.
A. So that is a factor of 3,000. And then the next factor?
Q. Half of them are men and half are women.
A. So that is a factor of 6,000.
Q. And then 1% of them.
A. So that is 600,000.
Q. In fact, on our calculation 500,000 --
A. Assuming that surnames and Christian names are not
correlated.
Q. Yes, exactly. Assuming they are independent variables.
We get 120 people out of 60 million, which is 1 in every
500,000. I'm not going to criticise your mental maths
in the witness box, just to get in the ballpark
together.
So that is 1 in every 500,000 people. Now --
MR JUSTICE FRASER: 1 in every 500,000 is a lady called
Penny Black.
MR GREEN: A lady called Penny Black.
MR JUSTICE FRASER: It is a different way of saying divide
60 million by 120.
MR GREEN: Exactly. And if you are at a dinner with 50
other people, 50 other females, let's say 50 other
people to start with.
A. So 25 females, let's say.
Q. Let's say 50 other people, right?
A. Yes.
Q. What are the chances of one person there being called
Penny Black? It is 50 over --
A. You mean exactly one or more than one?
Q. Let's say at least one.
A. At least one. I think it is the basic probability of 1
person being Penny Black times 50. I believe it is
that.
Q. Yes, exactly. So that is 50 over 500,000, isn't it?
A. Yes, 1 in 10,000.
Q. 1 in 10,000. What are the chance of two people being
called Penny Black?
A. That's approximately the square unless there is
a correlation of relatively -- if you are in a room full
of relatives, it is very different. But without the
correlation, I believe, let me think about this
carefully, I believe it is the square of the
probability.
Q. Yes. I mean, in fact there is a very fine point that
you have taken one person out of the UK and one person
out of -- but roughly it is the square?
A. Yes.
Q. Yes. So the chances of two people being called Penny
Black is 1 in 100 million, isn't it?
A. Two people at this dinner.
Q. It is one in 10,000 times 10,000. 10,000 times
10,000 --
A. I believe we are all right so far.
Q. It is not meant to be controversial. Just taking you
through it so we are agreed on methodology. And the
chances of three people being called Penny Black,
I promise you the last example, is 1 in 10,000x1 in
10,000x1 in 10,000, which is around 1 in a trillion,
yes?
A. I'm just trying to think about correlations.
Q. As independent variables. Don't get complicated. Just
as independent variables, that is the answer?
A. As long as we are in that assumption.
Q. Yes, we are, because it is terribly important, isn't it,
yes?
A. Well, obviously I don't know where this analogy is
going --
MR JUSTICE FRASER: I don't even think you need to worry
about that.
A. Good.
MR JUSTICE FRASER: But so far you are agreeing with
Mr Green's basic approach to --
A. Yes, it seems the probability theory is right,
basically. Yes.
MR GREEN: You may not know this, but in 2015 there was the
175th anniversary of the Penny Black stamp and
Royal Mail had an event to celebrate it at which they
invited people called Penny Black to attend.
A. Yes.
Q. Let's say 50 of those people who they invited to join
the dinner because they were called Penny Black came.
A. Yes.
Q. If you were at that dinner, what are the chances of the
other 50 people being called Penny Black?
A. It is obviously high, but we have violated the
assumption of no correlation.
Q. Yes, it is a different example?
A. It is absolutely, very different.
Q. So 50 people who are in fact called Penny Black have
been invited to join for dinner?
A. Yes. I mean --
Q. So the chances are likely to be at or about 100%, aren't
they?
A. Yes. I should say generally that probability theory is
what one uses in the absence of specific knowledge like
you have just put to me, and that specific knowledge
changes the whole ball game.
Q. The specific knowledge changes the whole ball game,
doesn't it, Dr Worden?
A. It does, absolutely.
Q. And you would accept that if we say, for the sake of
argument, the chances of the 50 people at the dinner,
you can imagine the chap coming out with the 50 credit
cards, "Penny Black anyone?" All the hands go up, yes?
MR JUSTICE FRASER: Because they are all being asked to pay
for the dinner they have been invited to.
MR GREEN: Your Lordship had the point, which made me
hesitant to give the example. But the point is that if
you look at the 100%, yes, on the one hand --
A. Yes.
Q. -- if we say approximately 100% chance of them -- there
might be somebody pretending to be called Penny Black
who is there interloping, but basically 100%, they've
all come because they've been invited because they are
called Penny Black, and one in a trillion for just three
of them, there isn't any bigger margin of probability
that we could illustrate by changing the assumptions.
A. Absolutely, and this shows that specific knowledge
overrides probability theory, when you have that
specific knowledge.
Q. It is essential. Do you know how GLOs work, Dr Worden?
A. GLO?
Q. Yes, that is a group litigation order?
A. Sorry, no, I'm -- I haven't looked into that, really.
Q. They are advertised for people to join. Let's look at
{C7/3/39}. That's the group litigation order. And at
page {C7/3/39}, this is the invitation/publication
notice to join the GLO.
The GLO is limited to people who satisfy those
criteria. Do you see that?
A. Yes.
Q. See what the third one is?
A. They claim to have suffered loss, yes.
Q. It is a pretty important specific bit of knowledge,
isn't, it Dr Worden?
A. I do not see how it fundamentally alters my calculation.
Perhaps you would like to spell that out to me.
Q. Well, you gave an example of people from the same family
being there, that would change everything. I gave you a
counter example of people being invited to join because
they are actually called Penny Black. This is
a situation in which the entire universe of claimants
has only joined the litigation at all because they
believe that they have suffered the very type of losses
we are concerned with.
A. They believe that, yes, but I'm asked as an expert to
examine that from a neutral point of view. I'm not
asked to take the claimants' case or the Post Office's
case. I'm asked to examine on my knowledge of Horizon
what the probability of certain events was.
Q. Okay, let's assume this: if the claimants were right
that they had been affected by such losses, yes, it
would be no surprise at all if they're all here,
would it?
A. I'm still at a loss to understand how that relates to my
analysis.
Q. Let me ask you this, Dr Worden: this is something
special about the claimants that you were searching
about in your quest, isn't it? They are the people,
they are the subsample of people who have sought to join
the group litigation?
A. But it would seem to me that I would then have to base
my analysis on an assumption that the claimants' claim
was true.
Q. Let's look at it the other way round. You would have
a factor to displace your assumption that people who
believed they have been adversely affected in this way
are equally likely to have been adversely affected as
people who believe they haven't been?
A. I don't see how making the assumption that the
claimants' case is true is an unbiased approach,
a balanced approach to assessing Horizon.
Q. Let's take that out of the picture.
A. No, this is my duty, that I have to make a balanced
assumption of the probability of those events not
assuming that either side's case is right when looking
at Horizon.
Q. But Dr Worden, you have reached that assessment by
starting, I would suggest to you, from a demonstrably
false premise that there is no special factor to
distinguish claimants from SPMs generally?
A. No, I have started from the assumption that a bug in
Horizon does not pick on claimants more than other
people. That is a different assumption.
Q. But in making that assumption you have ignored the
obviously relevant point that the definition of
a claimant is somebody who believes they have been so
affected?
A. I have not ignored that role in the point. I'm finding
great difficulty in following how this analogy relates
to the assumption I have made about how bugs in Horizon
act. That is the assumption. Not how claimants act.
Q. I will only ask you one more question about this before
I move to a different aspect of your report. Why would
someone who felt they had not been affected join the
group action?
A. I think somebody who felt they were not affected would
not join the group action. I do not see that as
relevant to the technical assumptions I made in my
section 8.
Q. Let's look at your scaling factor and your maths because
that's not a matter of opinion, is it? The advantage of
maths is that it is right or wrong?
A. That's very much the advantage from the court's point of
view and the expert's point of view.
Q. Let's look at the paragraph of your report Mr Coyne was
taken to at {D3/6/30}.
Just go back to page {D3/6/29} to give you context.
This is in the section 5.1 "Size of Claimants'
branches".
A. Yes.
Q. So this is the second of the assumptions I put to you
earlier, yes? You have assumed even distribution which
we have dealt with.
A. Even distribution of what?
Q. Of bugs across all subpostmasters, we have dealt with
that?
A. Bugs per transaction in Horizon.
Q. Yes. And you are now turning to your scaling factor, as
you have called it?
A. Yes.
Q. "... which expressed the average size of a Claimants'
branch (defined in terms of customer transactions per
day) divided by the average size of any branch across
the Post Office network defined in the same way."
A. Yes.
Q. The relevance of this is that because the claimants'
branches are smaller, smaller by number of transactions,
they are less likely to be hit by bugs. That is the
relevance.
A. That is right, that's the relevance. By about this
factor too.
Q. You say 0.37 in your first report and change it to 0.45
in your second report?
A. That is right.
Q. If we look at paragraphs 113 to 114 over the page, these
are the paragraphs that Mr Coyne was taken to {D3/6/30}.
You say four lines down "In my first report" because
this is your second report?
A. Where are we? Yes, got it.
Q. Paragraph 113, four lines down:
"In my first report, I made a calculation for each
year of the size of Claimant branches versus all
branches, and then took an average across the three
years."
A. Yes.
Q. Is that based on the data of 2001, 2007 and 2018?
A. Pretty much, I think, yes.
Q. Let's go back and see what you in fact did. {D3/1/149},
624.1.
A. Yes.
Q. You are referring at 624 to the figures in the
spreadsheet Mrs Van Den Bogard had?
A. Yes.
Q. You say:
"If this spreadsheet is accepted, it implies the
following about Claimants' branches:
"624.1. From summing rows of the spreadsheet, the
561 Claimants' branches carried out 558,000 customer
transactions per week ..."
Yes?
Then what you do is you subsequently compare that,
yes?
A. Yes.
Q. To get the number of customer transactions per day?
A. Yes.
Q. Now, there you are referring to the spreadsheet in
Mrs Van Den Bogard's evidence, yes?
A. Yes.
Q. And you actually say at 624.1 that you have looked at
the customer transactions per week in 2007?
A. Yes.
Q. Not an average over three years?
A. I have made an assumption of, if you like, linear change
to get at that figure. I mean, this figure is very
approximate and only needs to be very approximate, and
it was very approximate in my first report. I tried to
improve it a bit in my second report, but 2007, if you
have a uniform rate of change, 2007 is the average of
those three years.
Q. What you have actually done is take the figures for
2007?
A. Yes, as representative of the three years.
Q. But what you didn't do, you didn't take an average of
the three years by actually adding up the three years
and --
A. No, there was an extra assumption in there that the rate
of change between those two periods was the same, as the
average of the three years was the same as the 2007
figures. And that was, if you like, a hidden assumption
that I probably should have been clear about.
But as I say, the level of precision -- one is
always concerned with the level of precision one
needs --
Q. We have got that point.
A. -- in an answer, and I felt that the level of precision
I needed in this figure, even if we changed it to 1 it
wouldn't make much difference, really.
Q. I understand, I have got your point. But on precision
about what you said, it's not right what you say in your
second report. What you actually did was take the
figures from 2007?
A. Well, 2007 I took as representative --
Q. I understand that. It is not what you said you did.
A. No, I should have explained more carefully.
Q. And the figures you had for the network were 2017 and
then you derived some for 2003. That is right, isn't
it?
A. Where are we?
Q. We are in 625 and 626. At 625 is the figure for 2017.
A. Yes. Again, there is an interpolation here that
I didn't have a middle point so I took the average from
the first and the last point.
Q. Okay. So that's what you in fact did in your first
report, and so the change of method that you referred to
in your second report as a subtle statistical change of
method, from one type of averaging to another, it
wasn't, shall we say, fully explained in your second
report, the change of approach, was it?
A. Well, the change in method was explained and I can
re-explain that if you like. It is a matter of --
Q. We have only got so much time, Dr Worden, so let's look
at the analysis.
You refer at paragraph 623 to the spreadsheet in
Mrs Van Den Bogard's witness statement which she refers
to at paragraph 179 of her witness statement. That
spreadsheet is at {F/1837/1}. Were you in court for
Mrs Van Den Bogard's evidence where the original version
had wrongly allocated the figures across branches and it
had been changed but she didn't know it had been
changed?
A. I wasn't in court for that, no.
Q. Did you realise that had happened?
A. I was told about some changes and I believed that within
the margins of precision that I needed, those changes
did not make any significant difference.
Q. Okay.
Now, the spreadsheet -- can I just pause. When you
were told there had been a mistake in the first
spreadsheet, in data that Post Office had extracted, and
they had to change it, did that worry you at all?
A. I thought about it, but as I say, the precision I needed
in this number is a rather low precision, and to that
precision I felt it didn't make a significant
difference.
Q. Did you do the calculations on the spreadsheet or did
Mr Emery?
A. I did it.
Q. Okay.
A. I produced another version of it which I did the sums --
Q. Can we please go down to row 74. Do you see that
there's a gap in column E?
A. Little Waltham, yes.
Q. If we go down to rows 84 and 85.
A. Yes.
Q. Gaps there?
A. I was aware of these.
Q. You were aware of the gaps?
A. I was aware the data was patchy, yes.
Q. And in fact, Mrs Van Den Bogard explained in her witness
statement that the gaps had been filled, didn't she?
A. Mm.
Q. Yes?
A. Yes.
Q. After further investigation the gaps had been filled?
A. Yes.
Q. And you had intended to take the number of the filled or
unfilled column.
A. I can't remember the precise details of my calculation
because, as I say, I was looking for a precision which
was rather low. And therefore in terms of priorities
and what I put my effort to, I knew that other parts of
my calculation had greater imprecision than that, and so
I didn't spend a lot of effort on trying to refine this
figure.
Q. Okay. So will you take it from me that the column E
figure is 558,260, which is the figure you have used?
A. Column E?
Q. Do you see E is the one with the missing ones in?
A. Yes.
Q. And F has got the gap filled?
A. Yes.
Q. Do you see that?
A. Yes.
Q. And the sum of column E is 558,260 and with gaps filled
the figure is 575,803?
A. Sorry, can I have those two figures again? Are they on
this spreadsheet? They are not. You have done the
sums.
Q. They're not but we could sum them at the bottom. I've
got one here.
A. What are the two figures?
Q. Column E with the gaps in it is 558,260, which is what
you have used.
A. Yes.
Q. So you have used the columns with the gaps in it.
A. Yes.
Q. And Mrs Van Den Bogard explained another column was
there which had the gaps filled, and you have not used
that and that comes to 575,803.
A. So the difference is 4%.
Q. I'm just saying you used the wrong column.
A. Well, the kind of calculation one makes is I'm doing
an average of three numbers, one of them is going to
change by 4%, so the impact on the average is 1%. So
does that matter?
Q. No, Dr Worden, sorry, you used the 2007 figure as
a spurious average proxy for three years' averages on
the basis of an assumption that you didn't tell anyone
until you were being cross-examined. That's what
happened.
A. No, it is not what happened.
Q. Right. You explain what happened here.
A. What happens --
Q. Because that sounds like --
A. -- is when you are making an engineering estimate you
have in mind two things: one is the precision of your
result that you need to assist the court, or whatever it
may be, and the other is the precision that you can
achieve in your calculation.
Now, I was aware when I made this calculation that
the precision I could achieve, firstly the precision
I needed was rather small. It might have been to
a factor of 3, or something like that. Secondly, the
precision I could achieve was rather small for reasons
completely unconnected with this; other parts of the
calculation were less precise.
Therefore, you say: when I'm looking at this
calculation, how precise do I need to make it? And if
there's going to be a 4% difference in one of the
numbers that is added up to 3, is that a significant
difference? And you say no, and so it is, you know, not
worth bothering about that. And that is what people do
in engineering calculations and they do not worry about
spurious precision.
Q. Okay. Would you regard it as spurious precision to have
found out for yourself whether the gaps made a material
difference before you decided not to care about them?
A. You can do it by eyeballing.
Q. So you did eyeball the difference? You are saying you
can do it by eyeballing. Are you telling his Lordship
you did eyeball the differences and form a view that it
wasn't worth bothering about?
A. I formed a view that the changes --
Q. No, did you eyeball it and form a view, or is that
an answer of convenience?
A. I believe I did.
MR JUSTICE FRASER: Right, you mustn't overtalk. It is not
you, Dr Worden.
Mr Green, you really mustn't because I would like to
hear what the witness says and then you can follow
it up.
A. But as I say, the whole basis of this calculation is how
much precision do I need for different parts of it. And
I don't spend time struggling to produce high precision
in parts of the calculation that will not produce high
precision in the result.
Q. So are you saying you knowingly chose the column with
gaps in it because you eyeballed and decided not to
bother with the correct column?
A. I can't remember the precise sequence, but I decided as
a judgment about precision required that it was not
necessary to worry about these details.
Q. Why did you divide by 561? Because that's the number of
claimants? Is that the reason?
A. That's basically the reason.
Q. Did you notice that there were only 496 listed on the
spreadsheet or was that not a level of detail you needed
to bother with?
A. Again, I can't remember how exactly I treated that,
but --
Q. You just didn't notice, did you?
A. No, I did notice. I was aware of that difference.
Q. So you knowingly presented to the court a figure which
represented 496 --
A. I knowingly presented to the court a set of estimates
which I believe are sound and have adequate precision in
each part of the estimates. And the ultimate precision
that's required in the assessment is not very high.
Q. Can we go to row 1, please. Do you see what the
heading is?
A. Sorry, what are we looking for?
Q. The heading of the entire table.
A. Yes. The header row or the heading?
Q. It says "Volume of Customer Sessions", doesn't it?
A. Yes.
Q. Are you are aware that that's what it was both from the
heading and from the fact that Mrs Van Den Bogard said
it was customer sessions?
A. Yes.
Q. And you are also aware that a session is not confined to
one transaction; it could be three?
A. The average I believe is 1.7, but again, in the sources
of data that I had, there were ambiguities.
Q. Hold on a second, you have just told the court that the
average session contains 1.7 transactions. You just
said that.
A. There are various different terms used in different
places and there are issues of comparability of data,
I agree with that entirely.
Q. Hold on, you have scaled down the claimant branches on
the basis of one variable and one variable alone, namely
the number of transactions?
A. I have --
Q. To do that you have used the number of sessions for
claimants and the number of transactions for all
branches, and you know that they are different and you
know that on your evidence the difference is a factor
of 1.7?
A. No. When I was consulting various different sources
about volumes of transactions, there are various
different sources as well as Ms Van Den Bogard's
evidence, there are early documents and so on and so
forth, and I was aware that there are uncertainties of
definitions about these things and I was doing the best
I could and there are uncertainties in the result.
There is no question of it.
Q. Dr Worden, this is a serious factor because you well
understood, as an expert who has looked into transaction
integrity in session data, that there could be more than
one transaction in a session. You knew that, didn't
you?
Take it in stages. Did you or did you not know that
there could be more than one --
A. Well, the definition of transaction is rather fluid.
For instance, when you have a customer session, that is
genuinely packaged in -- what the 1.7 refers to, it is
unclear in different spaces, but what the 1.7 refers to
is generally how many things a customer does in one
session. But that still doesn't alter the fact that
a session in some terminology is one transaction because
it is a success unit in the database.
So there are these ambiguities and I agree there are
big uncertainties in my calculations.
Q. So in fact, even if we take your figure of 1.7, we
might -- none of this is explained in your report, none
of the uncertainties?
A. Well, this is not the big uncertainty in the analysis.
The biggest uncertainty in the analysis comes from the
number of bugs and the scale of each bug, and that is by
far a bigger uncertainty. And so I felt in terms of
priorities and length of explanation and so on and so
forth, it was not -- you know, the engineering approach
is not to drill down on things that you do not think are
the main uncertainty in your result.
Q. But you might have had to double, or nearly double your
scaling factor on your own evidence?
A. I put in a conservative (inaudible) factor from 0.45
to 0.5, which I will acknowledge I could have done.
Q. But you have left out a massive point the court would
need to know and certainly came up in cross-examination.
A. There are all sorts of uncertainties in this analysis
and I drew attention to what I felt were the main
sources of uncertainty.
Q. So just to clarify, wrap up, you knowingly took the 2007
column with gaps?
A. Yes, that's, as we discussed, an effect of 1 or 2%..
Q. And you eyeballed that to guess what the effect was?
A. Yes.
Q. You say that you noticed that there were only 496 lines
in the spreadsheet but proceeded nonetheless?
A. I believe I corrected that. In fact I believe what
I did was a rough correction which said just put in that
factor 491 over 561 or whatever it is basically.
Q. We don't see it anywhere in your calculation --
A. No, absolutely. I was making decisions about what's
worth drilling down and what's worth exploring and
I felt that the imprecision in the calculation, the
subjects I had to talk about at length, were other
imprecisions and that this claimants' scaling factor, it
doesn't make much difference if the claimant branch size
on average is the same as "other", because a factor 2 is
less than the precision I need in my final result in
order to be useful to the court.
So I was focusing my attention on what I felt to be
the biggest areas of uncertainty in my analysis.
Q. And there's no hint of those uncertainties mentioned in
your report?
A. No, the biggest areas of uncertainty, there is a lot of
mention of them.
Q. And there are a whole load of other points. Let's just
see, for example, you refer to, taking it very quickly,
you may take it from me you refer to 48 million
transactions. Again, it is not a big difference but the
actual figure is not in Angela Van Den Bogard's second
witness statement, it is in the first one. It is
47 million, not 48 million. You say Mr Coyne also
referred to that and he correctly recites it as 47.
This --
A. 47/48, in the context of my calculation I certainly took
a difference of 47 to 48 as being insignificant.
Q. So all of those errors, as we say they are, you say they
are fine? They are all against the claimants, aren't
they? If they are errors, which is a matter for his
Lordship, all of them have an effect against the
claimants, don't they?
A. What I have tried to do is present a numerical
calculation --
Q. Can you just answer the question?
A. Yes.
Q. They are. And if you are an independent expert and the
chances of you making a mistake one way or the other,
say against the claimants, is 50/50, yes?
A. Mm.
Q. The chances of you making the first mistake against the
claimants is a half?
A. Sorry? Which mistakes against the claimants are we
talking about?
Q. The four that I have just identified.
A. Right and what I would say is those can have an effect
and that effect, in my opinion, is balanced by the
claimant favouring factors, that you can see in table
8.4, which have an effect of a factor of approximately
30.
Q. Dr Worden, we are just talking about the primary maths.
A. Absolutely, the primary maths, the maths folds in errors
and changes made from various sources. There are
uncertainties and uncertainties and there are deliberate
claimant favouring factors which have injected, which
amount to a factor of 30.
Q. We will come to those. Just on this analysis, making
four 50/50 mistakes against the claimants, the chance of
6.25% of that being random; 1/16.
A. You are talking two to the four, is that what we're
talking about?
Q. Half of 50 is 25. Half of 25 is 12.5.
A. I do not see the relevance of that figure.
Q. Okay. It is just that the claimants are just unlucky
that all your errors are against them?
A. In this respect yes, but there are large numbers of
errors which are in their favour in the calculation and
I think one should take a balanced view of the errors in
their favour versus the errors against them.
Q. Dr Worden, would you accept that there is a difference
between an error that is not apparent on the face of the
report and a prominently flagged bit of generosity to
the claimants when you adjust a figure that you have
derived in their favour slightly? You understand what
the difference is, don't you?
A. I have tried to present a balanced opinion on this
number.
Q. I'm not asking you that because you have given that
answer about six times. I'm asking you do you
understand the difference between a concealed error in
your report, ie one that we can't see on the face of the
report, and somewhere where you say: I'm now going to be
enormously generous to the claimants and increase the
number to be really generous and conservative. You know
the difference between those two things, don't you?
A. I would not characterise it that way, I would say I made
some engineering assessments about what were the largest
sources of error in my calculation and I focused my
attention on those, and one made the typical engineering
decision that you put your priorities and your attention
where the big errors are and that's what I did.
Q. Let's very briefly, in your second report, we will just
go back to that quickly {D3/6/30}. This is where we
were looking at it and you explain that you have looked
across three years. We have only seen two years of data
for all branches. You used only two years data for the
all branches figure, didn't you? Yes?
A. I think that's the way I did it, yes.
Q. Okay. Let's look at the smaller branches scaling factor
methodology now, how you used the scaling factor. Let's
look at paragraph 630 in your first report {D3/1/150}.
We are looking at paragraph 630.
A. This is in the first report?
Q. Yes, it is. On the basis of the session data
transaction data comparison you say:
"So the claimants' branches, being generally smaller
than the Post Office average, have fewer transactions
per month and so are less likely to be hit by Horizon
bug in a given month."
Then you give the scaling factor.
A. Yes.
Q. Now that analysis depends on three things, doesn't it?
First, it depends on there being one variable by which
that scaling is to be measured and that's transaction
numbers?
A. Yes.
Q. Yes?
A. Mm.
Q. And it also depends on that varying directly in
proportion to bugs --
A. There is a linear assumption in there, yes.
Q. So the assumptions behind it are one variable, varying
directly in proportion to bugs and therefore reaching
your view that the scaling factor is how to calculate
the likely incidents of bugs on those claimant branches.
That is a fair summary, isn't it?
A. I think so, yes.
Q. Now, what we do have is data about branches which were
affected by actual bugs that we do know about?
A. We do.
Q. Yes. And if we were to take the Dalmellington bug, the
88 branches affected by the Dalmellington bug were tiny
outreach branches, weren't they?
A. Well, an outreach branch is tiny but it is an outreach
from a branch which -- I haven't looked into the size of
branches which had outreach branches.
Q. Yes and so to do this, even if you assume it is one
variable, which we will come to and so forth, if you
wanted to get even a sense of whether this was right, it
would be sensible to calculate -- to look at the number
of transactions done by the branches that were suffering
the identified bugs because you don't know how they
compared to average branch sizes, do you?
A. No, I don't.
Q. And that's not a calculation you did?
A. No, this is what I would call the next level of detail
down that one could have gone to look at the branches
affected by the known bugs or the acknowledged bugs~--
Q. Would you -- sorry.
A. -- and one could have calculated the average size of
those and I could indeed have factored that in my
calculation, I --
Q. Would you accept that's basic statistics to identify the
characteristics of the control?
A. Identify the characteristics of the control? To
the extent you can, yes.
Q. And you didn't do it?
A. As I say, it is an extra level of sophistication in my
calculation which I didn't do. I acknowledge that.
Q. But you are a bit of a legend in advanced statistics in
court, aren't you?
A. I don't know what that means.
Q. Well, you have given evidence in very prominent cases,
expert evidence, of advance statistical techniques; you
have a PhD in theoretical particle physics; and you have
come to this court in an important case to give expert
evidence about statistics?
A. Yes.
Q. Right. And you say what about why it is acceptable for
you to leave out that step of the analysis?
A. Because one makes a calculation at certain levels of
approximation based on the precision you require in the
result, and the precision I required in the result was
probably to within a factor of 3 or a factor of 10 and
I made the judgment that to achieve that precision
I didn't need to go to that next level of detail that
you have described to me.
Q. So if the average branches, if you had regard to them
having, say, 1/10 of the average transactions of normal
branches, you might have had to have a scaling factor of
x10 instead of x0.37 or 0.45?
A. There is a contradiction there. If my central
assumption is that branches are affected by bugs in
proportion to the number of transactions that they do,
then it is highly unlikely that the branches affected by
the three known bugs should be small in terms of their
numbers of transactions.
Q. That's why it throws up whether your assumption that it
is a uni-variable analysis based on transaction numbers
is right, isn't it?
A. Sorry, I don't understand the question.
Q. You have pointed out that the small branches being hit
would have helped you to identify whether you were
likely to be right that the only relevant variable to
consider was transaction numbers?
A. Sorry, I'm not really following the question.
Q. You have just pointed out when I gave you the
Dalmellington example, you said if it is lots of tiny
branches being hit, you said, well, that would question
my assumption that it is proportionate to transactions.
A. But I think it is unlikely that the Dalmellington
branches were tiny for precisely the reason I have just
stated.
Q. And it is right that a bug that affects an SPM when they
roll over -- yes?
A. Yes.
Q. If you had a bug which just affected rolling over?
A. Yes.
Q. A branch, however large or small, they all have the same
number of --
A. Same number of rollovers. Absolutely. But the branch
that is bigger has some bigger figures in its accounts,
so the effects of one bug on one rollover is likely to
be larger. So that --
Q. In number terms but not in instance and probability
terms?
A. Sorry, I do not understand.
Q. I'm just saying if a bug affects rolling over, just the
process of rolling over, the chances are not affected by
the number of transactions that they are otherwise
doing?
A. No, but the financial impact, which is the main
calculation, is affected.
Q. Yes, but I'm just trying to point out to you that it is
not based on the number of transactions as the frequency
or likelihood of bugs happening in all cases, is it?
A. But my main calculation was financial impact and I knew
that for financial impact larger branches had larger
numbers of transactions and they also had bigger numbers
involved in roll over, so the financial impact, it was
equivalent.
Q. But you deal with that separately. Your scaling factor
is about probability, isn't it?
A. Which scaling factor?
Q. The scaling factor that we have been talking about this
afternoon.
A. You say it is about probability --
Q. The premise of the scaling factor is that they are
equally likely to be hit; an average branch is equally
likely to be hit by another average branch, but
a smaller branch is less likely to be hit because they
do fewer transactions. That is the premise of your
scaling factor?
A. Well, the scaling factor applies to financial impacts.
Q. No, can you just answer the question because you know
what I'm asking. The scaling factor is designed to
control for a lower chance of a branch suffering
an impact of a bug, the incidence not the value, the
incidence of a bug by reference to the number of
transactions that happen in the branch?
A. Yes, we can agree --
Q. That is correct?
A. I can agree with you that if a bug affects roll over,
then a branch that is small is just as likely to be
affected as a branch that's larger, yes. Absolutely.
Q. And if a bug affects remming in the pouch --
A. Yes.
Q. -- if the pouch contains £100,000 or £2,000 for a
smaller branch, it doesn't have any impact necessarily
or any relation to the number of transactions they are
carrying out, does it?
A. I think it does because -- again we are going to a very
fine level of resolution here, but generally one would
expect a larger branch to have more cash remming.
Q. But that might be -- you haven't got any information
about that?
A. We are going down to the next level of resolution in the
calculation here and, you know, one has to cut off
somewhere in the precision of the calculation and the
place one chooses to cut off depends on the precision
you require in the answer.
Q. Finally, in relation to mathematics, we saw with the
spreadsheet at {F/1837/1}, if we can get that up, about
the column without the gaps filled been taken. Yes? We
have been over this. I'm just suggesting to you that
that was actually a mistake, Dr Worden. You just got
the wrong one. You misread -- go up to row 1 please.
What actually happened is you looked at the column
headings, saw 2007 gaps and thought that was the one
with the gaps in it and you made a mistake and took
column E. That's what really happened, isn't it?
A. I'm trying to remember what these columns do mean. You
seem to get the same figure in those two columns, so for
the level of precision I need, you can take either
column.
Q. What that first page doesn't show is all the gaps that
are on the other pages. I'm just putting to you that
actually the evidence you gave earlier isn't right, you
just made a mistake?
A. Well, there are imprecisions in that calculation we have
agreed that, and I felt that those imprecisions were
insignificant in the context of my overall calculation.
Q. It wasn't a deliberate choice to take the wrong column,
was it?
A. I made no deliberate choice to take the wrong column,
no.
Q. It wasn't a deliberate choice to take the column without
the gaps filled, was it?
A. I repeat that I don't think the gaps make a difference
in the context of the accuracy that I'm looking for.
Q. I'm only going to ask it one last time because we are at
4.29. The question is, was it a deliberate choice for
you to take the column that still had gaps in it?
A. It was not a deliberate choice, no.
Q. Was it a mistake?
A. It was a judgment that that issue was not important
enough for me to worry about it.
MR GREEN: My Lord, is that a convenient moment?
MR JUSTICE FRASER: Yes, it is. I have just got a couple of
requests it is just for some documents or some
references. Mr de Garr Robinson, can I just have a hard
copy of Dr Worden's corrections sheet.
MR DE GARR ROBINSON: My Lord --
MR JUSTICE FRASER: I don't need it now.
MR DE GARR ROBINSON: You can have it now --
MR JUSTICE FRASER: Well, if you have a spare one.
MR DE GARR ROBINSON: I believe I have got --
MR JUSTICE FRASER: I was reading it on screen but it is
just helpful and the other thing and this is --
MR DE GARR ROBINSON: This is clear, if I can hand it. My
learned friend has one.
MR JUSTICE FRASER: That is a different thing I think.
MR DE GARR ROBINSON: This is the corrections.
MR JUSTICE FRASER: So that is the corrections given in
chief. I think Mr Green is proffering a copy of the
signed joint statement. Thank you very much.
Then the only other reference which is one for
Mr Green, Mr Green can you just give me, tomorrow is
fine or by email in the morning, just a reference for
the covering letter that came with the MSC document that
you were putting this afternoon at {F/1834.14/1}.
MR GREEN: My Lord, one of them is at {H/325/1} which is
where it is explained that the documents with the new
MSCs are self-evident and don't need to be explained and
there I think is a couple of others which we will find
for you.
MR JUSTICE FRASER: If you maybe just pop them in an email
and copy into Mr de Garr Robinson and in the morning is
fine.
Thank you all very much. Actually when I say in the
morning I don't mean in the morning, actually for you I
do mean in the morning.
MR DE GARR ROBINSON: Really?
MR JUSTICE FRASER: I know we are not sitting tomorrow, so
I'm not asking you to do anything tomorrow. The thing
I was asking Mr Green to do by in the morning, I go back
to my original position, please do send it to me in the
morning.
MR GREEN: I will send it. I'm most grateful.
MR JUSTICE FRASER: Dr Worden it is going to be a bit more
difficult for you because you have one day interval, but
you know you are not to talk to anyone in the case and
we are going to come back on Thursday.
A. Yes.
MR JUSTICE FRASER: Is there anything else for today?
MR DE GARR ROBINSON: My Lord, no.
MR JUSTICE FRASER: I hope tomorrow goes as well as can be
expected in all the circumstances.
MR DE GARR ROBINSON: My Lord, that is very kind. Thank
you.
(4.32 pm)
(The court adjourned until 10.30 am on
Thursday, 13th June 2019)
(10.30 am)
MR GREEN: My Lord, my learned friend is about to call
Dr Worden. Just before he does, your Lordship asked us
about which days were which on the transcript.
MR JUSTICE FRASER: Yes.
MR GREEN: I have shown my learned friend this because
accidentally we only got one copy, but it is Days 9, 10,
11 and 13 that are not part of this trial, and we have
set them out on --
MR JUSTICE FRASER: I'm not sure that's necessarily correct
in the sense that they are -- but that's fine. So which
day is today?
MR GREEN: Today is Day 18, my Lord.
MR JUSTICE FRASER: Day 18, all right. That's excellent.
Thank you very much. Are you going to give a copy of
that to Mr de Garr Robinson?
MR GREEN: Yes, I have shown him it already and we will give
him a copy.
MR JUSTICE FRASER: Thank you very much.
MR DE GARR ROBINSON: My Lord, I call Dr Robert Worden.
DR ROBERT PEEL WORDEN (affirmed)
Examination-in-chief by MR DE GARR ROBINSON
MR JUSTICE FRASER: Do have a seat, Dr Worden.
A. Thank you.
MR DE GARR ROBINSON: Dr Worden, I see that you have
a formidable box there which I believe contains hard
copies of your reports. You may or may not want to put
the box on the ground; that's a matter for you.
What I'm going to do is take you to your reports and
ask you a couple of questions about them. First of all,
your first report on the trial bundle is {D3/1/1}. Is
that your first report?
A. Let me just go to the very beginning. That is.
Q. If you could go to page {D3/1/260}.
A. Could you tell me what tab that is?
Q. It should be at the end of the first tab.
A. I see, it is the affirmation bit. I'm sorry, the
signature bit. Got it.
Q. Is that your signature?
A. That is my signature.
Q. Then if I could ask you -- this may not be in the
bundles -- to go to {D3/1.1/1}.
A. Sorry, how do I correlate this thing with D3? I don't
think I do.
Q. It may not be in the hard copy bundles. I'm afraid I
don't know what's in those bundles. But you will see
a list of corrections there, yes?
A. The list of corrections I think is at the front.
Q. Are these some corrections to your first report that you
identified that needed to be made?
A. Yes.
Q. And this was added to the trial bundle back in April?
A. Yes.
Q. Now let's go to your second report. That's {D3/6/1}.
A. Let me just find my way around.
Q. I suspect it will be in a different bundle but it may
not.
A. I think it is right at the back of this one actually.
Let's have a look.
Q. Gosh, that's a very big bundle, isn't it?
A. Sorry, my fault.
Q. I wonder, my Lord, in due course whether it might be
appropriate to split them up.
A. I think it is a different bundle.
Q. Yes. I see that there are other bundles in the box.
A. Sorry, I will find my way around this soon.
Supplemental report. Got it.
Q. For the trial bundle it is {D3/6/1}. Is this your
second report?
A. This is my second report.
Q. Dr Worden, I have not found your signature page or,
indeed, an expert declaration. Do you give that expert
declaration in relation to your second report?
A. I do.
Q. And there are some appendices and financial calculations
which add to or correct some of the points made, and
calculations contained in your first report, aren't
there?
A. There's a series of corrections. I mean, the
calculations made in the first report were first
corrected in the second report and then a further
correction was made with the second expert joint
statement.
Q. I see. So subject to the corrections we have just
discussed, do you believe that these two reports, their
appendices and their attached financial impact
calculations to be true?
A. I do. They are my opinion.
Q. Joint statements. A point has arisen on the joint
statements that you and Mr Coyne have agreed. I'm told
there is an amendment that has been agreed to
paragraph 11.1 of the fourth joint statement. Is that
right?
A. That is correct.
Q. Do you have a copy of the amendment that has been
agreed?
A. Well, I had a few minutes ago when I signed it. I'm not
sure where it is now. It is probably in one of these,
isn't it? The problem is it is not amended in this one.
Q. Could you tell me what you did with the piece of paper
you signed?
A. I gave it to my assistant who gave it, I think, to
somebody else.
Q. Do we have a copy of the signed version of the amended
JS4?
MR GREEN: We have got copies if that helps.
MR DE GARR ROBINSON: Thank you. If that could be handed to
Dr Worden.
MR JUSTICE FRASER: I would like a copy, please. Is there
only one?
MR GREEN: My Lord, I think this is the one that's signed.
MR JUSTICE FRASER: Give it to the witness first.
MR GREEN: We will produce copies.
A. I think two were signed. I signed two.
MR DE GARR ROBINSON: That's the fourth statement which --
for the trial bundle, the current version of which is in
{D1/5/1}. I believe the amendment may be at page
{D1/5/10}. Is that right?
A. The amendment is 11.1, which is at page 10.
Q. Could you read out what the original version says and
what the changed version now says.
A. Do you want the full statement each time?
Q. It's just a sentence, isn't it, in 11.1?
A. Yes, it is a longish sentence. Start with the original:
"Evidence from several Peaks indicates that whenever
Fujitsu needed to make any change to data which impacted
branch accounts, they were concerned to seek permission
from PO to do so, and to ensure that PO took
responsibility for the resulting change."
{D1/5/10}
Q. Right. And what is the change that you have agreed?
A. The change that Mr Coyne and I agreed out there the
other day was that the word "whenever" is replaced by
"usually when".
Q. Just to explain, could you explain what you mean -- I'm
not asking you to delve into Mr Coyne's head -- by
"usually when"?
A. Yes. I mean, I agreed to the replacement of "whenever"
because "whenever" implies that the opposite never
happens, and I felt I didn't know strongly enough to say
that categorical statement. So "usually when" to me
means the great majority of cases, but I can't say with
my hand on my heart all cases.
MR DE GARR ROBINSON: Very good.
My Lord, I have no further questions.
MR JUSTICE FRASER: Just before -- and it might not be a big
point but it might be, the note that I made of when this
point was put to Mr Coyne was that the change was to
"usually" rather than "usually when". But your
understanding --
A. I noted that problem. I didn't say anything about it at
the time.
MR JUSTICE FRASER: This is last week?
A. Yes. I noted people said "usually" instead of
"whenever". I thought that is a problem, but I didn't
actually say anything.
MR JUSTICE FRASER: Right. But the change that you consider
should be made is that the single word "whenever" is
changed to "usually when"?
A. That I believe is the words we agreed.
MR DE GARR ROBINSON: That has been agreed, my Lord.
MR JUSTICE FRASER: The point that was put to Mr Coyne that
was agreed though was that the change was from
"whenever" to the word "usually".
MR DE GARR ROBINSON: Well, your Lordship will remember that
it was made clear that at that stage the actual wording
hadn't been --
MR JUSTICE FRASER: That's why I said it might not make
a difference, but my note of what was put was that the
change was to "usually". But there's now a signed
statement that has "usually when" on it; is that right?
MR DE GARR ROBINSON: Yes.
MR JUSTICE FRASER: Is that what the witness has been given?
MR DE GARR ROBINSON: Yes, my Lord.
MR JUSTICE FRASER: Can I have either a copy of that or that
at some point.
MR DE GARR ROBINSON: Yes.
MR JUSTICE FRASER: Thank you very much,
Mr de Garr Robinson.
Mr Green.
Cross-examination by MR GREEN
MR GREEN: Dr Worden, can we look at your CV, please. It
will come up on the screen. So when I call
a document --
A. Fine.
Q. It is {D3/4/1}.
A. Is it in this bundle, by the way? It should be,
I guess. Anyway, there we are.
MR JUSTICE FRASER: Just for the cross-examination generally
if you would like to see the hard copy in front of you
of whatever is on the screen please just say so.
A. Thank you, my Lord.
MR JUSTICE FRASER: You will be given the reference.
Right, Mr Green.
A. Here we are.
Q. So we have got your strong practical experience
explained there?
A. Yes.
Q. You have a PhD in theoretical particle physics?
A. Yes.
Q. That is right. And you have listed a number of
important and quite high value IT cases in which you
have been an expert?
A. That is right.
Q. You have a wealth of experience of being an expert?
A. I have. I have tried over the past 15 years to spend
not more than half of my time being an expert.
Q. I understand. Do you have any particular qualifications
in statistical techniques?
A. Well, I'm trained as a scientist and engineer, and as
such I have been using mathematics all my life. And
a part of that mathematics is probability theory and
statistics. I have a PhD which applies that
mathematical extensively. I have not any formal
qualification in statistics, but I don't regard
statistics as distinct from the other body of my
mathematical knowledge.
Q. Indeed if we look at page {D3/4/2} of that document --
it is just coming up now. There will sometimes be
a slight delay while the Opus operator loads the
page up.
A. Yes sure.
Q. -- if you look at the middle of the page you will see
July 1997 to July 2010 in bold. Do you see that?
A. Yes, right.
Q. The first paragraph begins:
"Acted for WPL ..."
Do you see that?
A. Yes.
Q. If you go to the right-hand side of that paragraph which
ends with "June 2010" and you come down two lines, you
see:
"Robert gave evidence on software development
methods, advanced statistical techniques ..."
Do you see that?
A. Yes.
Q. The phrase "advanced statistical techniques" reflects
the application of the expertise that you have just been
talking about?
A. Absolutely. I believe that the statistics I have
applied as a scientist particularly but also as
an engineer goes to the point of advanced statistical
techniques.
Q. And in statistics there are some matters which are
questions of approach, how should we go about looking at
something?
A. There are indeed.
Q. And there are other matters which are maths, which are
either right or wrong?
A. Yes, there is a lot of strain mathematics which dates
back to a long time.
Q. And you felt completely comfortable giving your evidence
on statistics and probability in this case?
A. Yes. I should say SAS is a statistics package, so it
has an advanced statistical method built into it and
I felt thoroughly comfortable examining those methods
and so on.
Q. Yes. And was it you who was doing the statistics
mostly, or Mr Emery, or a team effort?
A. No, Mr Emery was not involved in this dispute, it was
just solely me. Sorry, in this dispute I did the maths.
Q. You did the maths?
A. In SAS v WPL it was only me.
Q. I understand. But in this dispute, you and Mr Emery
worked together?
A. That is right.
Q. You did the maths?
A. Yes.
Q. Did Mr Emery check the maths?
A. No, he didn't actually. The maths is there in the
spreadsheets for anybody to check. It is not advanced
statistics, it is basic multiplication and division, and
it is laid out in the spreadsheets.
Q. Did you check it yourself?
A. Absolutely.
Q. Probably several times given you have checked for other
errors and made a correction?
A. Well, the correction was not a correction to the maths,
the correction was adjustments to the input assumptions.
Q. Let's just pause. I want to take you to one correction
you have made. It is at {D3/1.1/1}. This is
a corrections document to your report, isn't it?
A. Well, I haven't got -- I mean, the screen is not showing
me anything I can see is relevant at the moment.
Q. No, I'm just identifying the documents.
A. Right. This is corrections to my first report, yes.
Q. It says there just underneath what we call the tramlines
where it says "Corrections to Dr Worden's Expert
Report", it says you have identified some minor
corrections and clarifications?
A. Yes.
Q. If we go to page {D3/1.1/2} and if we look down at 761
towards the bottom, you say:
"the chances of the bug occurring in a Claimant's
branch would be about 2 in 10 million"?
A. Yes.
Q. And it should be:
"the chances of the bug occurring in a Claimants'
branch would be about 2 in a million."
A. That is right.
Q. So what was stated there was out by a factor of 10, and
that's why you've corrected it?
A. A correction had to be made. Yes, that is right.
Q. Did you notice any other errors?
A. No, I didn't because there were quite a few statements
of this "in a million" nature in my report. I made
careful checks of the main thread that led to my key
result in section 8.7, and that was my focus. And
similarly on the transaction corrections I had made
checks to that spreadsheet and, you know, I included the
spreadsheets with the reports so that you can see how
the arithmetic is done from the spreadsheets.
Q. And some of it you point out is basic arithmetic?
A. I think all of it is basic arithmetic.
Q. I'm very grateful.
Now, have you been involved in any group litigation
as an expert?
A. No, this is a new experience for me.
Q. I understand.
A. I think I haven't. No, I'm pretty sure I haven't.
Q. It is right, isn't it, you are not a behavioural
economist by training?
A. Absolutely, I'm not a behavioural economist.
Q. You know who Richard Thaler is?
A. No, I don't.
Q. He won the Nobel prize for economics in 2017 for his
work in behavioural economics. That is not your field
of expertise?
A. No, it is not my field of expertise. My knowledge of
economics is shallow and really my knowledge is much
more about business finance because I have been involved
in managing parts of various businesses.
Q. But you have referred in your report -- if we look at
{D3/1/107} and if we look at paragraph 415 there.
A. Yes.
Q. You have said there:
"I can go a little further than this ..."
This is referring to approaches to challenging
discrepancies:
" ... by making weak inferences about how a manager
of a small business, such as a Subpostmaster, needs to
prioritise his time in monthly balancing and other
evidence."
A. Yes.
Q. "The Second Witness Statement of Ms Angela Van Den
Bogerd at paragraph 187 says: 'Generally, when
discrepancies are of a value of several hundreds of
pounds, I would expect Subpostmasters to contact NBSC.'"
A. Yes.
Q. Now, what then follows is this statement:
"If this is correct, it is consistent with my
estimates."
That suggests you have made your own estimate of the
value at which a subpostmaster would contact NBSC?
A. Yes, I have made my own estimates, and you can see they
are rather more detailed than Mrs Van Den Bogard's
witness statement. In other words, she's put a point on
the spectrum. I have made some estimates about what the
spectrum is.
Q. Can we just be careful with what we are talking about,
because it does matter. You understand, particularly
with your background as a scientist, the difference
between an estimate and an assumption, don't you?
A. Absolutely. I think I do, yes.
Q. It is pretty fundamental. An estimate is based on some
facts --
A. I have made a number of -- well, have I been loose in
terminology between --
Q. Dr Worden, at present I'm just trying to clarify with
you what you have and have not done. I'm not going to
tie you to the words, but I'm going to show you the
words you use to clarify what you have in fact done or
not done.
A. "I assume the following, as best --"
Q. Dr Worden, would you allow me to take you through it?
A. I'm sorry.
Q. It will just help you to identify to show what you I'm
asking --
A. I'm just reminding myself that it is actually
assumptions I'm making in this paragraph.
Q. That's what I'm asking you, because there are a number
of ways of reading that paragraph. Because after the
quote of Mrs Van Den Bogard's witness statement, which
is in bold and ends with the words "paragraph 187)",
there's then this sentence:
"If this is correct, it is consistent with my
estimates."
Then there is a separate sentence which says:
"I assume the following, as best assumptions ..."
Now, the first question for you is this: when you
use the phrase, which we have also seen elsewhere, or
similar, "if this is correct, it is consistent with my
estimates", now, the use of the word "estimate" there,
I think we have agreed, tends to suggest something
different to "assumption", doesn't it?
A. Absolutely, yes.
Q. So can you please just tell the court did you make your
own free-standing estimate without regard to Angela Van
Den Bogard's witness statement of the levels at which
a subpostmaster would or would not challenge
a discrepancy by ringing the NBSC?
A. To clarify wordings, in my reports --
Q. Could you answer the question first?
A. In my report I made various estimates following from
an assumption. In other words, I made assumptions to
put into the maths which then constituted my estimates.
So that's the distinction in my mind between assumption
and estimate. And when I say it is consistent with my
estimates, it is consistent with estimates following
from my assumptions.
Q. Right. To clarify that sentence I'm asking you about,
where it says "if this is correct, it is consistent with
my estimates", you are not saying that you have any
independent basis upon which to support what Angela Van
Den Bogard says, are you? You didn't make any estimates
that confirm her witness evidence?
A. Well, there is a process you use in engineering of
ballpark estimates, which are imprecise -- or ballpark
assumptions -- I'm sorry if I get confused about the
wording here -- which are very approximate assumptions
you make in order to drive a calculation through. And
then you look back at the calculation and you say: is
the precision I have achieved in that calculation
sufficient or do I need to refine/revisit some
assumptions?
So you are always looking at the precision you need
to achieve in the result when you are concerned with how
precise should my assumptions or my estimates be. How
precise should my assumptions be in order to drive this
set of estimates which arrives at a number. How precise
do I need that number to be? I felt that given the
precision I needed or the court needed in the final
result, the precision of these assumptions, you know,
whether it is £300 or £400, for instance, was
sufficient.
Q. I will just take you back to the question for the
moment. We understand you made assumptions.
A. Yes.
Q. And you have set them out.
A. Yes.
Q. And you have said they are assumptions.
A. That is right.
Q. But you didn't make any fact-based estimate of what
would in fact happen yourself, did you?
A. I didn't go and interview a postmaster, for instance, or
anything like that. They were assumptions and they are
advertised as assumptions, and so if the court wishes to
change the assumptions and drive them through the
calculation, for instance, the calculations were
intended to have that ability, so that if the court --
I make some assumptions, if the court decides to find
something different then the court can drive those
different findings through my calculations.
Q. Okay. I will put it one last time. The answer is no,
isn't it, you didn't make --
A. I think I said no. Yes, it is no. I didn't make
a fact-based --
Q. Yes. To do so you would have needed the relevant fact
background, wouldn't you, to make an assessment of that?
You would have had to understand the process, what sort
of headwinds people faced when they sought to report
things, how they valued their own time etc?
A. To get precise numbers --
Q. To make an estimate not an assumption?
A. To get precise numbers I would have had to do that.
Q. Or any number that was an estimate rather than
an assumption, you would need some facts upon which to
make an estimate, wouldn't you?
A. Well, as I say, my view of the word "estimate" is
an estimate is what I made following my assumptions.
Q. No, Dr Worden --
A. Is that unclear? That I made some assumptions and put
them into some maths and that resulted in some
estimates.
Q. So the estimates that you refer to are the outcome of
having made assumptions?
A. That is right.
Q. I understand. And you accept that you had neither the
facts nor the relevant expertise to make estimates which
effectively fall within the field of behavioural
economics?
A. I accept that entirely.
Q. Thank you.
A. I won't say more.
Q. Thank you.
Just stepping back a minute, you have got a PhD in
theoretical particle physics, yes?
A. Mm.
Q. You have this wealth of experience as acting as
an expert and so forth?
A. Mm.
Q. You have made very detailed statistical calculations in
the report?
A. I wouldn't call them very detailed. I think you know
there are a number of multiplications there and
divisions and so on, and where the more detailed
statistics came in was estimating what sort of sample
size I would need to get more precision in the result.
Q. I understand. So you went into more detail in those
areas?
A. Occasional footnotes addressed these questions, yes.
Q. And you spotted the correction of the typo when you went
through your report?
A. Yes.
Q. Order of magnitude of 10?
A. Yes.
Q. Would you say, Dr Worden, you are a details man?
A. Physics people tend to focus; they don't like loads and
loads of facts, they like to focus on the core analysis
and make sure that's very right.
Q. I understand.
A. So they are detailed people in that sense.
Q. I understand. When we look at Mr Emery at {D3/5/1}, we
have got his CV.
A. Yes.
Q. And we can see he has obviously had a lot of IT
experience over many years. And if you go over the
page, please, {D3/5/2} we can see that he has a degree
in computing science from Imperial College?
A. That is right.
Q. And if we go to page {D3/5/5}, we can see where he
worked and over what period.
A. Yes.
Q. He was obviously at Logica for many, many years,
wasn't he?
A. That is right. We worked together at Logica in 1976, or
something like that.
Q. I understand. Is that how you have ended up together?
A. Well, obviously, you can see from his resume he has been
at Logica for a fair old time and he's been at
Charteris, and Charteris was a kind of spin-off Logica,
and I've been there. So we've overlapped quite a lot.
Q. Does Charteris still do some consulting for Logica?
A. Not for Logica. Logica doesn't exist anymore.
Q. When did it cease to exist?
A. It was taken over by CGI, I think, and ceased to exist
as a name about ten years ago, I think.
Q. What I really meant was I think at that point quite
a lot of people left?
A. Well, there have been various watersheds, yes.
Q. Was Credence originally a Logica solution?
A. That's a very good question. I haven't really looked
into that. Logica obviously had some role to do with
Credence but certainly I was nothing to do with
Logica at the time.
Q. Had you had experience of data warehousing when you were
working for Logica?
A. Yes, we -- a team that worked for me was involved in
business intelligence and that involved data
warehousing. And generally I have been doing databases
since the year dot and that includes data warehouses.
Q. Because data warehousing was quite big in the late
1990s, wasn't it?
A. It has been big since then, yes.
Q. In fact, Logica was quite strong in that area in the
early 2000s?
A. Well, I had left them by that point.
Q. I understand. That is right though, isn't it?
A. Logica was a bit of a jack-of-all-trades, actually.
Q. Okay. And just going back to Mr Emery, in your
experience of Mr Emery, I don't want to be unfair, but
is he a details man?
A. I think he is more than me, yes.
Q. And are there any parts of the expert report that he has
not checked? I think you suggested earlier he might not
have checked the maths?
A. That is right.
Q. Is that right?
A. I don't think he has checked the maths.
Q. But otherwise would he have read through --
A. He certainly read through the whole report and the
appendices.
Q. And you have probably read through it several times as
well?
A. I'm afraid so.
Q. I'm afraid so. How many times?
A. How many times I read through my report? Yesterday
and -- about five times, I suppose.
Q. Five times. So with your eyes on it and Mr Emery's eyes
on it, we have four eyes, haven't we?
A. Yes.
Q. Probably other people who I want ask you about have
looked at it as well, lots, and some of the people who
have looked at it looked at it many times, yes?
A. Well, I have looked at it many times, Mr Emery has
looked at it many times and I suspect the lawyers have
as well.
Q. Can I ask you about how you have approached the
assumptions upon which you have given your opinions in
your report.
Could we look, please, at page {D3/1/153}. We are
going to look, if we may, at paragraph 650. Do you have
that?
A. Yes. "Receipts/Payments Mismatch".
Q. Now, this is looking at the receipts and payments
mismatch bug at the bottom of page 153.
A. Mm.
Q. You see the heading there "8.6.1"?
A. Yes.
Q. And you say:
"This issue is cited in paragraph 5.6 of Mr Coyne's
report."
A. Mm.
Q. "It involved a bug in Horizon which was triggered by a
rare circumstance (which one would not expect to be
exercised in testing) and which had an effect on branch
accounts."
A. Yes.
Q. "If Mr Godeseth's evidence about this bug is not
accepted, I shall revise my opinions accordingly. They
are based on written evidence - particularly on a
written analysis by Gareth Jenkins ..."
Yes?
A. Yes.
Q. "... as well as the Second Witness Statement of
Mr Godeseth."
A. Yes.
Q. You make a consistency point there. Can we just go back
now you have seen the whole paragraph?
A. Yes.
Q. I just want to ask you about three lines down in that
paragraph, 650, on the right-hand side the following
sentence:
"If Mr Godeseth's evidence about this bug is not
accepted, I shall revise my opinions accordingly."
Now, pausing there. Let's just only look at this
example in your report. Let's just focus on this.
In this example what you have done is premised the
opinion that you give on the court accepting
Mr Godeseth's evidence.
A. That is right.
Q. But you have fairly said that if the court does not
accept that evidence "I will revise my opinions
accordingly."
A. Yes.
Q. The content of this report in this respect, we will come
to other examples later, focuses on what the
consequences would be if the defendant's factual
evidence is accepted?
A. Generally, if findings -- if things happen in oral
evidence which go against witness statements then
I would need to come back and say what's the impact.
Q. Yes. You understand, don't you, that the moment at
which everyone finds out whether factual evidence is
accepted or not is when the judgment is handed down?
A. Absolutely, yes.
Q. So you are proposing to revise your opinions after that,
are you?
A. No. That's a good point. What I was saying there is if
it becomes evident, all I can do is make assumptions.
The court will find findings and all I can do is make
assumptions and drive them through my opinions and try
and assist the court that way.
So all of this is assumptions I have made based on
the evidence I have seen, and the court may find
differently.
Q. Can I pause there. You were assuming there that
Mr Godeseth's version was true?
A. I was assuming mainly Mr Jenkins' written analysis which
Mr Godeseth's evidence confirmed.
Q. Let's just go with that for the moment. Have a look, if
you would, please, at {D3/1/260}. It is paragraph 1194.
A. Here we go, yes.
Q. You recognise this page because you were shown it --
A. Yes.
Q. Now, come down four lines and look at the right-hand
side about three-quarters of the way across.
A. Yes.
Q. You see "I have not assumed that any particular version
of events is true ..."
A. Yes.
Q. Yes? Now, what you say, if we look at the transcript,
please, at the bottom of page 26 of today's transcript,
just a few moments ago --
A. Can I get that? Yes.
Q. See the bottom of 26, the question is:
"So you are proposing to revise your opinions after
that, are you?"
And you say?
"Answer: --
A. Sorry, I haven't got the right line.
Q. It is the bottom of page 26, halfway down, line 23. Do
you have that?
A. Yes, I have got that.
Q. Dr Worden, if I'm taking you too fast at any point let
me know --
A. No, it is just me finding my way, that is all.
Q. Not at all, it is not always straightforward.
Line 23:
"Question: So you are proposing to revise your
opinions after that, are you?
Your answer is:
"Answer: No. That's a good point. What I was
saying there is if it becomes evident, all I can do is
make assumptions. The court will [make] findings and
all I can do is make assumptions and drive them through
my opinions and try and assist the court that way.
"So all of this is assumptions I have made based on
the evidence I have seen, and the court may find
differently."
A. Yes.
Q. Do you feel that overall in your report you have
provided opinions on the footing that the claimants'
evidence is accepted as often as you have provided
opinions on the footing that the defendant's evidence is
accepted?
A. Well, my analysis of the claimants' evidence is mainly
contained in my supplemental report, and I explained
there that individual claimants' evidence, particularly
individual subpostmaster evidence, I did not feel able
to make strong use of that and I gave the reasons in my
supplemental report.
So my opinions have little dependence on that. And
the core of my opinions, the numerical estimates I make,
those estimates have been designed -- or the process and
the method has been designed so that if -- so that to
make my assumptions evident where those assumptions come
in, and if the court finds something different from my
assumptions the court can go to the spreadsheets and
re-do the method for itself.
So I have tried to make the dependence of what the
court may take from my reports as little dependent on my
assumptions as possible. I'm not sure if I'm answering
the question.
Q. Let me give you another example. This is a slightly
different situation; let's look at that as well. It is
on page {D3/1/206}. It is where you are dealing with
transaction corrections.
A. Yes.
Q. They were set out, you understood, in Mr Smith's witness
statement?
A. Evidence was in that witness statement which I thought
bore on the issue of erroneous transaction corrections,
yes.
Q. Let's have a look. If we go over the page, please,
{D3/1/207} we can see that you say, 934:
"I proceed on the assumption that these figures
(which are the only ones available to me) are accepted
by the court. If they are not, a different calculation
along the same lines may possibly be appropriate."
A. Yes.
Q. We will come back to what available evidence there was
about TCs later. For now, you will accept that again is
proceeding on the footing that the evidence of the
defendant will be accepted?
A. Yes, I don't see that evidence of the defendants is very
relevant to this calculation.
Q. I understand.
Now, you have been in court, haven't you, for the
whole of the trial, I think?
A. Not the whole trial.
Q. Did you miss any days?
A. I think I must have missed some days, yes.
Q. Did you hear the defendant's evidence?
A. I certainly heard Mr Rolls' evidence.
Q. No, the defendant's evidence.
A. Sorry, the defendant's --
Q. I will say Post Office.
A. I heard all of that, yes.
Q. You heard all of that?
A. Pretty sure I did.
Q. Have you had any changes of heart about anything in your
report having heard it?
A. Changes of heart?
Q. Anything you want to change about what you said?
A. Well, had there been some fundamental change I would
have felt obliged to communicate it to the court.
Q. Of course, because you take your duty very seriously?
A. Yes, I do. I can't think of any major change of heart,
but there may well be things that you bring me to and
I say there's an adjustment here. But to come back to
the point about precision, I am conscious of how precise
my numerical results have to be in order to be of
assistance to the court.
Q. Well, that's appreciated.
A. And therefore that level of precision is my yardstick
for saying: has my opinion changed? Does the court need
to be informed about it? And so on and so forth.
Q. I see. Let's look at some particular facets of your
evidence. Let's look, if we may, please, at {D3/1/239},
which is paragraph 1086 in your report. Do you have the
page there? It is the bottom paragraph.
A. Yes.
Q. You say, second line, do you have halfway through the
second line the words "when Post Office"?
A. Yes.
Q. "When Post Office is investigating anomalies reported by
subpostmasters, they use Credence and their other
management information systems in the first instance
..."
Just pausing there, it is pretty important to
identify with care and clarity what systems and
information are available to Post Office, isn't it, in
this case because it is one of the issues?
A. As far as the experts can find out it is important, yes.
Q. Exactly, so far as the experts can find out. We will
come back to that in a minute, but let's just focus on
this.
So you agree it is important to identify what
information and systems is available as far as the
experts can find out. Just recapping on that sentence
in the third line of that paragraph, so just remind
yourself what I was asking you:
" ... they use Credence and their other management
information systems in the first instance ..."
Yes?
A. Yes.
Q. There is a footnote there, footnote 41?
A. Yes.
Q. There you say:
"The Witness Statement of Ms Tracy Mather,
16 November 2018, is consistent with my understanding."
A. Yes.
Q. Pausing there, did you get your understanding materially
from her witness statement or did you form
an understanding from something else that was consistent
with her witness statement?
A. I formed an understanding from something else, which is
largely my understanding of things like POLSAP, things
like Credence and my experience of how --
Q. And the opportunities you have had to look at that since
you were instructed?
A. To look at what?
Q. To look at the overall system, documents and so forth?
A. Yes, that was combined with my experience of how
businesses worked.
Q. Of course. So let's look at what Ms Mather said in that
witness statement to which you have referred.
It is at {E2/8/3} and it is paragraph 12.
A. This is not one I have looked at very recently.
Q. Don't worry, I will take you through what it says. It
is a short paragraph.
A. (Reads to himself) Yes.
Q. Let's take it is this stages, firstly as an information
tool, I think that's agreed?
A. Yes, it is also called POLMIS in certain documents.
Q. Yes. That's the same ...?
A. My understanding is that POLMIS and Credence are
different names at different times for the same system.
Q. Same system. Because "POLMIS" stands for Post Office
Limited Management Information System?
A. That is right.
Q. And if we look at the second sentence, it says it is
designed to work alongside other applications?
A. Yes.
Q. So there are a number of different things they can
look at?
A. Yes.
Q. "It is used to help understand what has happened in
a branch as it records all keystroke activity performed
in that branch by the user ID, date and time ..."
Do you see that?
A. Yes.
Q. That was your understanding when you wrote your report,
wasn't it?
A. Yes.
Q. Had you formed that understanding from any other
documents that you had seen or was that understanding
only formed from reading Ms Mather's witness statement?
A. No, I had formed it from other documents.
Q. Can you remember what other documents you had seen which
showed that Credence recorded keystrokes in the branch?
A. To that level of keystrokes in the branch -- no, to go
into that level of -- not level of misunderstanding,
level of understanding, to go to that level of
understanding I think Mrs Mather's witness statement was
useful to me. But the point that Credence is an MIS and
Post Office use it to determine what happened in the
branch and so on, those points were evident to me from
my reading of quite a lot of evidence.
Q. But you will forgive me, Dr Worden, if I ask you the
precise question again, which is this: in relation to
Credence recording "all keystroke activity performed in
that branch", was that your understanding when you wrote
your report or not?
A. I did not go as far as understanding all keystroke
activity. In other words, I felt that Credence was
a fairly comprehensive record of things that had gone on
in the branch, but the phrase "all keystroke activity"
would not have occurred to me before her witness
statement.
Q. So --
A. Because that is a very detailed level of information,
and my experience of MIS is that you are taking all
sorts of slices of information, and really drilling down
to the keystroke is very detailed.
Q. But you would accept, Dr Worden, that whether factually
you could look at the keystrokes' activity performed in
a branch or whether you could not, as Post Office, do
that, would be important, wouldn't it?
A. No, I do not think so. For the purpose of knowing that
Post Office could investigate events in the branch using
their MIS, I did not need to know that it went down to
the keystroke level.
Q. Why not?
A. Because MISs typically deal with more summary
information and they are designed based on the
requirements, and if the requirement doesn't require to
go to keystroke level that's not a part of the MIS. And
I didn't feel myself that going to keystroke level was
necessary for the level of investigation that I believed
needed to be done in Post Office.
Q. So you made a judgment about what needed to be done by
way of investigation and you made a judgment that
keystrokes were not necessary to the level of
investigation that you formed a judgment about?
A. Yes. To me, keystroke level is very, very detailed, and
from my experience of MISs, keystroke level is not often
required in an MIS. An MIS typically deals with more
summary information and ways of getting different slices
of information, and so on and so forth. Going down to
the keystroke level is, to my mind, very detailed.
Q. We will come back to this page in a moment, but can we
just go, please, to {C1/1/2}.
The document you are now looking at is not the
document that was on the screen before.
A. No, this is --
Q. These are the Horizon Issues which you probably
recognise?
A. Yes, that is right. Yes.
Q. And if you look halfway down you will see "Operation of
Horizon" in capitals and in bold?
A. Yes.
Q. Under that, there is one subheading, "Remote Access",
and there is another heading, "Availability of
Information and Report Writing". Do you see that?
A. Yes.
Q. Under that, there is issue 8, and issue 8 says:
"What transaction data and reporting functions were
available through Horizon to Post Office for identifying
the occurrence of alleged shortfalls and the causes of
alleged shortfalls in branches ..."
Yes?
A. Yes.
Q. So it was a specific Horizon Issue that you were
instructed to report on?
A. Yes.
Q. To identify what transaction data and reporting
functions were available through Horizon to Post Office?
A. Yes.
Q. And the reason that you weren't interested in whether or
not keystrokes were available was the one you have just
given to the court a moment ago: that you formed
a judgment about what would be necessary, and on the
basis of that judgment you formed a judgment that
keystrokes wouldn't be important?
A. Well, broadly that is correct. I should say also
there's a kind of prioritisation that the expert has to
do, how far he can drill down in particular areas and
what drilling down is used for.
Q. And you felt, if we go back to Ms Mather's report -- it
is {E2/8/3}, it is a short paragraph and it makes quite
a key point about keystroke activity, doesn't it?
A. It does.
Q. And she positively said that it records all keystroke
activity in that branch by the user?
A. Yes.
Q. And when I started asking about this it did sound as if
you were saying you had separately formed the view from
other documents that that was possible?
A. No, I had said I had separately formed a view from other
documents that Credence was used to investigate what
happened in the branch.
Q. Right. Let's go back, please, to the paragraph
I originally asked you about, which is {D3/1/239}. It
is paragraph 1086. You will remember it is at the
bottom of the page.
I asked you about the second sentence of that
paragraph:
"When Post Office is investigating anomalies
reported by Subpostmasters, they use Credence and their
other management information systems in the first
instance ..."
Yes?
A. Yes.
Q. You give the footnote at 41?
A. Yes.
Q. And you say:
"The Witness Statement of Ms Tracy Mather ... is
consistent with my understanding."
Yes?
A. Yes.
Q. And that's the witness statement we have just been
looking at.
A. Yes.
Q. Which specifically says that Credence records all
keystroke activity.
A. Yes.
Q. So you accept that this would suggest to anyone reading
it, whether it is the court or the claimants or anyone
else, that you agreed with her description of what
Credence could do?
A. No, that's not quite what I said.
Q. You say it is consistent --
A. Her description is consistent with my understanding and
went beyond it.
Q. You didn't spell out anything -- should we read
throughout your report where you say "consist within the
my understanding" as meaning "there is some aspect in
which my understanding coincides with what I'm referring
to"?
A. I think we read "consistent" in the ordinary sense of
"consistent".
Q. I see. But on the face of Mrs Mather's statement you
would agree, wouldn't you, that Post Office had access
to information that the SPM doesn't have if she is
right?
A. Well, whether or not she is right, Post Office has
information that the branch doesn't have.
Q. Of course we will come to that in more detail. I'm just
focusing on the keystrokes.
If she is right about the keystrokes being available
to Post Office, they have effectively a record of what
buttons were pressed in the branch?
A. Yes.
Q. And the SPM doesn't have them?
A. I believe that is right, yes.
Q. And so you were here I think when Mrs Van Den Bogard
gave her evidence, weren't you?
A. I certainly read the transcripts, but --
Q. Let's refresh your memory. Sorry, I interrupted you,
Dr Worden.
A. I do not think I was actually in court when Mrs Van Den
Bogard was --
Q. Okay, let me show you the transcript then, to be fair to
you. {Day6/69:1}, please.
(Pause)
We are going to look at lines 10 to 23. There is
a question:
"Question: ... the point is this, that the Credence
report would show the specific keystrokes by the
operator in branch, wouldn't they?"
Do you see that?
A. And she says yes.
Q. She says yes, and then:
"... so if one wanted to get to the truth about what
keystrokes Mr Patny had pressed, one could have obtained
a Credence report to identify that and that would
have --
"Answer: Have used a Credence report, yes.
"Question: And that would have shown all the
keystrokes pressed --
"Answer: Yes.
"Question: -- so you could follow exactly what he
had done in sequence?
"Answer: Yes."
Do you see that?
A. Yes.
Q. Were you here for the evidence of Mr Patny?
A. I do not think I was.
Q. Where we looked at the internal Post Office documents
where he had been promised a Credence report and it
didn't seem to come?
A. I do not think I was here for that evidence, no.
Q. Now if we look, please, at page {Day6/104:1} because
Mrs Van Den Bogard was re-examined about that.
(Pause)
We are going to look on this page at lines 2 to 10.
A. 104, 2 to 10.
Q. Just to orientate you, this is my learned friend
Mr de Garr Robinson, leading counsel for the
Post Office, re-examining Mrs Van Den Bogard about the
section of cross-examination I have just shown you. Do
you understand?
A. Yes.
Q. Okay.
"Question: Now, the reason why I'm putting this to
you is because Mr Green put to you that Credence shows
every single keystroke that's pressed by the postmaster
and you accepted what he said and he put it to you on
the basis that this is what Mrs Mather said. Could I
ask you this question: is it the case that Credence
gives an account of every single key that's pressed by
the postmaster in branch?
"Answer: That's not my understanding, not every
single stroke.
"Question: I'm grateful."
Do you see that?
A. Yes.
Q. So in re-examination there was a different picture.
A. Yes.
Q. But I think you said you were reading the transcripts.
Did you notice this?
A. I didn't notice this. I'm afraid I didn't, no. I can't
say I have read every line of the transcript because it
would take a day to do it for each --
Q. I understand.
Okay. Let's look at page {Day6/149:1}.
A. Right.
Q. Because Mrs Mather was the person whose witness
statement you had actually positively referred to,
wasn't she?
A. By me? Yes.
Q. Yes, it is in your report. You say look at that --
A. She's consistent with my understanding, yes.
Q. Yes. And if we look on page 149, look at lines 7 to 20.
Again, this is my learned friend, leading counsel for
Post Office, re-examination Mrs Mather this time.
Sorry, examining Mrs Mather in chief. That's before she
is cross-examined.
A. Right, okay.
Q. Yes?
A. Mm.
"Question: And now, Mrs Mather, I have to ask a
question which is born out of sheepishness on my part.
Everyone else in court will understand why, but you
won't. I would like to ask you about paragraph 12 of
your statement please ..."
That is the one we have just been looking at.
A. So that is --
Q. We saw what it said. It was perfectly clear, wasn't it?
A. The keystroke --
Q. Yes, that is right.
"Question: Here you are describing Credence and you
say:
"'Credence is used as an information tool.
It is designed to work alongside other
applications. It is used to help understand
what has happened in a branch as it records all
keystroke activity performed in that branch by
the user ID, date and time.'
"Could I ask you to explain what you mean
by the phrase 'It records all keystroke
activity'?"
And she says:
"Answer: What I actually meant was the
transactional data as in sales and non-sales."
A. Right.
Q. So that was a bit of a surprise.
A. Yes. I mean --
Q. Hence your reaction by saying the intonation of the word
"right" in your answer which doesn't always come out on
the transcript?
A. Yes. This, if you like, Mrs Mather's witness statement,
down to keystroke, that was consistent with my
understanding but it went beyond it. And now what's
happening here is it doesn't go beyond my understanding
as far as it did before.
Q. So it recedes back to what your unspecified --
A. It comes closer to what I would have expected from
an MIS.
Q. I don't want to be unfair, but were you sort of guessing
based on experience about what an MIS would have?
A. Yes, I have lots of experience of what an MIS has.
Q. So that is the understanding you are referring to?
A. That is right, yes.
Q. So --
A. Also what I can infer from the documents I have seen,
you know. I put those together with my experience and
one imagines what Post Office would need in order to
investigate what went on at the branch, and my
understanding is they would need what Mrs Mather says
here, transactional data. They would not need
individual keystrokes. That's my -- what my experience
tells me.
Q. So is it right, then, that you hadn't appreciated
whether or not keystrokes were recorded in branch either
way despite referring expressly to Mrs Mather's
statement when you made your first --
A. I said Mrs Mather's statement was consistent with my
understanding, and it was, and it went beyond my
understanding. And now we have seen that the extent to
which it went beyond my understanding varied at the
time.
Q. It is fair to say that the finesse that you are now
making was not expressly explained in your first report,
was it?
A. What I said in my first report was it is consistent with
my understanding, and it was and it is.
Q. But what she said about keystrokes couldn't have been
consistent with your understanding about that because
you had no idea?
A. If somebody tells you something that is consistent with
what you think and goes beyond it, then it is
consistent. That's all I'm saying.
Q. So the court should have that in mind for every use of
the word "consistent" in your report?
A. I think I'm using the word "consistent" in the ordinary
sense of the word.
Q. I understand.
A. That things can be consistent if they overlap in
a consistent way.
Q. Let's look at a different example. For all those
reasons you didn't feel there was any need to change any
aspect of your report even on Horizon Issue 8 which
asked about what information --
A. No, there was always this trade-off that one wanted to
write twice as much as one could write, and there was
always this editorial trade-off of, how much to drill
down and what was useful to the court.
Q. But isn't it quite useful just to list to the court what
the information they have is, in answer to the question
"what information do they have", scope: they have this,
they have this, they have this, they did have this, they
didn't have that and they did have the other?
A. I was making decisions about how much detail is needed
here and there, and I made that decision not to do
a detailed list of this and this and this, and perhaps
I was wrong. I'm always having to make these decisions
about how much detail do you go into here in order to
assist the court.
Q. But you would accept, Dr Worden, that it is a binary
issue: do you have access to the keystrokes in the
branch or do you not?
A. That is a binary issue.
Q. So that could have been accommodated in your report by
the addition of a few words in that footnote?
A. It could have been.
Q. Or one line in your report?
A. It could have been. Yes.
Q. Let's move to a different example. I'm not going to go
into the detail of this because we are going to deal
with remote access later. I want to ask you what you
say at {D3/1/244}. Do you see that? If we can go to
{D3/1/244}.
A. Paragraph number?
Q. Let's just get the correct page up. There we are. It
is paragraph 1114. Do you see that's under the heading
"Transaction injection in Old Horizon"?
A. Yes.
Q. You say there:
"Mr Godeseth says that, in the Old Horizon
system ..."
Which is Legacy Horizon?
A. Mm.
Q. "... The SSC could also inject transactions and that
those transactions were clearly distinguished from those
entered at the branch because they would have included a
counter position greater than 32 when no branches would
have had such a high number of counters ..."
Yes?
A. Yes.
Q. And if we go forward, please, to page {D3/1/245}, so
turn over the page, at paragraph 1119 you are
considering what Mr Roll's evidence was.
A. Yes.
Q. You say:
"Mr Roll worked in the SSC, and I established above
that (during his tenure with Fujitsu) certain SSC users
had the ability to transact injections, although these
would have become visible to Subpostmasters. So, in my
opinion, Mr Roll could not have made these changes to
branch accounts 'without the Subpostmaster knowing'."
Yes?
A. Yes.
Q. Now, we know in fact now that Mr Roll was right about
that and we will come back to that when we do the remote
access piece. But at the moment I just want to ask you
about this paragraph and how it is expressed.
A. There is a typo. "Transact injections" is a bit --
I think it is "inject transactions".
Q. Don't worry about that, no one is going to complain
about that. What you say is that Mr Roll worked in the
SSC, yes?
A. Yes.
Q. Service support centre?
A. Yes.
Q. And then these are the words I want to ask you about:
"... I established above that (during his tenure
with Fujitsu) certain SSC users had the ability to,"
inject transactions I think it can be fairly be --
A. Yes.
Q. "... although these would have become visible to
Subpostmasters."
Yes?
A. Yes.
Q. So in your opinion, Mr Roll could not have made these
changes to branch accounts without the subpostmaster
knowing?
A. Yes.
Q. Now, that was a disputed issue of fact between Mr Roll
and the Post Office, wasn't it?
A. Yes.
Q. And you say that you have established it above.
A. Yes. Can we go back to the above?
Q. Let's go back to the above and I will show you. It
seems to the reader possibly that it is paragraph 1114,
which is {D3/1/244} and following?
A. Yes.
Q. Because this is the section where you basically say what
Mr Godeseth says: a counter position greater than 32,
yes?
A. Yes.
Q. And you say it accords with your experience and so
forth. And then you point out at 1116 that it couldn't
be done without the subpostmaster's knowledge, yes?
A. I say "thus". Let me see what leads before "thus".
(Pause)
Q. So what you have actually done, as we can see between
1114 and 1118, and then over the page on 1119, is you
have looked at a disputed issue of fact, you have
accepted Mr Godeseth's factual evidence in the face of
Mr Roll's account, you say you have established over the
page {D3/1/245} that effectively what Mr Godeseth says
is correct because you used the word "established". And
on that footing you have given the court your opinion
that Mr Roll is wrong. That's what you have done.
A. Yes.
Q. And you shouldn't have done that, should you?
A. I think I shouldn't. I think the word "established" was
too strong.
Q. Well, there are two points. (1) It is misleading to say
you have established it because all you have done is you
have said "I accept Mr Godeseth's evidence" and proceed
on that footing as if it is fact. That's the first
problem.
A. As I say, I think "established" was too strong and I was
wrong to say that.
Q. Well, okay. Let's go back. I took you to
paragraph 114. It is on {D3/1/260}. We had a look at
this didn't we, 114?
A. What paragraph are we talking about?
Q. Sorry, I misspoke, paragraph 1194.
Right-hand side, four lines down. That declaration
is not true in respect of what you did where you said
"establish", is it?
A. I think there are occasions in my report I have made
mistakes and this was a mistake, yes.
Q. And the particular type of mistake that we are talking
about is one that you have expressly said at the end you
have not done?
A. That's what this says, I think. I think on that
occasion, on this counter 32 issue, I did, although
I said at the time Mr Godeseth's evidence is consistent
or accords with my experience in that paragraph, I did
later at 119, or whatever it was, go on to draw
a stronger inference which I should not have done.
Q. You have based your opinion accepting Mr Godeseth's
evidence and on that basis you have rejected Mr Roll's
evidence of fact on a disputed issue, hotly disputed and
central to these proceedings. The answer to that is
yes?
A. Well, this has developed over time and what I have said
is that my statement in the first report was too strong,
and --
Q. But pausing there, your approach, it is not just the use
of the word "established", I'm putting to you fairly and
squarely that your approach was wrong. There were two
witnesses saying opposite things. As an expert you
should have said: if Mr Godeseth is correct, this; if
Mr Roll is correct, it undermines the proper access
controls that I mention in my witness statement as my
final countermeasure and undermines the reliability or
robustness, or whatever it is, of the Horizon system, or
whatever your view was if Mr Roll was correct. That's
what you should have done, isn't it?
A. I have tried in my reports to --
Q. Can you just answer the question: do you accept that's
what you should have done?
A. I'm just prefacing an answer. I have tried in my
reports generally to acknowledge the limitations of
factual witness evidence. On this occasion I was wrong
and I should have done so.
Q. Can you think of any prominent example in your report
where you proceed on the footing that the claimants are
right about anything?
A. Well, this depends on the claimants' evidence being
relevant to my opinions, and I can't think of any
occasions where individual subpostmasters' evidence, for
instance, is relevant to my opinions.
Q. But we know from this example that Mr Roll's evidence
was the evidence that was put forward by the claimants.
And you rejected it.
A. Yes.
Q. And we know that that's an important example because we
now know the answer that you were wrong to reject it,
don't we?
A. We do.
Q. So it was both wrong in its approach and wrong in its
outcome, wasn't it?
A. The outcome was wrong, no doubt. As I say, my approach
has been to try wherever possible to qualify factual
witness evidence with the statement: this is my
assumption and this is why I'm using an assumption to
proceed.
I have tried wherever I can to make that
qualification and we agreed on this occasion that
I should have made it more careful.
Q. I can only put it one more time: on this your approach
was wrong and your conclusion was wrong?
A. Yes.
MR GREEN: My Lord, would that be a convenient moment?
MR JUSTICE FRASER: Yes, we will have 10 minutes.
Dr Worden, you know this, I know, because you are
experienced and you also heard me reminding Mr Coyne
many times, but now you are in the middle of your
cross-examination you are not allowed to talk to anyone
about the case. We are going to have a 10-minute break
for the shorthand writers. Please don't feel you have
to stay in the witness box. I always encourage
witnesses to move around and stretch their legs. But we
will come back in at 11.55 am.
(11.46 am)
(A short break)
(11.55 am)
MR GREEN: Now, Dr Worden, just picking up from where we
left off just before the break, we had been looking at
{D3/1/245}, which is your first report, at
paragraph 1119. And I'm going to ask you about that.
Yes? Do you remember?
A. Yes, I do indeed.
Q. I think you may have mentioned that you address that in
your second report?
A. Yes.
Q. Between the first and second reports there had been
a further statement from Mr Roll and a further statement
from Mr Parker, hadn't there, Mr Parker's second witness
statement?
A. That is right.
Q. And you would have read those?
A. Yes.
Q. To inform yourself about whether the approach you had
taken in your first report was right or not?
A. Yes.
Q. It is the only fair way to do it, isn't it?
A. Well, obviously I did --
Q. Yes, you must have done. And paragraph 27 of
Mr Parker's second witness statement explained that what
Mr Roll had described was possible. Do you remember
that?
A. If we go to that --
Q. Let's have a look at it. It is on {E2/12/9}.
There's Mr Parker's second witness statement. Go to
page {E2/12/9} of that. If we look at paragraph 27, and
this is Mr Parker referring to paragraph 20 of Mr Roll's
second witness statement?
A. Yes.
Q. "... Mr Roll describes a process by which transactions
could be inserted via individual branch counters by
using the correspondence server to piggy back through
the gateway."
Yes?
"He has not previously made this point clear. Now
that he has, following a discussion with colleagues who
performed such actions I can confirm that this was
possible."
Yes? So we are now in a position where a point
favourable to the claimants has not only been set out in
the claimants' evidence, but agreed by the Post Office.
That is correct, isn't it?
A. Well --
Q. On the evidence.
A. The difficulty I have with this, or one difficulty
I have, is that I don't understand the phrase "piggy
back".
Q. Well, let's pause there. Just look at what he actually
says:
"He has not previously made this point clear."
Line 3:
"Now that he has, following a discussion with
colleagues who performed such actions ..."
So it is not theoretically possible. I have spoken
to people who were actually doing this. Having done
that, I can confirm that this was possible.
A. Yes.
Q. Right. Now, pause there. The premise that you
proceeded upon, which we have looked at in your first
report, was that Mr Roll was wrong, that it was
possible. This is a yes or no answer.
A. No, it is not that simple. In other words, that
I believe that what I did not accept in my first report
was that it could be done without the subpostmaster
knowing --
Q. Precisely.
A. And I was never very clear on -- I never got to the
bottom of the issue of what the subpostmaster might
know. Now, piggy backing says this was possible, but
I was not clear whether that would make it evident to
the subpostmaster.
Q. No, Dr Worden, the way it went was that Mr Rolls said it
was possible without the knowledge of the subpostmaster?
A. Yes.
Q. And that's what you were specifically responding to in
your first report?
A. Yes.
Q. And that's what you rejected which we dealt with before
the break, yes?
A. Yes.
Q. And you rejected it on the basis of Mr Godeseth's
evidence that it would be done through a counter number
higher than 32?
A. Right.
Q. So that it would be visible to the postmaster if they
noticed the counter number?
A. Counter number higher than 32 was one way in which the
postmaster could find out.
Q. Well, that was the way --
A. No, not the way. That's not what I believe.
Q. Well, that's what Mr Godeseth was saying, wasn't it?
A. That's what Mr Godeseth said at the time, yes. But --
Q. Let's go back to --
A. -- fundamentally, I believe that whether the postmaster
knew or not is not a simple matter of counter 32 or not.
Q. Dr Worden, you know where this cross-examination is
going, don't you?
A. No, I don't.
Q. Let's have a look and then I will bring you back to what
you have just said.
Mr Parker in paragraph 27 specifically says that
what Mr Roll has been describing --
A. Can I read paragraph 27 carefully, then. {E2/12/9}
This was possible --
Q. Without using counter 32, yes?
A. But this paragraph, as far as I can read it, does not
refer to the subpostmaster's knowledge.
Q. Okay. Well, I will have to come back and trace through
everything in a minute. Let me take you forward to what
you say in your second report and we will do it all over
again carefully. But {D3/6/20} is your second expert's
report. Okay?
A. Yes.
Q. So you were commenting, yes, on paragraph 82 at the
bottom of the page?
A. Yes.
Q. You are specifically commenting on paragraph 20 of
Mr Roll's witness statement.
A. Mm.
Q. Yes?
A. Mm.
Q. And you say:
"[He] addresses a factual point about injection of
transactions."
So when you say that, you know that that is
a factual matter in dispute, don't you?
A. Yes.
Q. You do. Second line:
"He says: 'Sometimes we had to ask for a specific
person to log in to the counter before injecting
transactions so that the software would not detect any
discrepancies. A transaction inserted in this way would
appear to the subpostmaster as though it had been
carried out through the counter in branch."
A. Yes.
Q. Then you add the words {D3/6/21}:
"He ... goes on to disagree with my paragraph 1119."
Which is the one --
A. Can we just go back to that paragraph again so we can
compare them side by side?
Q. Yes, we can.
MR JUSTICE FRASER: It might be sensible to have that in the
hard copy.
A. Yes, that's what I'm trying to do.
MR JUSTICE FRASER: Is that what you are looking for? It
is --
MR GREEN: 245.
MR JUSTICE FRASER: I don't know if Dr Worden has
a paginated one.
MR GREEN: I understand.
A. Sorry about this.
MR JUSTICE FRASER: Don't worry.
(Pause)
A. Now, this is chapter 11, is it?
MR JUSTICE FRASER: Does that document have a bold number in
the bottom right-hand corner, a bundle page number?
A. Yes, it does.
MR JUSTICE FRASER: If you turn to {D3/1/245}, is that the
page you want?
A. Right, okay, these would have become visible to
subpostmasters without the subpostmaster knowing were
there. We are there.
Q. So the genesis of this is that at 1119 you explain that
you have established above that these would become
visible to subpostmasters and rejected Mr Roll's
account. That's what you did at 1119. You can see
that?
A. That is right, that is correct.
Q. You have accepted that the approach and the conclusion
were both wrong?
A. I was overstrong.
Q. That the approach was wrong?
A. Yes, I have accepted that.
Q. And that the conclusion was wrong?
A. I have accepted that.
Q. And you also accepted that it is clear from the evidence
that the conclusion was wrong --
A. I accepted that I was wrong in accepting Mr Godeseth's
evidence over Mr Roll's evidence in my first report.
Q. But we now know, I put to you, that those changes can be
made without the knowledge of the subpostmasters?
A. Well, it is without the knowledge of the subpostmaster
that's the nub of the issue.
Q. So you write paragraph 1119, then we get Mr Roll's
second witness statement, and Mr Roll -- we won't go
there now but it is {E1/10/6}. Then we get Mr Parker,
his second witness statement?
A. Which doesn't refer to knowledge of the subpostmaster in
the paragraph you took me to.
Q. Okay. Let's go back to Mr Roll at {E1/10/6}.
Now, the passage that you have now accepted between
114 and 119 was perhaps not as it should have been?
A. Yes.
Q. Is the very passage to which Mr Roll is referring in
paragraph 20 on this page, isn't it?
A. Can I read this long paragraph, 20?
Q. Please do. (Pause)
A. Okay, I have absorbed that paragraph now.
Q. And he specifically explains two-thirds of the way down
in that paragraph "therefore". Do you see "and
therefore"?
A. "... and therefore not a number greater than 32,"
absolutely.
Q. "... and not in a way which would distinguish it in any
logs as having been inserted by Fujitsu rather than by
the subpostmaster or an assistant."
A. Yes.
Q. Then at the bottom of the page, five lines up on the
right-hand side, "A transaction inserted". Do you see
that?
A. Yes.
Q. "A transaction inserted in this way would appear to the
subpostmaster as though it had been carried out through
the counter in branch."
A. Yes.
Q. Couldn't be clearer, could it, what he is saying?
A. It is pretty clear but --
Q. "I therefore disagree with Dr. Worden's conclusions that
these transactions would always have been visible to
subpostmasters ..."
And he references the specific paragraph he is
looking at:
" ... if he means to say that they would be shown to
subpostmasters as transactions inserted by Fujitsu,
rather than as transactions which appeared to have been
created in branch."
A. Yes.
Q. Right. And that is what, when we go to Mr Parker's
second witness statement -- let's look at that again.
It is {E2/12/9}. Mr Parker, paragraph 27, we have
already looked at it.
A. Yes.
Q. Mr Parker is referring to paragraph 20. That's the
paragraph we have just looked at, isn't it, Dr Worden?
A. That is correct.
Q. "... Mr Roll describes a process by which transactions
could be inserted via individual branch counters by
using the correspondence server to piggy back through
the gateway. He has not previously made this point
clear. Now that he has, following a discussion with
colleagues who performed such actions I can confirm that
this was possible."
A. Yes.
Q. So the conclusion that it is not possible to do so, we
can see at this point has been reached by Mr Parker as
well as --
A. The question is what do we mean by "it"?
Q. Dr Worden, you knew that a central issue, not only
a central issue legally but a very high-profile issue in
the case, was the extent to which Post Office had remote
access to the counters, didn't you? You knew that?
A. Yes, and what I'm talking about, what I was talking
about was the extent to which this could happen without
the knowledge of the subpostmaster.
Q. And that's the --
A. And we agreed in the joint statement that more or less
Fujitsu or Post Office could do anything.
Q. Let's just take it in stages. That turns out to be the
case, but we will come back to that separately.
At the moment what I'm asking about is the approach
that you have taken to disputed issues of fact in your
reports. And this is one of them. We have identified
what went wrong with paragraphs 1114 to 1119. I'm going
to ask you now to look again; I have shown you Mr Roll's
second statement, I have shown you Mr Parker's second
statement. Look now, please, again. Go back to where
we were before, {D3/6/20}.
This is your second expert report?
A. Sorry, I'm not there yet.
Q. Don't worry. Do you see at the top --
MR JUSTICE FRASER: We are not there yet.
MR GREEN: Sorry. {D3/6/20}, "ER2 Worden". Do you see
that?
A. Yes.
Q. "Charteris" at the top right-hand corner. This is your
second report. We are going to look, please, at
paragraph 82 again.
A. Right.
Q. "In his paragraph 20, Mr Roll addresses a factual point
..."
And you accepted earlier that that was a factual
point that you knew to be in dispute?
A. There are two factual points in dispute. There is
whether SSC could inject transactions and whether they
could do it unknown to postmasters.
Q. Quite. And the witness statements we have been looking
at specifically deal with whether they could do it
unknown to subpostmasters.
A. Well, what they specifically deal with is a particular
mechanism for counter 32 by which subpostmasters might
find out. And that, in my opinion, is probably not the
only mechanism.
Q. Let's take it in stages. The answer you just gave was
very interesting. It was in two parts. You said
that -- give me one second -- they were dealing with
counter 32, yes? Those witness statements, the
counter 32?
A. Dealing with counter 32. What do you mean by that?
Q. I will give you your exact words in a minute when my
realtime starts working again, but the point you were
making is that the witness statements were referring to
this counter 32 point that had been raised by
Mr Godeseth.
A. Yes.
Q. And the way in which they were dealing with them,
Dr Worden, is that they were saying: actually, as well
as counter 32, which would be capable of showing
a subpostmaster that it had been injected by Fujitsu,
there was also a separate way, which would not. That's
what those witness statements were addressing, isn't it?
A. There was a separate way which would not do the
counter 32 thing, but then the question is whether that
separate way is unbeknownst to the postmaster.
Q. Well, Mr Roll specifically says it is unbeknownst to the
postmaster --
A. Where does he say that? Can we go back to that
statement?
Q. We can do it one last time. I think I am going to have
to move on. Mr Roll, paragraph 20, {E1/6/1}.
A. Before we move on, I would like a short opportunity to
try and explain what I was doing here.
Q. Let's go back to your second witness statement. If you
are going to explain something, can you explain why it
is in your second witness statement that you did what
you do.
Let's look at D3 --
A. I would really like to explain that.
Q. We will do it in stages. I will ask the questions and
you give the answers and explain what you need to
explain to his Lordship about how you have adopted this
approach as an expert.
MR JUSTICE FRASER: Can you each try not to talk over one
another. It is not just you; it is Mr Green as well.
MR GREEN: I apologise.
MR JUSTICE FRASER: The transcribers need to get everything
down, so one at a time.
So where are we going?
MR GREEN: {D3/6/20}.
A. This is a long paragraph, or is it?
Q. No, this is in your second witness statement --
A. Report.
Q. Sorry, your second expert report. I do apologise. And
I am trying to take you to paragraphs 82 to 85. We
never get past 82. So let's just see what paragraph 82
says.
A. Yes. It summarises apart from the end of Mr Roll's
paragraph 20.
Q. If we go over --
A. Could I get to my supplemental report so I can get over
the page?
Q. Of course. Please get any documents you wish,
Dr Worden. I don't want to stop you.
A. Where are we? Can somebody tell me the chapter?
Q. It is your second report. Your second report is -- you
should be able to find it -- {D3/6/20}.
A. 20?
Q. Page 20.
A. Right, okay.
Q. Do you have that?
A. Got it nearly. Sorry, I'm not there yet. Not there
yet.
Right, okay. Got it.
Q. You got that page?
A. Yes.
Q. Okay. And I was asking you about paragraph 82 at the
foot of that page.
A. Yes.
Q. I will just wait until you finish pouring the water?
A. Sorry.
Q. Don't apologise. I just don't want to -- if you want to
try and turn the pages. Now, we were looking at the
foot of page 20, weren't we?
A. Yes.
Q. "Mr Roll addresses a factual point about injection of
transactions. He says: 'sometimes we had to ask for
a specific person to log in to the counter before
injecting transactions so that the software would not
detect any discrepancies. A transaction inserted in
this way would appear to the subpostmaster as though it
had been carried out through the counter in branch."
{D3/6/21}
A. I think his actual 20 has a kind of "if" after it.
Q. Let's look at what you are citing.
A. Let's look at what I say, let's do that.
Q. You say:
"He then goes on to disagree with my
paragraph 1119."
Which is the one we have already explored in --
A. And that's where I said I didn't think it could be done
without the knowledge of the --
Q. That's where you say you disagree it could be done
without the knowledge of the subpostmaster.
A. Yes.
Q. You knew that a factual witness who had actually worked
at Fujitsu, he is not a claimant, he has volunteered to
come and give independent evidence, was saying that this
was possible?
A. He was saying it was possible without the knowledge of
the subpostmaster. My opinion is --
Q. Can we take it in stages, please, because you knew that
he was saying that and you also knew by the date of this
report that Mr Parker had agreed, didn't you?
A. No. Mr Parker had agreed it could be done. He did not
agree that it could be done without the knowledge of the
subpostmaster.
Q. Okay.
MR JUSTICE FRASER: Can I just ask for a point of
clarification because it has not come up on the
LiveNote.
I think in your answer you just said to Mr Green "He
was saying it was possible without the knowledge of the
subpostmaster". By "he", you mean Mr Roll?
A. Yes.
MR JUSTICE FRASER: Then you said "my opinion is" and
I thought you said different to that, but it didn't go
on the -- is that right or did you --
A. Yes, my opinion is different to that.
MR JUSTICE FRASER: Your opinion is different to what
Mr Roll said?
A. I'm not sure it contradicts what Mr Roll actually said
because Mr Roll said there was this counter 32 mechanism
which made it clear. My opinion is yes, that would have
made it clear, but there may be other mechanisms like
actually seeing the guy doing it.
MR JUSTICE FRASER: Okay.
MR GREEN: If we read down through this section of your
witness statement, you say:
"It seems to me that I require further factual
information before I can comment on this evidence.
Which 'specific person'? Under what circumstances? How
frequently? Until I have that information, it remains
possible in my view that any transaction which 'would
appear to the subpostmaster as though it had been
carried out through the counter in branch' might only be
a transaction that he had given his consent for, as the
'specific person' - and which had in effect been made on
his behalf."
So what you are saying is you are reading Mr Roll as
saying that sometimes they asked for somebody in the
branch to be logged on, yes?
A. Well, I'm not reading Mr Roll, I'm saying I want to know
more about what Mr Roll says.
Q. And if we go to paragraph 85, you say that:
"In his paragraphs 27 - 34, Mr Parker provides
detailed and specific commentary on Mr Roll's
paragraph 20, using his knowledge and the appropriate
contemporary documents, where they have been found.
Here he acknowledges that Fujitsu could insert
transactions into branches by a piggy back process. I
am not yet able to comment on Mr Parker's evidence or
the documents he cites."
Why were you able to comment on Mr Godeseth's
evidence and apparently establish a disputed fact, but,
when Mr Parker gave evidence favourable to the
claimants, unable to comment on his evidence?
A. It was a time issue. You know, did I have other
evidence to drill down, did I have time to drill down?
But Mr Parker's evidence went to the point of whether it
was possible and subsequently the experts have agreed
that it was very difficult to say anything was
impossible. Mr Parker's evidence goes to that. The
issue that we are discussing is whether it is possible
without the knowledge of the subpostmaster.
Q. What evidence would you have needed in front of your
eyes, beyond Mr Parker agreeing with Mr Roll, to be able
to say: on the basis of that, I accept that it was
possible to inject a transaction without the
subpostmaster knowing?
A. Well, that's a bit hypothetical. I mean, I can
conjecture that if Mr Parker had said "I agree piggy
backing was possible" and "I agree that the
subpostmaster would have known nothing about it", then
that would have done it.
Q. Pausing there, Dr Worden. Do you accept that by
refusing to comment on Mr Parker's evidence when he
agrees with Mr Roll, that contrasts very strikingly with
accepting Mr Godeseth's evidence as establishing a fact?
A. I was not refusing to comment on Mr Parker's evidence.
What I was saying was that Mr Parker and Mr Roll seemed
to agree that something was possible, so I'm not going
further on that. What I don't know about is whether
something was possible without the subpostmaster
knowing.
Q. Well, I suggest to you, Dr Worden, that it is bizarre to
deal with that evidence in that way.
A. I don't agree.
Q. Let's look at the joint statement at {D1/5/6}. Now,
that, 10.6, is where the subtopic of remote access is
being dealt with. Yes?
A. Yes.
Q. And Mr Coyne is dealing with it. Where do you say to
the court that you have reflected your view in the light
of Mr Roll's further evidence and Mr Parker's further
evidence in this agreed statement? Or didn't you?
A. I would need to -- let me pore through that joint
statement, but basically my recollection of the joint
statement is -- which one is it?
Q. Dr Worden, how about this: is that something you would
like to have a glance at over the luncheon adjournment?
Just in case there's any particular parts that come to
mind?
A. It might be useful, yes. But as I say, my opinion in my
second report was that I wasn't clear from the evidence
I had seen, so whether this could be done without the
subpostmaster's knowledge, and I would have required
more evidence from either Mr Roll or Mr Parker to make
myself clear on that. In the joint statement I do
recall that the experts agreed that you cannot say
anything was impossible, that, you know, it is difficult
for the experts to say at this distance that certain
things were not possible. However, that was a separate
issue from whether a thing could be done without the
subpostmaster's knowledge.
Q. Okay, let's just go back one last time to {D3/1/245}.
So even though you say that the experts were agreed that
they couldn't say things were impossible, it is the
precise effect of what you conclude at 1119, isn't it?
A. 1119 --
Q. You say:
"... Mr Roll could not have made these changes to
the branch accounts ..."?
A. Can we look at the agreement on the joint statement to
see how that meshes with 1119? I can't remember where
it is in the joint statement.
Q. Go back to {D1/5/6}. You have referred to
paragraph 1114 --
A. Sorry, I'm not there yet.
Q. If you look at the screen. Do you see 10.6?
A. That's Mr Coyne's statement.
Q. Yes, Mr Coyne said that. And there's a reference there
in the right-hand corner of that box to
paragraph 1114 --
A. Sorry, it is not clear to me who put in the reference.
If Mr Coyne made a statement I think probably what
happened is he put in the references.
Q. Okay. What I was keener on was perhaps you can have
a glance at the joint statement over lunch?
A. I will do, yes.
Q. And if there was somewhere where you thought the court
should fairly look to see any revision of the way your
views were expressed as at 4th March, then perhaps you
can direct the court to that.
A. Yes, sure.
Q. Now, when you are doing that, could we just go back,
please, to your second witness statement --
MR JUSTICE FRASER: Report.
MR GREEN: I'm so sorry. I apologise.
Dr Worden, your second report, second expert's
report, at {D3/6/21}.
A. Yes.
Q. Having said that the experts couldn't say anything was
impossible --
A. Well, I really would like to see the wording of that
joint statement.
Q. No, no, but I'm asking you about what you put in your
second report before the joint statement.
A. Yes.
Q. Okay. And in the your second report in paragraph 84 you
have actually -- this appears to be the thing -- one
thing that you are able to say is impossible. So you
say:
"Therefore --
A. I say:
"Mr Roll's ... evidence does not cause me to alter
the opinion ..."
Q. Yes, the one that we agreed was incorrect in approach
and conclusion.
A. We agree now, but at the time I wrote the second report,
on the basis of Mr Roll's evidence and Mr Parker's
evidence, I didn't see reason to change that opinion.
Q. I would suggest to you that you were inexplicably
reticent to accept something that was contrary to
Post Office's interests.
A. No, I was reticent. Not inexplicable, I was reticent
because I had not seen sufficient evidence to convince
me that these things could be done without the knowledge
of the subpostmaster.
Q. But Dr Worden, would you accept that the approach you
have taken there contrasts very, very strikingly in how
you approach Mr Roll's evidence with the approach you
took at 1119 in your first report when you accepted,
effectively, Mr Godeseth's evidence?
A. I have accepted that my approach at 1119, that use of
the word "established" was wrong and my approach was
wrong, and we have established that -- you know we have
done that before the interval --
Q. My question is: do you accept the contrast is very
striking?
A. I think the court will have to -- I accept that my
attempt to make my position clear that I'm not trying to
find findings of fact, I'm not trying to find one
witness or the other, I accept that on this occasion
I fell short of that.
Q. Do you mean only in your first report or also in your
second report?
A. Not in the second report. I believe the second report
was -- you know, I believe that the evidence from
Mr Parker and Mr Roll, and if we look again at Mr Roll's
paragraph 20, I believe that the question that I asked
in paragraph 83 that I want further evidence, I believe
that was a valid approach.
Q. When you say in paragraph 84 -- because we have dealt
with what you accept about your first report at
paragraph 1119 several times, haven't we?
A. Yes.
Q. So in your second report, you say, notwithstanding what
you have now accepted about 1119, in your second report
that you were right not to change that view.
A. I hadn't seen sufficient evidence to change that view.
Q. But do you say you were right not to change that view
when you revisited the entire piece in your second --
A. At the time I believe it was right not to change that
view because I hadn't seen evidence that convinced me
that this change could be made without the knowledge of
the subpostmaster. That's where I am.
Q. Dr Worden, do you accept that that betrays a complete
failure to appreciate the need to consider the situation
both on the basis of whether the claimants' evidence is
right as well as on the basis that the defendant's
evidence is right?
A. No, I don't accept that.
Q. Do you accept that you have failed to make obvious
observations on the basis of evidence of the claimants
with which one of the defendant's evidence has in fact
agreed?
A. I accepted what I accepted about the first report,
1119 --
Q. No, on the second report.
A. On the second report, no, I don't accept that.
I believe that Mr Roll's evidence about counter 32 and
sometimes we had to get somebody's permission and so on,
that left me in a position of doubt where I was not able
to change my opinion.
Q. How did you pursue any genuine doubts that prevented you
from saying anything helpful to the claimants about
that?
A. What do you mean pursue?
Q. Did you ask anyone at Fujitsu?
A. I mean, I had to write this report and this was the
position at the time of this report. I was not
expecting a third witness statement from Mr Roll
addressing that there might have been that, for
instance. I just wrote down what I felt was appropriate
to write down in the second report.
Q. Let's pause there. You identify your sort of, as it
were, standing on the island of not having enough
information to say anything, and the question is, what
did you do about it because it is obviously an important
issue? Did you ask for any further information
yourself?
A. I do not think I did, no.
Q. Because not only are you entitled to ask for further
information from those instructing you, but the courts
made a specific direction that you can apply for
directions, and you also have the ability to request
information formally, don't you?
A. Absolutely. But let me try and explain the sources of
my doubt. Counter 32 is clear cut. If counter 32 is
there, the subpostmaster sees it. But in my opinion,
there may be many other mechanisms whereby the
subpostmaster may observe that something has been done,
and that includes simply observing that somebody had
been in his branch and he had given his permission to do
something, and that was the area of doubt which was
raised by Mr Roll's statement.
Q. So pausing there, Dr Worden. Did you understand that
what was being agreed was that a Fujitsu person would
actually go into the branch? Is that what you think
this is all about? It is, isn't it?
A. This was my doubt, right, which person, what
circumstances. And if we look at Mr Roll's
paragraph 22, you know, we had to ask for a specific
person to log in and that means somebody at the counter
in the branch doing something specific.
Q. Because the position is that they can't use that unless
someone is logged in at the branch. And my question to
you is slightly different. Did you understand this
entire exchange as being about someone from Fujitsu
turning up at the branch and saying "Jeff, can you log
in, please"?
A. No.
Q. In the branch?
A. I didn't understand the entire exchange in full. It
left doubt in my mind that something had to be done at
the branch, somebody had to log in at the branch to make
this thing happen. I didn't understand the full detail
of that so I was left in doubt.
Q. Because you made references in your answers to seeing
someone actually make a change at the branch?
A. Well, the whole question I was asking consequent to
Mr Roll's witness statement is: what does he mean?
Q. But the entire issue is about remote access, not at the
counter, isn't it?
A. Remote access means not at the counter, but this seems
to have been facilitated by something at the counter.
Q. I understand, but it sounded as if you were saying you
understood a Fujitsu person would go into the branch and
be seen to make an alteration?
A. No, I didn't mean that.
Q. So if you have referred to that in the transcript we
should disregard that; that's not what you had in mind?
A. What I have referred to is that Mr Roll's paragraph 10
to 20 leaves me in doubt as to what actually happened.
Q. Let me go back to the original question from which we
have gone round a little bit. What steps did you take
to clarify the ambiguity which bore directly on one of
the identified Horizon Issues? Did you take any steps
at all?
A. I didn't take any further steps beyond this report.
Q. Could you tell his Lordship why not?
A. Priorities and, you know, whether I felt that further
investigation would get me further. It was kind of my
feeling that it was a difficult area and that -- put it
this way, sorry, it is very hard to explain this, but
there are levels of depth and complexity in the way
Horizon actually works which the experts have not been
able to plumb, if you like, and there is a whole lot of
detail about how a transaction might have been
identified. For instance, there is a PEAK that talks
about counter 11 and 12. That's not counter 32 but that
will maybe give an indication.
To my mind there was a kind of swamp of difficult
questions there and I was not going to -- I felt,
rightly or wrongly, going to make progress in that area.
Q. Even in circumstances where in your first report you had
concluded that Mr Roll was factually wrong?
A. I was taking the situation as at my second report, and
as I say, I wrote down what I believed I could conclude
at the time and I did not have much expectation that
I would be able to conclude more.
Q. Dr Worden, what is weird is you conclude very forcefully
one way in your first report and then you are not
prepared to reach a conclusion which would have been
correct the other way in your second?
A. I do not think that is weird. I'm trying to go with the
evidence and provide what I know to the court.
Q. But you are going only with the defendant's evidence in
report 1 and then when the defendant's evidence agrees
with the claimants', you stop going with anyone's
evidence?
A. No, I have to go where I think the evidence takes me and
I have to try and be balanced and neutral about it. And
sometimes I may fail to do that, I'm sorry, but I'm
trying to actually assess the evidence. And to my mind,
in my second report, what Mr Parker said and what
Mr Roll said did not sufficiently convince me that
I could write down, yes, it could be done without the
knowledge of the subpostmaster.
Q. What threshold of proof were you applying when you are
faced with the defendant's evidence, agreeing with the
claimants' and you are still unconvinced? Beyond
reasonable doubt? 99%?
A. It is not a question of threshold of proof, it is
a question of what opinions I can usefully offer the
court.
Q. But you had offered one and you are refusing to change
it in your second expert report. It is not that you
haven't offered one at all. Dr Worden, if we were there
we would all understand. You say too complicated, it is
absolutely impossible to plumb the depths of this
system, as you mentioned, I can't give an opinion one
way or the other. That is not what happened. You give
a very clear opinion which we've analysed with some
care --
A. And I've agreed it was overkill.
Q. Right. And then you refuse to change the overkill
opinion in any way because you say there's some doubt.
A. Well, we established today that my approach in my first
report was wrong.
Q. Yes. And you stick to it in your second report, that
the point --
A. No, I wrote in my second report -- in my second report
I wrote what the evidence persuaded me and convinced me
about, and that was I hadn't seen sufficient evidence
that this thing could be done without the knowledge of
the subpostmaster. I was trying to be balanced and fair
and that was what I was doing in my second report.
Q. So this exchange that has gone on a bit longer than
I expected, but this is helpful to the court in
understanding your approach to being balanced and fair,
what you have done here is a fair example, is it, of you
being balanced and fair throughout your reports? That
is right?
A. I don't see that I can comment to how this example
relates to the whole of my reports.
Q. Do you think you have been especially fairer elsewhere
in your reports? You have given this, you have
described this approach as balanced and fair --
A. I'm trying not to go beyond the evidence.
Q. Just focusing on the balance and fair point, let's not
go back to the evidence. You say that this is
an example of you being balanced and fair?
A. Yes.
Q. You are a statistician, yes? You understand what
a representative example means, don't you?
A. Yes.
Q. Is this a fairly representative example of you being
balanced and fair in your approach in your reports? It
is, isn't it?
A. I think this is a fair example of me trying to assess
the evidence in front of me and trying to draw
conclusions as far as I can and not trying to go beyond
the evidence, and that is fair, I believe.
Furthermore, if you look at the balance of my
reports, there are many, many occasions where I have
tried not only to be fair on the claimants but to be
biased in my numerical estimates in favour of the
claimants.
Q. We will come to those.
A. Absolutely.
Q. So, one final question and we will move on. You would
have no problem with the court using this as an example
of your approach overall?
A. What's written in the supplemental report, I have no
problem with the court using it as an example of my
approach.
Q. Even if it were a totemic example of your approach --
A. Totemic is not a word that occurs --
Q. A very good example of how you approached the evidence?
A. I would hope the court will take into account all sorts
of examples of my approach in assessing whether I've
been balanced or not.
Q. I understand. Let's move on.
You have given your evidence that you didn't follow
up your doubts on this point. More widely, have you
spoken to any of the Fujitsu witnesses in the course of
preparing your reports?
A. There was one early conversation.
Q. Who was that with?
A. With Gareth Jenkins, and it was a phone call with
lawyers present and I was trying to clarify --
Q. Well, if lawyers were present I'm not going to ask you
about the content, if it was in any way privileged. So
my learned friend will indicate if he has got any
problem with that?
MR JUSTICE FRASER: Was the conversation with Gareth Jenkins
on the phone as well?
A. It was on the phone.
MR GREEN: Okay. What about anyone else from Fujitsu?
Because there are lots of witnesses have been here, you
have all got a room, everyone has been going in and out,
you have had the opportunity to speak to people from
Fujitsu, haven't you?
A. Well, the defendant's lawyers set down a set of ground
rules for what --
Q. I don't want to ask you about what they told you, I just
want to ask you have you spoken to anyone from Fujitsu?
Mr Godeseth? Mr Parker?
A. We talk about the weather, yes. I mean --
Q. No, about the case?
A. About the case, no.
MR DE GARR ROBINSON: My Lord, I'm perfectly happy for my
learned friend to ask about the ground rules that the
defendant's lawyers set down if my learned friend
wishes to.
MR GREEN: And has anyone from Fujitsu provided any written
comments or observations on your report, because we
heard Mr Godeseth had had some comments which originated
from Mr Jenkins on one of his statements?
A. It is obvious that when I did analysis of various KELs,
Fujitsu did the same analysis and that came out in
a witness statement which I saw at the same time as
everybody else. That was the only instance.
Q. But they haven't actually provided any comments for you?
A. Not for me, that went into my reports.
Q. But you did know you were entitled to ask questions of
Fujitsu, didn't you?
A. Well, that always happened through the Post Office
lawyers.
Q. No, of course. And you have in fact asked questions
through Post Office lawyers of Fujitsu, haven't you?
A. I have done, yes.
Q. If we look, please, at {H/302/1}. It is a letter of
29th May, and if we go, please, to the second page
{H/302/2}, we can see halfway down:
"Question from Dr Worden."
A. Yes.
Q. And there are questions and Fujitsu responses?
A. Yes.
Q. You hadn't actually agreed those questions with
Mr Coyne, had you?
A. No.
Q. Was there a reason for that?
A. It was just my -- not curiosity, but there were various
questions in my mind that were unanswered to me from my
examination of the documents. I was not aware that
Mr Coyne was interested in the same questions.
Q. Okay. So you followed up things that you had some
uncertainties --
A. I found those documents puzzling in various places.
Q. So on this puzzling issue you did follow up by asking
questions?
A. Yes, because I felt answering the questions could be
productive, I felt there was a big chance that, you
know, I would get an answer which would actually resolve
that uncertainty.
Q. Okay. If we go back a page {H/302/1} to the first page
of this document, we can see that the date on the letter
was 29th May.
A. Yes.
Q. And can you remember roughly when you asked the
questions?
A. I can't, I'm afraid. I mean --
Q. Was it a few days -- because 29th May is the Wednesday
before we restarted the trial.
A. I think it was rather earlier than that. I mean, I have
read the low level design of the transaction correction
tool a considerable time earlier and I think my
questions about it had nothing to do with the trial
starting.
Q. Okay. On the face of it, let's look at {H/324/1}, on
3rd June in a reply to a letter from Freeths, it says:
"Dr Worden asked the questions which are set out in
our letter on 19 March 2019."
Does that sound about right?
A. Yes, that probably is ...
Q. If we just go back to {H/302/2}. Just under the heading
it says:
"Dr Worden has asked a number of questions ..."
A. Yes.
Q. It seems these are the ones you asked back in March,
yes? And it then goes on to say:
"We have taken instructions from Fujitsu ..."
Yes?
A. Yes.
Q. "... in respect of these questions and set out below
both the questions asked and responses given. These
responses are being disclosed to Dr Worden at the same
time as this letter is being sent to you."
So did you get the answers for the first time last
Wednesday?
A. I believe so.
Q. Okay.
A. Again, I'm not brilliant on these sort of dates and
chains of events, but I believe so. I mean --
Q. Dr Worden, can I ask you this. Mr Coyne made lots of
requests, didn't he?
A. Yes.
Q. Lots?
A. Quite a lot, yes.
Q. And although they were put in a form of a joint request
document, in fact he was making lots and lots of
requests for information which only he was making?
A. Yes.
Q. Was there a reason why you didn't support any of them?
A. Well, basically I think it became evident, and it is
evident now, that two experts took very different
approaches, and my approach was top down, understand the
architecture and work down through things like KEL. So
as far as I was concerned, I had plenty of information
to go on. And Mr Coyne's information requests didn't
strike me as things that, yes, I have got to really see
that because I was -- had a different priority. I was
trying to do a top down understanding of the
architecture and top down look at robustness, and so on
and so forth.
And so I had plenty of documents to look at,
basically. So I think it is the different approaches
taken by the two experts that led to lack of overlap.
Q. That's your explanation for not having supported any
requests?
A. I do not think I supported any requests --
Q. Sorry, not having supported the very many requests he
had to make?
A. Yes, I felt his interests were different from mine and
I really had plenty to do.
Q. And you initially only looked at KELs, didn't you, you
didn't actually look at PEAKs?
A. No, I looked at KELs and I looked at PEAKs where they
were relevant. I felt that KELs were a more distilled
form of information. I felt they were sufficient in
many ways, especially when you go and look at the PEAKs.
Q. You looked at one or two PEAKs referred to in the KELs
you looked at?
A. Yes. Look at some of them where you feel the KEL
doesn't tell you enough.
Q. And Mr Coyne was looking at PEAKs and KELs?
A. I think in the first report he was looking mainly at
KELs, like me, and in his second report he turned on
PEAKs.
Q. And --
A. For instance, in my report Mr Coyne had commented on
nine PEAKs in his first report and I commented on those
in my first report as well.
Q. I understand. You heard Mr Coyne was asked whether
anyone provided documents for him to look at?
A. Mm.
Q. Ie ones that he hadn't found or his assistants hadn't
found. Were you provided with any documents, particular
documents by anyone to look at, have a look at this?
A. Very early on there was a tranche of about 75
architecture documents that were put to me, but I think
they were in the disclosure and that was when I was
doing my high level early exploration. Other than that,
I think the Post Office lawyers have been really trying
to get a very level playing field.
Q. Other than that?
A. Mm.
Q. So when we look at {C5/9/2}, it is a document of 30 May.
Just to give you the chronology, 30 May is the Thursday
before we start the trial, 31 is the Friday. Sorry,
2018, I do apologise. This is 30 May 2018. I was
confusing it with another document. This is 30 May 2018
when there was a question about whether or not Mr Coyne
had had the same documents as you had in early 2018.
Yes?
And if we look on page {C5/9/2}, the answer is in
the final paragraph, three lines up from the bottom:
"Dr Worden commenced work in around February 2018
and has had broadly the same information and documents
as Mr Coyne."
Yes?
A. Yes.
Q. Mr Coyne had actually asked you in an email, hadn't he?
If we go to {F/1792.2/1}, do you remember Mr Coyne asked
you in an email had you:
"... had access to additional resources that [he]
had not had access to? ... any document repositories
other than those formally disclosed? Or access to PO or
Fujitsu staff who had imparted any knowledge?"
A. I'm afraid my knowledge of the precise chronology a year
ago is not very -- now we have got it --
MR JUSTICE FRASER: I do not think it can be F.
MR GREEN: It is {F/1792.2/1}. You might need to refresh.
MR JUSTICE FRASER: That's 1792.1.
MR GREEN: We will come back to it. If you had been sure
that Mr Coyne had had the same access that you had, you
would be able to answer that straightaway, wouldn't you?
A. I was just trying to get on with the job of drilling
into this stuff.
Q. Fair enough.
A. And I'm not a good witness on precise blow-by-blow who
said what.
Q. Dr Worden, don't worry. Let's move on. Let's look
at --
MR JUSTICE FRASER: On the basis it is 12.58 am and you have
asked the witness to look at something over the short
adjournment, I think we ought to stop now.
Mr de Garr Robinson made effectively an offer about
the ground rules which I don't intend to say anything,
other than if you are going to ask any questions about
it and there is a document, it is probably best to make
sure that there are copies rather than pose it as
a memory test. But it is completely in your ballpark.
MR GREEN: My Lord, yes.
MR JUSTICE FRASER: And do you just want to remind the
witness what it is you asked him to do over the short
adjournment?
MR GREEN: Yes.
Dr Worden, I think you were very kindly going to
look at the joint report.
A. JS4?
Q. JS4. Have a look there to look at where you feel you
have commented helpfully or as you felt appropriate in
relation to that remote access issue.
MR JUSTICE FRASER: Now, Dr Worden, just to be helpful to
you, do you have a hard copy there of the joint
statement?
A. Yes, I do.
MR JUSTICE FRASER: You have, all right. Because it is
probably asking too much of a witness to do it on
screen.
A. No.
MR JUSTICE FRASER: But if you have got a hard copy it is
much easier.
We will come back at 2 o'clock. Thank you all very
much.
(1.00 pm)
(The short adjournment)
(2.00 pm)
MR GREEN: My Lord, I have already shown my learned friend
this. This is Mr Coyne's mobile telephone on which
a missed call from Dr Worden is recorded at 1.40 pm.
Can I just show your Lordship so your Lordship has
seen it.
MR JUSTICE FRASER: Yes.
MR GREEN: Dr Worden, you understand you are not supposed to
talk to anyone --
A. Yes, I noticed -- this was a pocket call. I noticed my
phone was ringing somebody -- I didn't even notice what
the number was -- and I stopped it.
Q. I understand. You will understand why --
A. Absolutely, of course.
Q. Now --
MR JUSTICE FRASER: I think the modern expression for that
is a bum call; is that right?
A. I understand it is.
MR JUSTICE FRASER: I think pocket call is a much better way
of putting it.
MR GREEN: Dr Worden, the ground rules you were given you
mentioned before lunch, could you just explain what they
were?
A. Basically that conversations between me and Fujitsu
about the case had to happen with a lawyer present who
would intervene if anything crossed any boundary.
That's basically it.
Q. I understand. So you weren't able to call to Fujitsu
without a lawyer being there?
A. Yes.
Q. To see what was being discussed. And can we get up,
please, on the Opus system -- I think it is
{F/1792.3/1}. I think I may have thought it was 1792.2.
This was the email I was referring to. This is May 2018
halfway down.
A. Yes.
Q. You didn't respond.
A. I don't recall the email. I will accept that I didn't
respond.
Q. Okay. Can I ask you this: when you were given the
ground rules, were you given the ground rules before or
after you spoke to Mr Jenkins with Womble Bond Dickinson
on the line?
A. No, the ground rules were for the trial.
Q. For the whole trial?
A. Yes.
Q. Can I just ask you whether you can remember whether you
were given the ground rules before you spoke to
Mr Jenkins or after?
A. No, Mr Jenkins was a year ago.
Q. Yes. But the ground rules were for the trial only?
A. Yes.
Q. So they didn't apply before the trial started?
A. No, but conversation before the trial was Mr Jenkins'
phone call, nothing else.
Q. I understand. So there was no prohibition on you
talking directly to Fujitsu people prior to the ground
rules --
A. No, that's not correct. The position was I wanted
a clarification in my understanding of the receipts
payment mismatch, and a phone call was set up with WB
present. Very shortly after that, and that was
May 2018, Mr de Garr Robinson said we have to be
absolutely whiter than white about this, no direct
contact with Fujitsu at all, and so everything was
through Post Office lawyers --
Q. Thereafter.
A. Yes.
Q. I understand.
Can we now look quickly at the transcript, please,
today's transcript, at page {Day18/9:1}. Do you have
page 9 at the top?
A. Yes, I have it.
Q. And it might not be quite the right page. You said
earlier in your evidence that you had had plenty of
information to look at and you were taking a top down
approach. Do you remember that?
A. Yes, I remember that.
Q. And you had looked at the sort of high level documents
of architecture --
A. Looked at a lot of architecture documents, yes.
Q. Can I ask you to look, please, at {F/1611/1} and we will
look at the first page to see what it is. This is
a Post Office board agenda from 31 January 2017. Do you
see that?
A. Yes.
Q. If we go to page 100 of that document, please
{F/1611/100}. Now, this is in the context of looking at
the Horizon architecture, and can I take you back to
page {F/1611/87} first, please. Just look at the --
this is a "Technology Strategy Update"?
A. Yes.
Q. I just want to take all these documents reasonably
briefly. The first question is, is this the sort of
type of document that you would have had regard to when
compiling your report, or not?
A. Well, it is certainly not the type of document I was
interested in in the early stages of putting together
what I called the foundation sections of my report.
Q. What about later stages?
A. Well, I would like at these documents typically in
response to seeing references in Mr Coyne's report.
Q. If we look at the second line, at the context there?
A. Context?
Q. Do you see under "Context"?
A. "IT not fit for purpose", that is right.
Q. It is fair to say, isn't it, that where Post Office
themselves are recognising the limitations to their own
system, that is likely to be a fair and sensible view?
A. This is Post Office talking about the whole IT estate.
Q. Yes.
A. And they put their phrase "not fit for purpose" in there
and they were obviously discontent, yes.
Q. And if we look, if we may, please, at page {F/1611/98}
A. Sorry, could I check the year of this?
Q. 2017.
A. 2017, right. Okay.
Q. If we look at page 98 you will see paragraph 25 down at
the bottom.
A. "There are tensions in each contract." Yes. (Pause)
Q. You see Fujitsu:
"... a 6 year fixed contract signed with PO which
continues to invest in legacy and obsolescence where FJs
own strategy globally is to move to cloud."
Do you see that?
A. Yes.
Q. So that's their own perception of the system and the
situation they are in?
A. That's PO's perception of Fujitsu's strategy, yes.
Q. Yes. If we look at page {F/1611/100}, please. This has
got senior people like Tim Parker, the chairman, Angela
Van Den Bogard, also attending, various others. On
page 100 if we look at paragraph 35, for example, do you
see there "For Retail"?
A. Yes.
Q. They are specifically talking about HNG-X, which is the
Horizon system?
A. Right.
Q. So it is not the general IT environment, is it?
A. No, that is Horizon specific --
Q. It says:
"The Horizon (HNG-X) platform is at the end of its
life and needs replacing."
Yes?
A. Yes.
Q. That's fair comment, isn't it?
A. It is interesting, that. I mean, 20 years is a long
life and that's what they were saying, yes. Personally
I think it is doing quite well, but there we are.
Q. "Previous attempts to move away from HNG-X platform,
specifically with IBM, have been unsuccessful."
A. Yes, I didn't know about that.
Q. Did you see this document at all, can you remember?
A. I have seen reference to this end of life quotation at
some time, but I haven't read this document in detail.
Q. You didn't follow this one up?
A. No.
Q. Let's look, please, at {F/1603/5}. So that document was
31 January 2017, this is 17 January 2017.
MR JUSTICE FRASER: I think we are at 1603, but did you want
to be at 1603 --
MR GREEN: Yes, 1603, this is 17 January 2017. If you look
at "Credence", do you see that?
A. Yes. (reads to self)
Q. Do you see "Credence" underlined?
A. Yes, I'm just reading that paragraph (Pause).
Q. Do you see where it says Accenture picked up a difficult
pass?
A. Yes, and additional costs at the moment.
Q. Starting to see light at the end of the tunnel. Then
Fujitsu. Do you see that heading?
A. Yes. Not hold the power.
Q. "FJ see the contract as a cash cow, so need to persuade
them that working with POL to migrate to cloud
technology is to their benefit against a 'too good'
contract."
Did you see this document at all?
A. I didn't read these pieces -- this page.
Q. Looking at it now, does that chime with a sense of
anything you have seen?
A. Well, that Fujitsu paragraph, it sort of chimes because
I have always had the impression from the governance
structure and the documentation and so on that Fujitsu
were not short of budget, really.
Q. They weren't, okay. And at least not short of budget
coming from Post Office?
A. That is right.
Q. If we go forward, please, at {F/1586/3}. Perhaps you
should see the first page, November 28, 2016. Then if
we look at page 3 as it comes up?
A. That's interesting.
Q. Yes?
A. Okay.
Q. If we look at the first column:
"Horizon software was developed in 1996, originally
as a DSS IT project."
Yes?
A. Yes.
Q. There was a heavy part of it which was DSS, but also for
the Post Office is what happened, isn't it? You know
about that?
A. I hadn't gone into the detail of that prior project but
I know it existed.
Q. DSS wanted a big system and Post Office went to be
a joint procurer. Then the DSS dropped out and the
Post Office was left holding the contract?
A. I'm just aware of that pre-history but have not gone
into it.
Q. "At the time Horizon was one of the first electronic
Points of Sale.
"Horizon was created before the internet had any
real effect on Retail or Banking.
"It was built as a 'closed' system & designed based
on paper processes, is clumsy & operator unfriendly."
Now, were you aware of that recognition, either in
this document or anywhere else, that Post Office
internally recognised that it was operator unfriendly?
A. Well, this is a document in 2016 talking about 2001.
Q. No, I'm just asking you a clear question, Dr Worden. It
says it is clumsy and operator unfriendly. Were you
aware from this document or from any other document that
that was Post Office's internal view of Horizon?
A. I was not aware from more contemporary documents and
I had not read this one.
Q. Thank you. Did you say from more contemporary
documents?
A. No, I mean I do not think I have seen a document before
2000 -- I mean, the case for moving the HNG-X included
"Let's get rid of Escher", and so on. It didn't include
"God, the interface is clumsy".
Q. Okay, so you weren't aware of that?
A. I wasn't aware of that --
Q. -- in forming your views you have expressed?
A. No.
Q. Let's look at the third column, if we may, "IBM deal
ends". And the only bit I would like you to look at
really there is halfway down, third bullet point:
"However, whilst modernising the 'front end' is
relatively straightforward, as the project developed we
realised that shifting the 'back-end' is extremely
difficult."
A. Yes.
Q. So they reverted back to Fujitsu.
A. Yes.
Q. And then if we look on the right, "HNGA":
"Whilst HNGA runs on updated Windows software,
fundamentally its architecture is the same as HNG-X."
That is fair, isn't it?
A. Yes.
Q. "This means that it remains a 'closed' and inflexible
platform that retains the complexity of transaction
journeys and operational requirements - not Simpler to
Run!"
A. Yes.
Q. I think it appears to follow from not having read this
document that you may not have appreciated that when you
were putting your reports together?
A. No. But if I may comment briefly on the shifting the
back end extremely difficult, that seems to me to chime
with my understanding that the bulk of the investment
actually was in the back end and that's where the
complexity was.
Q. And the front end was the bit the subpostmaster was
using?
A. Yes, and I would guess, and it is a bit of a guess,
that, you know, in terms of lines of code, complexity
and so on and so forth, the back end is more than the
front --
Q. So more had been invested in the back end than in the
front end that the subpostmasters had been using?
A. I feel broadly the back end is more complex.
Q. {F/1557/1}, please. This is a month earlier than the
previous one, 22 October 2016.
MR JUSTICE FRASER: Just before you go there, do you know
what the expression "thin client" -- how would you thin
client --
A. Thin client generally nowadays corresponds to a client
that's just a web browser and it goes through to some
internet site. I think that's pretty much the meaning
now.
MR JUSTICE FRASER: That's perfect.
Mr Green.
MR GREEN: This is a document from October 2016, and on
page 1 under "Context" would you look at paragraph 3:
"Our back office"?
A. Struggles, yes.
Q. "Our back office also struggles with the complications
of dealing differently with each of our many clients,
heavily manual processes ..."
Do you see that?
A. Yes.
Q. "... reconciling disparate sources of data,
retrospectively financial controls and a lack of
flexibility."
A. Yes, and this is referring to the Post Office bank.
Q. Yes. And it says:
"This backlog of challenges, poor support contracts
and a lack of skills have led to a prohibitive cost of
change preventing the improvements that should occur as
a part of business as usual."
Do you see that?
A. Yes.
Q. Now, first of all, did you broadly pick up that
impression from anything else or not?
A. There were various other documents, like finance roadmap
and so on, which conveyed a general impression to me
that the Post Office back office in its IT and its use
of POLSAP and various different SAPs and so on, and the
business processes around it, that they had to interface
with the clients, but that back office process was more
complex than the Horizon back office.
Q. I understand the complexity point. I'm talking about
the satisfactoriness point. Did you get that
impression?
A. Well, obviously from these documents Post Office was
unsatisfied about something.
Q. You hadn't picked that up yourself?
A. Well, I picked it up when I started looking at the
documents cited by Mr Coyne.
Q. Right. Did you expressly deal with that anywhere?
A. Well, it seemed to me not really directly part of
Horizon.
Q. Okay. I do not think you have dealt with these other
documents we have been to already, have you?
A. None are very familiar.
Q. Okay.
A. But as I say, I did start looking at these Post Office
IT strategy documents when Mr Coyne cited them.
Q. Let's go a bit earlier, 29th August 2016. {F/1522/1},
please. Let's look at paragraph 1. Can you look at the
right-hand side three lines down:
"The Back Office process and applications remain
complex, unreliable, expensive to maintain and not
suitable for today's business."
A. Yes.
Q. Did you look at that document?
A. Again, I can't recall looking directly at this document.
I'm not brilliant on looking at this --
Q. I understand.
A. -- etc, but it doesn't surprise me.
Q. And it doesn't surprise you that they mention that it is
unreliable?
A. This is the Post Office back office --
Q. Yes.
A. -- things like POLSAP and all sorts of stuff --
Q. All things they --
A. Much more than that, the whole of Post Office business.
Q. Yes, the whole thing, and that didn't surprise you that
it was unreliable?
A. They say unreliable. It didn't surprise me it was
unwieldy.
Q. You said this didn't surprise you a moment ago. You are
just rowing back on that a bit. Bit worried it is
damaging to Post Office to say so?
A. Unreliable, good question. Well, they say that and that
seems to be part of the package, yes.
Q. You said it is a good question, the reason why I'm
asking is reliability of the overall process at least on
one view is a question which may be of relevance in this
trial, isn't it? Including reliability of the back
office aspects. Is that fair?
A. Well, there has been a bit of prohibition about looking
at Post Office business processes.
Q. Who prohibited you?
A. Well, people said that's out of -- I got a flavour that
Post Office business processes have been out of scope.
For instance in TCs, the process for creating TCs has
not been a thing the experts have looked at.
Q. You sort of said people said and then you said you got
a flavour?
A. Yes.
Q. Can I ask you, did you form a view yourself about
whether that was out of scope or were you told not to
look at it?
A. I was told that it was out of scope, and I found some
difficulty with that because in a sense things like
robustness of Horizon actually depends on all sorts of
things and one has to try and assess the consequences of
certain things, where the causes, how business processes
work and so on, are out of scope. And it is a bit of
a blurry boundary.
Q. We will come back to it later, but you refer to
transaction corrections and transaction acknowledgements
and things as very important countermeasures in your
report?
A. That is right, yes.
Q. So they were central to your analysis?
A. They are essential to my analysis.
Q. And they rely on Post Office back office systems that
we've been talking about?
A. The approach I took to that is transaction corrections
are a corrective measure, they are a countermeasure. It
is obvious they work a very large proportion of the
time, and I tried to work out numerically the small
proportion when they didn't work.
Q. We're going to get to that in some detail, but for
today's purposes I'm only asking about relevance. So
where we are is you were told it was out of scope. You
had difficulties with that because you felt that that
was an important countermeasure --
A. Well, I think I reconciled it to myself in that certain
causes were out of scope but the effects were in scope.
Q. I understand.
A. Does that make sense?
Q. We will explore whether it makes sense or not. Against
that background you hadn't had any particular regard to
a document we see at 1522?
A. Well, the view I have taken is that the Post Office
whole IT estate is much bigger than Horizon.
Q. I understand. But it is right, just to be clear, that
the Post Office back office estate, yes -- back office
systems that we have been talking about just now --
A. Yes.
Q. -- are relevant to the issue of data integrity,
transaction corrections, transaction acknowledgements
and those things?
A. Indeed they are. As, you could say the same thing,
certain client IT systems are relevant.
Q. Absolutely. So if you had concerns about data integrity
coming in from Camelot on National Lottery, that would
be relevant as well?
A. Yes.
Q. Okay. Let's look at {F/555/1}, please. Let's go a bit
earlier. Let's go before 2010. My screen has gone
completely --
MR JUSTICE FRASER: It is because the F folder is so big.
Give it a second, I daresay it will appear in a moment.
MR GREEN: Thank you very much.
So this is a document from December 2009,
7th December 2009, and if we could go to page
{F/555/10}, please. This is the 2009 appraisal,
internally:
"Horizon - Current State."
It says, as at 2009:
"13 year old design and technology to satisfy
a different business."
A. Yes.
Q. "Evolved rather than designed"?
A. Where does it say that?
Q. Third bullet point.
A. I see. I was looking at the footnotes, sorry.
Q. Just look at the big points, the ones they thought were
really important.
A. (Reads to self) Yes.
Q. And "slow and expensive to use". Do you see that in the
middle?
A. Which year was this assessment?
Q. This was 2009.
A. And that's what they are saying there, yes.
Q. Had you looked at this document?
A. Again, I do not think I have seen this actual document.
Q. Well, let's look at the bottom four lines, please:
"Horizon is also a system that is wrapped up in
'barbed wire' - making changes difficult and costly -
test everything!"
Do you see that?
A. Sorry, where is "test everything"?
Q. If you look at the bottom four lines of the text, at the
bottom of the page, four lines up from the bottom --
A. (Reads to self) Yes, I must admit this is different
from my own understanding of Horizon.
Q. This is different from your own understanding?
A. Yes.
Q. Okay. And:
"Design was optimised at the time to minimise costs
..."
Do you see that, at the bottom?
A. "... ([especially] network) - offline -- what's their
reference to offline working?
Q. ?i think it's probably because internet access was more
expensive in those days.
A. The question then is what was it at the time: was it
original Horizon or HNG-X? This is before HNG-X, so
this is the original design. Right, okay.
Q. Original design. But you fairly accepted that that's
different to the view of it you formed in your reports?
A. Yes.
Q. And you probably hadn't seen this document?
A. I hadn't, no.
Q. Can we look now at two individual documents relating to
individual subpostmasters which we know about because
they were lead claimants in the November trial.
Let's look, please, at {F/68/1}. You will see that
this is an email and I would like to take you over the
page to page {F/68/2} where the chain begins, if I may,
please. You will see there that it is from
Frank Manning. Do you see halfway down?
A. From Frank Manning, yes.
Q. See that? And we look up two lines, it is to Sue Lock?
A. Yes.
Q. The subject is "Horizon matters --
A. Yes.
Q. -- Barkham SPSO"?
A. Yes.
Q. And it explains in the third paragraph:
"The balances are a mess ..."
Do you see that?
A. "Balances are a mess", yes.
Q. "... (in pre-Horizon times - the Postmistress virtually
achieved a clean balance every week) ..."
Do you see that?
A. Yes.
Q. Do you see "but" in bold?
"... I worry that something like 25 re-boots in one
day is having an effect overall."
A. Yes.
Q. And if you look at the second paragraph on that email
you will see he says:
"I visited there today & was too scared to accept a
cup of tea in case the Horizon system crashed cos the
electricity supply is still a live (excuse the pun)
issue."
A. Yes.
Q. Now look at the bottom big paragraph, before "help
please", the one that starts "Need your best offices".
Do you see that?
A. "Need your best offices", yes.
Q. "Need your best offices to get this case to a proper
solution - she keeps getting promises of attention - but
nothing is actually being done now to clear up the
problem. It is Horizon related - the problems have only
arisen since install & the postmistress is now barking &
rightly so in my view."
A. Yes.
Q. "Help please."
Now, it is fair to assume, isn't it, that where
Post Office internally recognised that Horizon was
likely to blame, or in this case positively asserts that
it definitely is, it is quite likely to be right?
A. Well, I don't know who Frank Manning was or what kind of
evaluation he had done. I mean, it seems he has made
a visit to the site.
Q. Indeed. Do you think --
A. And on that basis he had seen that apparently there are
lots of power cuts.
Q. 25 reboots in one day. Do you think that would have
an effect to --
A. Well, it seems -- I'm just guessing because, you know,
I haven't studied this evidence in detail, but it seems
to me electricity supply is something to do with the
reboot.
Q. Yes. But do you accept it is quite possible that
a problem with the electricity supply causing 25 reboots
might well be the source of problems in her accounts and
balances on the basis of this document, or have you got
some other explanation that you would like to --
A. "Balances are a mess ..."
Q. You can't diagnose it at this distance in time?
A. No.
Q. I'm only putting to you do you think it may be fair that
that may well be right?
A. It is certainly possible.
Q. But is it fair to accept that it may well be right on
the face of the document?
A. I accept it may well be right.
Q. Let's go back to page {F/68/1}, please. This is to
Kevin Cox from Sue Lock. It says:
"Frank came to see me about this office and we
discussed it with Sanjay and said that she needed to
prove that it was Horizon that was causing all these
power failures in the office."
A. Yes.
Q. "Can you tell us please how we can now get this resolved
as it appears now it is a direct consequence of the
installation and not anything that has happened in
steady state."
A. Yes, that what it says.
Q. It is fair to assume that may well be right?
A. Well, it is a puzzling paragraph because Horizon causing
power failures --
Q. I'm not suggesting Horizon is causing the power failure.
A. That's what this paragraph says.
Q. I'm suggesting that power failures and lots of
recoveries are quite likely to cause balances to be in
a mess at this branch, aren't they?
A. All I'm saying is that the phrase "Horizon was causing
these power failures" indicates to me some possible
confusion about cause and effect.
Q. You don't know this, Dr Worden, but what in fact
happened is this had been all installed as the system.
A. My general view would be that Horizon has been built to
be robust against power failures, and therefore the
prior expectation is that there can be a number of power
failures which do not cause discrepancies in accounts.
Q. So you would not have expected that outcome on the basis
of your understanding of the Horizon system --
A. My understanding is that power cuts happen approximately
once per branch per month, I believe, or was it once per
branch per year. I'm sorry about that, but they do
happen and Horizon has to be robust against them.
Q. And what we have seen is not consistent with your
understanding of its robustness?
A. That is right.
Q. If we look, please, at another document, last example,
{F/99.1/1}. Now, this is an audit in 2001 of Mr Bates'
branch. You probably read his name. He is the lead
claimant.
A. I think I know what he looks like.
Q. Bates v Post Office?
A. I think he is here, actually.
Q. Yes. This is his branch. If we could go forward,
please, to page {F/99.1/4}, if you go down to "Cash
Management" heading, do you see that? Halfway down?
A. Yes.
Q. Go down to the next heading "Control Gaps - High Risk".
A. Yes.
Q. The cash holdings were 25% higher than the ideal target
holding.
A. Yes.
Q. And then underneath "Comments":
"A correct assessment of cash holdings could not be
made because the Horizon system intermittently adds the
previous days cash holdings to the daily declaration."
A. So that is what the auditor is saying?
Q. Yes. The auditors are very familiar with the system,
aren't they?
A. I am sure they are.
Q. And that is also not consistent with your understanding
of how Horizon works, or should work?
A. I haven't seen, I think, other evidence about adding
previous day's cash --
Q. I'm not asking whether you have seen other evidence, I'm
asking is that consistent or inconsistent with your
impression of Horizon's robustness or reliability?
A. That is inconsistent.
Q. Can we now move on, please, to documents recently
disclosed, so new. So the Friday before the hearing was
31st May, Thursday was the 30th and the Wednesday was
the 29th.
A. Fine.
Q. So let's start with documents closed on Friday, 31st May
or uploaded to the bundle on the 31st. I will just show
you.
A. It is uploaded to the bundle we are talking about.
Q. It is in the bundle. {F/1834.14/1}
A. I have tried to look at some of the things recently in
the bundle. I haven't done it all.
Q. Dr Worden, we just want to know what you have looked at,
what you haven't and what you made of it, that's all.
So {F/1834.14/1}.
A. Yes.
Q. This document as a document was first disclosed on
31st May, but it contains information from the MSC
spreadsheets which have been disclosed a few months
prior.
A. Right, okay. Yes.
Q. Now, have you seen this document before?
A. No, I haven't.
Q. Do you have any idea why it has been uploaded?
A. No, I don't.
Q. Can I show you just how it works. If you look at the
top of this one, you have the MSC number in the middle?
A. Yes, middle of the page.
Q. There is the heading, 043J -- I probably should have
said -- 043J0348236. This then explains:
"Below is the extracted text relating to," this MSC,
"taken from Rows 284594 to 284654," in one of the three
documents to which one had to put together the MSC logs?
A. Yes.
Q. Did you find the MSC logs in their original form easy to
use?
A. I didn't find them too problematic. They were clumsy.
They are monster spreadsheets, but what I found is one
of them, I think it is this one that the extract is
from, was fairly narrative and you could go to
a specific MSC just using an Excel search and you could
read the narrative. And, indeed, I actually built
a JAVA tool to filter the MSCs and get a smaller
spreadsheets of things I wanted to --
Q. Okay. Let's look at this one. We can see what's being
done because the third paragraph says:
"The proposed change is to insert 1 dummy
Transaction Acknowledgement (TA) row for Branch 74005
into the Branch Database (via an SQL INSERT statement)
with transaction amount value zero and a quantity of 1."
Yes?
A. Sorry, I'm not there at the moment.
Q. It is the third paragraph at the top.
A. I see --
Q. Third paragraph --
A. (Reads to self)
Q. Okay?
A. Yes.
Q. Then do you see "Justification for the change and
urgency"?
A. Yes.
Q. "POL have requested this issue be resolved," and there
is a POL incident reference, Q17628223, yes?
A. Yes.
Q. And from there we can see the text about it, and if we
go over the page {F/1834.14/2}, at the top of the page:
"POL approval required (Y or N) ..."
Do you see that?
A. Yes.
Q. And it says:
"No," so POL approval isn't required, "- POL aware
..."
And gives the same reference that we have seen on
the previous page?
A. Yes.
Q. So what we learn from this MSC on a fair reading is that
POL have asked for something to be sorted out, this was
how it was going to be done, POL was aware, Post Office
was aware and so they didn't need any approval because
they were already aware, they had asked them to do it.
Yes?
And we can see the other questions that follow on
that page. Go forward one page, please, {F/1834.14/3}.
You see:
"SSC (managed by Steve Parker)."
At the top?
A. Yes.
Q. We go over the page again {F/1834.14/4}, you have to
note here at the top it says:
"Below is the extracted text relating to MSC
043J0348236 taken from Rows 121273 ..."
A. Yes, this is another one of the spreadsheets.
Q. Yes. So what they have done is they have put the rest
of the information about the MSC on --
A. On the same MSC number.
Q. On the same MSC number. So they are right to do that,
aren't they?
A. Yes, absolutely. Absolutely.
Q. That is the only right way to do it?
A. That is the correct way to do it.
Q. And you can see expected impact, testing, security
implications:
"Do users need to be informed of the change?
"No."
Etc.
Do you see that?
A. Yes. And this was inserting a TA?
Q. Yes.
Then let's look on page {F/1834.14/5}, please. It
says:
"The information relating to [this MSC] cannot be
easily extracted from," the complete data spreadsheet?
A. Well, there are three of them. Are we talking about
different ones of the three now?
Q. Yes. So they have two in there, but not the third one
because it is too difficult to get out.
Now, if we now compare that one with one of the
others also that was provided on Friday, 31st May in
a similar format, {F/1834.12/1}, the same principle. It
explains the text below, yes?
A. Yes.
Q. "Below is the extracted text". Top line:
"Below is the extracted text ..."
Do you see that?
A. Yes, that is the one that I have looked at --
Q. So you looked at this one, have you?
A. No, not this particular MSC, but I'm saying the source
of my information on MSCs is mainly the narrative one.
Q. Let's look at what the non-technical overview says. It
says on the right-hand side:
"This causes discrepancies in Stock declarations
report."
Do you see that?
A. Yes.
Q. So:
"Older unused branch declarations are being picked
up in Branch Declarations Report in the same trading
period ... [or] balance period ... (BP) from previous
year occurs again."
A. Yes.
Q. So it is picking those up and causing discrepancies in
the stock declarations report?
A. Yes.
Q. At the bottom under the second line of asterisks you
will see:
"Steps to carry out the fix."
A. Yes.
Q. And it is:
"Logon to BDB Host Node 1 as 'oracle' user."
A. Yes.
Q. Then there and over the page we see the scripts being
used {F/1834.12/2}.
A. Sorry, can I just read that? (Pause)
Q. I'm not going to ask you about the scripts, I just want
to show you aspects of the document.
A. Yes.
Q. If you look at the bottom --
MR JUSTICE FRASER: Of? Bottom of page 2.
MR GREEN: On the bottom of page 2. There is a mandatory
field:
"POL approval required (Yes or no)."
A. Yes.
Q. We see there "no" on the next page {F/1834.12/3}.
A. I see, on the next page, right.
Q. So it looks as if there POL approval was said not to be
needed?
A. Yes.
Q. And if we look further down we can see there is
a "delete SQL" being applied?
A. Yes.
Q. And this is effectively to try and remedy the stock
discrepancies problems, isn't it?
A. It is remedying a problem in BRDB, yes.
Q. And we go over the page to page {F/1834.12/4}, please.
And it says:
"Who will action this change ..." at the bottom.
Then go over the page to page {F/1834.12/5}, "RMGA
UNIX"?
A. Yes.
Q. So that is the end of that one. So we don't have there
the data from the other spreadsheet that we did have on
the previous one we looked at. Do you remember the
first one had two spreadsheets output?
A. Yes.
Q. Now, what we now need to do is we need to look in the
spreadsheet itself at {F/1843/1}.
A. Can I get it on this screen?
MR GREEN: It will come up.
MR JUSTICE FRASER: It might take a moment, but it will
appear currently.
A. Previously I didn't get spreadsheets on those screens at
the back there for some reason.
MR GREEN: What we are going to do, Dr Worden, is we are
going to look on the second spreadsheet that we haven't
got reproduced within the Womble Bond Dickinson document
and just see what's in there.
Because what we are looking for is on row 90474 --
don't scroll, please. Could you go into the A1 thing
and type "A90474"?
A. There we go.
Q. There we go. And scroll down a tiny bit, if you would,
so we can look below that?
A. I see, yes. So this is all the impact stuff?
Q. This is all the impact stuff which is not included in
the document prepared by Womble Dickinson.
You see there 1:
"What is the expected impact on the live/production
service of implementing this change?
"None.
"2. What level of testing ...
"N/A.
"3. What are the security implications ...?
"None.
"4. Do users need to be informed ...?
"No."
Regression path, and then:
"8. Has the change been implemented before? If yes
give details of the previous implementation impact and
success criteria:
"N/A.
"9. Taking into account all of the above what are
the worst case risks implementing the change?"
You see that?
A. I have read through. I am not quite in sync with you
but I have read through that case.
Q. Number 4 is:
"Do users need to be informed of the change," and
the answer is "no".
A. Yes.
Q. Now, if we go, please -- and close that spreadsheet --
to {F/812.1/1}.
Now, Dr Worden, this is a PEAK when it comes up. It
is {F/812.1/1}.
A. So this is 2011. Is it referring to the same MSC?
Q. It is referring to the MSC, but I'm going to -- because
it is just about to be 2.50 and it is a new document,
I'm going to ask his Lordship if that's a convenient
moment to have a break and perhaps you can have a glance
at it if you want to.
A. I mean, his Lordship said something about paper copies.
Now, would it be possible to have paper copies of PEAKs
or is that a bit difficult?
MR JUSTICE FRASER: Even my paper copies of PEAKs have
disappeared, although I'm expecting them back at some
point. But I am sure --
MR GREEN: My Lord, I do not think it is going to be easy
for him to look at various pages.
MR JUSTICE FRASER: How many pages is this PEAK?
MR GREEN: It is reasonably long. It is eight pages long.
MR JUSTICE FRASER: Does anyone on your side of the court
have an unmarked copy of the eight pages?
MR GREEN: No, I don't think so.
MR JUSTICE FRASER: So what is it you would like the witness
to do?
MR GREEN: If he could look at page 2 and just read that,
that would be helpful.
MR JUSTICE FRASER: Over the 10 minutes or now?
MR GREEN: Over the 10 minutes if that's not --
MR JUSTICE FRASER: Would you prefer for us to break for
10 minutes to that you can look at it carefully, or
would you -- it really is not an issue one way or the
other, it is whichever you find easier to do.
A. If it is convenient for a break now, it is nice to look
at one. But I suspect there will be loads of PEAKs
where I won't get a break.
MR JUSTICE FRASER: At the moment I'm unclear whether that
is the case or not.
MR GREEN: Shall we break now, my Lord?
MR JUSTICE FRASER: No, we will. I'm just looking at
logistics if this is going to be replicated tomorrow or
the day after.
MR GREEN: My Lord, yes.
MR JUSTICE FRASER: It is just some people are used to
working on screen, some people are used to working on --
MR GREEN: I understand.
A. And I have a particular way of looking at PEAKs.
MR JUSTICE FRASER: That is entirely understood, Dr Worden.
So after today, counsel is going to make sure that he
has got a hard copy for you.
A. Excellent, thank you.
MR JUSTICE FRASER: Who knows, by then I might even have my
hard copy file back, so I might be able to do it both on
screen and hard copy.
For the moment we will break until 3 o'clock for the
shorthand writers and for you also to have a look at it.
If, when you have read that page, you want to look
through subsequent pages, just ask either of the counsel
who will be standing there on the claimants' side of the
court and they will ask the Opus people and they will
flick it forward on the screen.
Mr de Garr Robinson, does this seem like a sensible
way forward?
MR DE GARR ROBINSON: My Lord, yes.
MR JUSTICE FRASER: All right, I will come back in at
3 o'clock.
(2.51 pm)
(A short break)
(3.00 pm)
MR GREEN: Dr Worden, so if we can just start on the first
page of that PEAK, please, if you go back a page
{F/812.1}.
A. First page, okay.
Q. We have the PEAK number at the top, which is PC0211010?
A. Yes.
Q. And at the bottom of the page under the "Impact
Statement", it says user "unknown", 23rd June 2011. Do
you see that?
A. 23rd June 2011. Is that right?
Q. User unknown, date 23rd June 2011. Then the problem is
explained:
"Branches will be forced to declare stock when they
don't want to. Apparent reappearance of withdrawn stock
may cause spurious discrepancies."
Now, Dr Worden, to give you context we have only
found this document by trying to trace through what was
happening on that MSC.
A. Yes. Could I ask about the linkage to the MSC?
Q. Yes, I will show you. It is page {F/812.1/3}.
A. If we go back to the MSC and the beginning of the MSC,
I think --
Q. Can I take it in stages and just show you --
A. Well --
Q. Page 3 at the bottom, do you see 24th June 2011?
A. Yes.
Q. Last yellow box?
A. "Reference Added", that is the last page of the PEAK.
Q. That is the last page of the PEAK.
A. And it refers to that MSC.
Q. It refers to that MSC.
A. Fine.
MR JUSTICE FRASER: I think it is on the first page as well,
isn't it?
MR GREEN: I think it is also.
MR JUSTICE FRASER: At the top.
MR GREEN: Yes, it is MSC on the first line under the
references value. Do you see that?
A. Yes, and that is the MSC we just looked at?
Q. That is the MSC we just looked at, okay?
A. That's my puzzlement, actually.
Q. But I'm just asking you --
A. We will come to my puzzlement when we do.
Q. Okay. For the moment you can see that the MSC refers to
this PEAK?
A. The PEAK refers to the MSC, I will check --
Q. Sorry, you are quite right. The PEAK refers to the MSC.
A. Yes.
Q. And we have seen what the problem is: spurious
discrepancies at the bottom?
A. That's the PEAK problem, yes.
Q. Now, pausing there, had you identified this document in
your researches?
A. Well, I can't say I have seen this PEAK before, but when
you go to page 2 in the Anne Chambers piece on page 2,
it is one of these recurring every year when the same
trading period comes around effects, which is just like
the suspense account book.
Q. This is a slightly different one, but --
A. It is similar.
Q. But pausing there, I was asking you had you seen this
PEAK.
A. I do not think I had, no.
Q. Right. If we go over the page to page {F/812.1/2}, you
will see in the yellow box --
A. I don't have colours.
Q. Second box down:
"PC0208335," which we won't look at but for his
Lordship's note is {F/773/1}, "addresses the need to
remove old branch declarations from the BRDB so they are
not picked up and reused when a stock unit reaches the
same TP/BP a year on."
Do you see that?
A. Yes.
Q. So there had been a previous PEAK about it?
A. Yes.
Q. Then she says:
"I have now found that there are some old
declarations which were created but never updated, which
are not removed by the SQL used for both the MSC and the
code fix."
Do you see that? We will look at what that means
and what it is referring to in a moment:
"In particular there are 36 stock declarations which
will almost certainly give branches problems when
balancing between now and the end of 2011."
A. Yes.
Q. And you would agree, wouldn't you, that's not how
Horizon is supposed to work?
A. Yes.
Q. The options discussed are:
"Can we either:
"a) change the fix already in the pipeline ...
(I know this is probably not possible) ...
"b) schedule and produce a fix which also targets
these extra declarations ..."
Do you see that?
Then the next paragraph, bottom of that box:
"If the MSC can be run by 1st July, it should not be
necessary to contact POL or any branches to get them to
take action to remove the declarations themselves (since
we confirmed to POL earlier this year that we had
already taken steps to prevent further instances, this
would be a good idea)."
A. Yes.
Q. So what they are trying to do is get it done quickly so
that the fact that the previous fix didn't fully work
doesn't show up?
A. I think that is right.
Q. That's fair, isn't it?
A. Yes. I think that the reference to the previous MSC
wasn't done quite right and it left some bits over.
Q. Yes. So the previous fix left some bits outstanding?
A. Yes.
Q. Because it wasn't done quite right. Post Office Limited
thought it had been fixed and we know that the problem
first occurred, if we go back to page 1, in
February 2011?
A. Right, okay.
Q. Look at the bottom of the page, first occurred 2011?
A. Yes.
Q. Go forward again {F/812.1/2} and "updated" at the bottom
of that page, Anne Chambers says:
"Apparent reappearance of withdrawn stock may cause
spurious discrepancies."
That is the phrase we see throughout.
Then if we go over the page, please, to page
{F/812.1/3}, it says there at the second box down:
15:07:16 20 "23rd June 2011 at 17:06:26 User: John Simpkins
"POL are aware of the declarations issue which was
not completely fixed by the previous release, I do not
yet think that they realise that fix was not complete so
a quick resolution to this would be good."
A. Yes.
Q. So what he is saying is: they think we fixed it last
time, if we fix it quickly they may not realise it?
A. That's pretty much what was on the previous page as
well, I think.
Q. Yes, and that tends to confirm it?
A. Yes.
Q. So if we go to page {F/812.1/6}, please, and we look at
the top box:
"RISKS (of releasing or not releasing proposed fix):
"What live problems will there be if we do not issue
this fix?
"Incorrect branch declarations belonging to a
previous year will be picked up and cause stock account
discrepancies to a live trading branch."
Do you see that?
A. Yes.
Q. So the position we have now is that Post Office doesn't
know the previous fix hasn't worked?
A. Yes, Fujitsu had told Post Office we fixed it.
Q. We fixed it --
A. And they weren't quite right, so they are going to tidy
up before it happens.
Q. So the NBSC haven't been told the fix didn't work?
A. Whether it is Post Office or --NBSC would be part of the
chain somehow, T.hey won't know.
Q. And they are going to try and fix it before Post Office
realised it?
A. They are going to try and head it off at the pass before
these branches run into it.
Q. So neither the users nor Post Office were informed about
this fix being done on the face of these documents?
A. That is right. What they are saying is we haven't
completely sorted it, we can sort it before any damage
is done.
Q. That's what they hope to achieve.
A. Yes, they are hoping to do that.
Can I mention my puzzlement about the MSC, or is
that --
Q. By all means, if it is helpful to the court?
A. The MSC, when I read it I thought it was to do with
a transacting acknowledgement for zero pounds, and that
seems to me to be a different view from this.
Q. No, there were two separate MSCs. The first one
I showed you that was very fulsome explaining PO
approval not needed, yes? The fulsome one was for TA
zero pounds?
A. I was just confused.
Q. I understand. The second one didn't have all the PO
approval detail from the second spreadsheet.
A. Right, okay.
Q. So we had to go and find it in row whatever 100,000, dig
it out, then search for the PEAK and, through that, find
out what actually happened, which I have just put to you
and you agreed with.
A. So the second MSC didn't actually say what was going to
be done. I was confused between the two.
Q. I understand. Don't worry.
A. Okay.
Q. Dr Worden, can we move now, please, to the approach you
have taken to Horizon Issue 1. Can I ask you, please,
to look at page {C1/1/1}. It will come up on the
screen. There's the top of the page. There is
a definition of the Horizon system. Yes?
A. System? Yes, right.
Q. "The Horizon system ..."
A. Okay.
Q. "... shall for the purposes of this list of issues mean
the Horizon computer system hardware and software,
communications equipment in branch and central data
centres where records of transactions made in branch
were processed," as defined in the pleadings, yes?
A. Yes.
Q. And that doesn't include the manual processes of
deciding whether to issue TCs or not, does it?
A. Right. That is correct.
Q. When we look at Issue 1, the question that's posed is:
"To what extent was it possible or likely for bugs,
errors or defects," any of those, "of the nature
alleged" and so forth "to have the potential to (a)
cause apparent or alleged discrepancies or
shortfalls ..."
A. Yes.
Q. And it doesn't say "to cause" without being first
qualified by "to have the potential", does it? The
words "to have the potential" are important, aren't
they?
A. What's the distinction between having the potential and
actually doing? I mean, if it doesn't actually do, then
in some sense it didn't have the potential.
Q. Well, let's put it this way: something might have the
potential to do something which you can identify, but
you can't identify whether it has actually done it;
that's fair?
A. Yes, that's fair.
Q. So whether it has actually done it and whether it has
the potential to do it are two different things in that
respect?
A. In a sense the word "potential" reflects a lack of
knowledge. If I knew more I would know whether it is
going to happen or not, but I know less so it is
potential.
Q. Yes. As you fairly accepted earlier, that the Horizon
system by its nature has been quite difficult to plumb
for the experts?
A. To plumb in that sense, yes.
Q. Indeed, you refer to a swamp and so forth. I mean, it's
not the most transparent task you've been engaged in?
A. It is a very big system to understand in a year, yes.
Q. So just coming back to this, what we can say with
confidence, if something actually does cause
a discrepancy in a branch --
A. Then it did had the potential.
Q. -- then it plainly had the potential?
A. Yes.
Q. So potential is necessary but not sufficient to have
an actual impact?
A. Yes, I agree with that.
Q. So it says to have the potential, and then there's two
parts to it: cause apparent or alleged discrepancies or
shortfalls relating to subpostmaster's branch accounts
or transactions?
A. Yes.
Q. Not necessarily in their branch accounts, but relating
to --
A. Ah --
Q. -- branch accounts, bare branch accounts or
transactions?
A. Discrepancies relating to branch accounts. That could
be interpreted very broadly.
MR JUSTICE FRASER: You are both talking at the same time,
I'm afraid.
A. Sorry.
MR JUSTICE FRASER: Go on, Dr Worden.
A. I'm saying discrepancies relating to branch accounts
could be interpreted very broadly and it might be taken
to include all back end discrepancies because they
relate to accounts.
MR GREEN: Okay. (b):
"... undermine the reliability of Horizon accurately
to process and record transactions ..."
Just an example of (b), if a scanner doesn't
correctly scan in a pouch that undermines the
reliability of Horizon accurately to process and record
the transaction, doesn't it?
A. It does, and for better or worse I took my role to be
concerned with lasting effects on branch accounts rather
than transient ones.
Q. Well, you did two things, didn't you? You focused on
actual impact rather than potential?
A. I tried to derive an upper limit on actual impact.
Q. And you added in lasting?
A. I added in lasting.
Q. And in fact we see where you define the issue in your
report, you also focussed on the impact on claimants'
accounts?
A. Yes.
Q. That was your analysis?
A. That was what I felt would be useful to calculate.
Q. But you saw in Issue 1 that it said subpostmaster's
branch accounts?
A. Yes, I agree.
Q. So you took your own course in your own sense?
A. Well, yes, it is in two stages really. The first stage
is I felt extent in Issue 1 required numbers, and for
a second stage I needed some calibration of those
numbers, and those are two steps we can look at in
return, if you like.
Q. In all of this you were still clear that in the Horizon
system, transaction corrections, the manual process of
reconciliation and issuing TCs was not in?
A. The manual process, the cause was not in, yes, but the
effect was.
Q. I see. The helpline was not included in the system,
was it?
A. First line support was not included but the experts did
look at evidence about it.
Q. To have understood a bit about -- we have touched on
this before -- about the processes and the scope for
errors, you need to know a bit about the helpline if you
are going to look at whether people get through and how
often?
A. That is correct.
Q. Did you know whether people on the helpline were reading
out scripts when people rang up about things?
A. Which helpline are we referring to?
Q. NBSC helpline.
A. That is the business helpline.
Q. The first port of call, as it were. Did you know if
they were reading out scripts --
A. I'm sure they had prompts of various kinds.
Q. But you didn't know, was the answer?
A. I didn't know because I hadn't been asked to look in
detail at the MSC.
Q. You heard the helpline referred to as the "hell line"
at all?
A. I didn't see that phrase.
Q. You didn't. Can we look at {F/1257.1/1}, please. This
is a forum post in which Mr Tank had made a post. If we
look at page {F/1257.1/6}, please?
A. Can I ask what forum?
Q. It was a subpostmaster's forum.
A. Internal Post Office thing.
Q. No, a subpostmaster's forum that they used to use.
A. Right.
Q. Okay. {F/1257.1/6}, if you look at the second box down.
A. It is very dim on my screen.
Q. The second box down says:
"When I balanced on 17th September I had
an unexplained loss of £176.74. I paid it in rather
than go through the stress of the hell line. It has not
come to light."
A. Okay.
Q. It is only a yes or no answer to identify what
information you have had regard to --
A. Right, I haven't read this before.
Q. You have not seen this before?
A. No.
Q. Had you seen any other factual evidence about people
deciding whether to pursue things through the helpline
or not of this sort?
A. Well, what I have particularly looked at is the helpline
logs for one or two weeks. Now, that's not directly
about the decisions made by the postmaster.
Q. If we look at {D3/1/201}, which is paragraph 904
Just look at 904 at the bottom. The last two lines,
couple of lines, you say:
" ... although I have no knowledge of the business
processes for creating TCs used by the central
Post Office departments."
A. Yes.
Q. And that is correct isn't it, you don't?
MR JUSTICE FRASER: Is that the correct word?
A. I said yes.
MR JUSTICE FRASER: I'm sorry, it didn't come up.
MR GREEN: Now, we mentioned that you placed quite a lot of
emphasis on transaction corrections.
A. Yes.
Q. And essentially what you say is where there are
discrepancies in the branch accounts that have to be
corrected, the transaction corrections process is what
then makes that happen?
A. It is a very important part of the correction, yes.
Q. And it is very important for your analysis as well,
isn't it?
A. It goes to my analysis of robustness. It is an example
of user error correction, which I think is quite
an important one.
Q. And because you placed a lot of emphasis on it for
robustness, if the Horizon system -- it is a question
for the judge, but if the Horizon system was as it
appears in the definition at the beginning and didn't
include the transaction correction process itself, you
would have to answer the question, well, for every time
a TC is issued, that may be because a discrepancy has
arisen in branch accounts that needs to be corrected and
that may be due to the Horizon system?
A. Sorry, I haven't quite understood the question.
Q. Your answer to the robustness of Horizon is heavily
dependent upon TCs and, you say, countermeasures which
effectively reflect that process, and you effectively
say that those are important because that's how
discrepancies are corrected?
A. Yes. I'm trying to analyse robustness and the
correctness of errors in asking how effective are the
countermeasures in countering the effects of errors
often when I don't know the causes. It might be
something out in the client's system, or whatever.
Q. And it is right, isn't it, that the answer to the
robustness question on your analysis would be very
different depending on whether you included the
corrective impact of TCs or not?
A. Absolutely. Yes.
Q. Fundamentally different, in fact?
A. Yes, very different.
Q. It is right, isn't it, that you and Mr Coyne took
slightly different approaches to the Horizon Issue 1 and
what it meant, because we see, if we look at joint
report 2 at {D1/2/28} at 1.9, it says:
"The experts have differing views on 'branch
impact'. Mr Coyne refers to any discrepancy that caused
a loss (or gain) within branch accounts that needed
corrective action as an 'impact to branch accounts'.
Dr Worden only considers an effect or impact on branch
accounts where a discrepancy loss (or gain) was not
rectified by a correction such as a Transaction
Correction."
A. Yes.
Q. There was a fundamental difference about that and then
there was also a difference about transient and lasting
inaccuracies in branch accounts, wasn't there?
A. Well, I think that is the same difference, really. "Not
rectified" means lasting, whereas "rectified" means
transient.
Q. Well, we will come back to that point later, Dr Worden,
but that rather depends how long the TC process takes,
doesn't it?
A. It does, absolutely.
Q. So if the TC process takes a long time, then you might
say something was lasting, even notwithstanding years
later it might be corrected?
A. That is not how I understood lasting --
Q. How did you understand lasting?
A. There were delays in the TC process which might be due
to client organisations or might be due to all sorts of
things, and they could be at the outside, I believe,
several months.
My definition of lasting did not depend on TCs
coming in within a certain timeframe. If TC never came
in, that would be lasting, but if TC took several months
to come in that is not what I would call a lasting
effect. A lasting effect is permanent; it is at the end
of the day, you know, he has lost money forever.
Q. So in fact, 1.9 and 1.10 in the report {D1/2/29} are
really directed at the same thing.
At 1.10 what you say -- this is your comment:
"... transient inaccuracies in branch accounts which
needed some form of correction, have arisen so
frequently and from so many causes that to list them is
not useful; and that evidence of each correction being
carried out is unlikely to persist to this day."
That is your position?
A. Yes, that's my view.
Q. Now --
MR JUSTICE FRASER: Just so I'm clear, the example you just
gave to counsel of a transaction that took several
months --
A. A TC.
MR JUSTICE FRASER: I beg your pardon, a TC, a transaction
correction, would that be a transient inaccuracy?
A. That would be a transient inaccuracy, yes.
MR JUSTICE FRASER: That's very useful.
A. The branch accounts would look wrong for that period.
MR GREEN: So now that was one big tectonic difference of
approach between you and Mr Coyne.
A. Yes.
Q. The other one was the identification of particular bugs
and their impacts?
A. We ended up in different places on what we thought were
actual bugs, yes.
Q. I understand. It is right, isn't it, that in seeking to
reach some compromise given the differences of approach,
you managed to agree what we find at 1.15?
A. Yes.
Q. And that was the context in which that sort of
compromise agreement was reached?
A. Yes.
Q. Now, can we look please at page 9 of your first
statement at {D3/1/9}. You describe section 28, at
paragraph 29 of that page. Yes, you describe what
section 8 in your report is about.
A. Yes.
Q. And you describe section 8 as addressing Horizon 1, yes?
A. Yes.
Q. "... the extent to which bugs in Horizon may have
affected the Claimants' branch accounts"?
A. Yes, I've stuck in the word "claimants" there.
Q. You stuck the word claimants in because that's what you
in fact do in section 8?
A. It is mainly what I do in section 8, but I could come to
that.
Q. Now, can we look at how the reasoning in section 8 works
so we understand it. You'll appreciate, Dr Worden, that
there's a difference between the parties on what the
right approach might be?
A. Absolutely.
Q. But I am not going to proceed from here on to analyse
your approach on the basis of how you did it. I'm not
suggesting I agree with your approach, but I'm trying to
understand how you have done it in case you turn out to
be right. Do you understand?
A. I might do.
Q. And it is really to focus on the methodology you have
adopted.
A. Yes, right.
Q. You have made a number of assumptions in section 8 of
your first report, haven't you?
A. Yes, I have, to arrive at a number I made some
assumptions.
Q. It is meant to be an easy question. You are meant to go
yes and then I ask you the next bit.
A. Yes.
Q. Let's look at a couple of them, please. {D3/1/176},
paragraph 769.
A. 769?
Q. Yes.
A. "... would affect all branches ..."
Yes.
Q. Yes. So one of the assumptions you have used for your
model is that all branches, whether claimants or
non-claimants, are effectively --
A. Yes.
Q. There is another assumption if we look at {D3/1/148} at
paragraphs 621 to 622, and I'm going to suggest to you
it is basically the same point underlying it because on
the premise that it affects, there is an even
distribution of bugs, you have then considered a scaling
factor, haven't you?
A. Yes.
Q. And you specifically say at 622:
"It seems implausible to me that there is some
special factor about Claimants' branches, which makes
them more prone to --"
A. Much more prone.
Q. Quite right:
" ... much more prone to bugs which one would expect
to strike any branch at random."
Then you give quite careful consideration to this,
don't you, because it is important, in your report?
A. I have thought about it. It is in the appendix,
I think.
Q. Absolutely, so let's trace it through:
"Nevertheless, I have considered the possibility
carefully in Appendix F."
You say {D3/1/149}:
"I have shown that that there is no significant
difference between Claimants' branches and other
branches, in proneness to bugs in Horizon."
Pausing there. Bearing in mind what I asked you
about this morning, when you say "I have shown there is
no significant difference between Claimants' branches
and other branches", would it be fair to say that when
you say "shown" you mean thought about it, can't think
of a reason and therefore concluded that was the case?
A. I think that's fair, yes.
Q. That's what we see in appendix F.
If we look at {D3/2/207}, paragraph 430. Let's look
at the start of 429, if we may, on page 207. In your
appendix F, 429, you say:
"I have tried to think of possible differences of
this sort, and I have only been able to find one
candidate difference."
So you wracked your brains and the only thing you
could think of is then in paragraph 430?
A. That is right.
Q. And what you come up with, you say:
"It might be said that Claimants tend to make more
errors than non-Claimants ..."
Yes?
A. Yes.
Q. " ... and that these human errors particularly trigger
bugs in Horizon. Or it might be said that certain bugs
in Horizon are successfully handled by non-Claimants,
but tend to cause Claimants to make errors, which cause
losses."
Then you discount both of those?
A. Yes, I say -- the argument that follows is that they are
both second order effects.
Q. I understand.
Now, you heard Mrs Van Den Bogard give evidence, or
you may have heard her, where she accepted that
Post Office may suffer from UEB or user error bias
sometimes?
A. I didn't hear that evidence, but --
Q. Do you think that might be a fair suggestion against you
there?
A. I don't see how that's relevant to this consideration.
Q. You don't think -- you have approached this on the basis
that the claimants are probably making more mistakes?
A. I put that up as a possible hypothesis --
Q. As the only thing you could think of?
A. As the only thing I could think of.
Q. I understand. You then, if we go back very kindly to
your first report at {D3/1/149}, you have proceeded on
the basis of paragraph 623 that claimants' branches are
smaller and have fewer transactions?
A. Yes.
Q. And therefore they are less likely to be hit by
a Horizon bug in a given month?
A. That's the assumption.
Q. Have I fairly summarised those key planks of your
reasoning?
A. I think that's all fair.
Q. Your calculations then follow from those assumptions?
A. Yes.
Q. You reach conclusions which are likely if not calculated
to suggest that the claimants' claims are so implausible
as to be wholly unrealistic, don't you?
A. They are not calculated to suggest -- I did the numbers
as what comes out.
Q. What you conclude is that the claimants' claims are so
implausible as to be wholly unrealistic --
A. I conclude that the maximum amount of shortfall in
claimants' branches which could arise from bugs in
Horizon is approximately 0.5% of their claimed
shortfalls. That is one way of stating the number
I come up with.
Q. So through the prism of looking at the claimants
specifically rather than subpostmasters in general, you
have effectively reached an opinion that it is absurdly
unlikely that their claims can be right; is that fair?
A. Yes. Could I add a little bit? And that is I have
taken two approaches to Horizon Issue 1, one of which is
that I think numbers are important to give meaning to
phrases like "likely" and "risk" and so forth, and you
have to define the scope of those numbers. Because if
you say it is likely in the whole lifetime --
Q. I understand what you have done. We will come to the
detail.
A. But I'm coming to the second stage. The second stage,
I have said I think it is useful for the court to
calibrate those numbers with respect to claimants, but
I may be wrong in that. And if I'm wrong in that, then
the court can convert my numbers from a claimant basis
to an old postmaster basis. I have given the conversion
factors in my report, so if I'm wrong about emphasis on
claimants the court can convert the numbers to something
else.
Q. But Dr Worden, I'm going to follow through what you have
done.
A. Yes.
Q. And ask you at each stage whether you think it is right.
A. Yes.
Q. Or wrong.
A. Fine.
Q. Now, I would like, if you would just bear with me while
we go through a relatively simply statistical exercise,
if that is all right?
A. Yes, fine.
Q. And I would like you first to consider how one would
assess the chances of meeting somebody called Penny
Black in the UK. If you bear with me.
A. Yes.
Q. There are 60 million people in the UK of whom, as far as
we can tell, roughly 24,000 have the surname Black?
A. Yes.
Q. And of those, assume half of them are men and half of
them are women, so that is 12,000. Let's assume 1% of
all women with the surname Black are called Penny. So
1%, 120 people, are called Penny Black in the UK, let's
assume.
That is a consequence of that --
A. Sorry, we started with 60 million? Let me just track
the numbers.
Q. 24,000 are called Black.
A. So that is a factor of 3,000. And then the next factor?
Q. Half of them are men and half are women.
A. So that is a factor of 6,000.
Q. And then 1% of them.
A. So that is 600,000.
Q. In fact, on our calculation 500,000 --
A. Assuming that surnames and Christian names are not
correlated.
Q. Yes, exactly. Assuming they are independent variables.
We get 120 people out of 60 million, which is 1 in every
500,000. I'm not going to criticise your mental maths
in the witness box, just to get in the ballpark
together.
So that is 1 in every 500,000 people. Now --
MR JUSTICE FRASER: 1 in every 500,000 is a lady called
Penny Black.
MR GREEN: A lady called Penny Black.
MR JUSTICE FRASER: It is a different way of saying divide
60 million by 120.
MR GREEN: Exactly. And if you are at a dinner with 50
other people, 50 other females, let's say 50 other
people to start with.
A. So 25 females, let's say.
Q. Let's say 50 other people, right?
A. Yes.
Q. What are the chances of one person there being called
Penny Black? It is 50 over --
A. You mean exactly one or more than one?
Q. Let's say at least one.
A. At least one. I think it is the basic probability of 1
person being Penny Black times 50. I believe it is
that.
Q. Yes, exactly. So that is 50 over 500,000, isn't it?
A. Yes, 1 in 10,000.
Q. 1 in 10,000. What are the chance of two people being
called Penny Black?
A. That's approximately the square unless there is
a correlation of relatively -- if you are in a room full
of relatives, it is very different. But without the
correlation, I believe, let me think about this
carefully, I believe it is the square of the
probability.
Q. Yes. I mean, in fact there is a very fine point that
you have taken one person out of the UK and one person
out of -- but roughly it is the square?
A. Yes.
Q. Yes. So the chances of two people being called Penny
Black is 1 in 100 million, isn't it?
A. Two people at this dinner.
Q. It is one in 10,000 times 10,000. 10,000 times
10,000 --
A. I believe we are all right so far.
Q. It is not meant to be controversial. Just taking you
through it so we are agreed on methodology. And the
chances of three people being called Penny Black,
I promise you the last example, is 1 in 10,000x1 in
10,000x1 in 10,000, which is around 1 in a trillion,
yes?
A. I'm just trying to think about correlations.
Q. As independent variables. Don't get complicated. Just
as independent variables, that is the answer?
A. As long as we are in that assumption.
Q. Yes, we are, because it is terribly important, isn't it,
yes?
A. Well, obviously I don't know where this analogy is
going --
MR JUSTICE FRASER: I don't even think you need to worry
about that.
A. Good.
MR JUSTICE FRASER: But so far you are agreeing with
Mr Green's basic approach to --
A. Yes, it seems the probability theory is right,
basically. Yes.
MR GREEN: You may not know this, but in 2015 there was the
175th anniversary of the Penny Black stamp and
Royal Mail had an event to celebrate it at which they
invited people called Penny Black to attend.
A. Yes.
Q. Let's say 50 of those people who they invited to join
the dinner because they were called Penny Black came.
A. Yes.
Q. If you were at that dinner, what are the chances of the
other 50 people being called Penny Black?
A. It is obviously high, but we have violated the
assumption of no correlation.
Q. Yes, it is a different example?
A. It is absolutely, very different.
Q. So 50 people who are in fact called Penny Black have
been invited to join for dinner?
A. Yes. I mean --
Q. So the chances are likely to be at or about 100%, aren't
they?
A. Yes. I should say generally that probability theory is
what one uses in the absence of specific knowledge like
you have just put to me, and that specific knowledge
changes the whole ball game.
Q. The specific knowledge changes the whole ball game,
doesn't it, Dr Worden?
A. It does, absolutely.
Q. And you would accept that if we say, for the sake of
argument, the chances of the 50 people at the dinner,
you can imagine the chap coming out with the 50 credit
cards, "Penny Black anyone?" All the hands go up, yes?
MR JUSTICE FRASER: Because they are all being asked to pay
for the dinner they have been invited to.
MR GREEN: Your Lordship had the point, which made me
hesitant to give the example. But the point is that if
you look at the 100%, yes, on the one hand --
A. Yes.
Q. -- if we say approximately 100% chance of them -- there
might be somebody pretending to be called Penny Black
who is there interloping, but basically 100%, they've
all come because they've been invited because they are
called Penny Black, and one in a trillion for just three
of them, there isn't any bigger margin of probability
that we could illustrate by changing the assumptions.
A. Absolutely, and this shows that specific knowledge
overrides probability theory, when you have that
specific knowledge.
Q. It is essential. Do you know how GLOs work, Dr Worden?
A. GLO?
Q. Yes, that is a group litigation order?
A. Sorry, no, I'm -- I haven't looked into that, really.
Q. They are advertised for people to join. Let's look at
{C7/3/39}. That's the group litigation order. And at
page {C7/3/39}, this is the invitation/publication
notice to join the GLO.
The GLO is limited to people who satisfy those
criteria. Do you see that?
A. Yes.
Q. See what the third one is?
A. They claim to have suffered loss, yes.
Q. It is a pretty important specific bit of knowledge,
isn't, it Dr Worden?
A. I do not see how it fundamentally alters my calculation.
Perhaps you would like to spell that out to me.
Q. Well, you gave an example of people from the same family
being there, that would change everything. I gave you a
counter example of people being invited to join because
they are actually called Penny Black. This is
a situation in which the entire universe of claimants
has only joined the litigation at all because they
believe that they have suffered the very type of losses
we are concerned with.
A. They believe that, yes, but I'm asked as an expert to
examine that from a neutral point of view. I'm not
asked to take the claimants' case or the Post Office's
case. I'm asked to examine on my knowledge of Horizon
what the probability of certain events was.
Q. Okay, let's assume this: if the claimants were right
that they had been affected by such losses, yes, it
would be no surprise at all if they're all here,
would it?
A. I'm still at a loss to understand how that relates to my
analysis.
Q. Let me ask you this, Dr Worden: this is something
special about the claimants that you were searching
about in your quest, isn't it? They are the people,
they are the subsample of people who have sought to join
the group litigation?
A. But it would seem to me that I would then have to base
my analysis on an assumption that the claimants' claim
was true.
Q. Let's look at it the other way round. You would have
a factor to displace your assumption that people who
believed they have been adversely affected in this way
are equally likely to have been adversely affected as
people who believe they haven't been?
A. I don't see how making the assumption that the
claimants' case is true is an unbiased approach,
a balanced approach to assessing Horizon.
Q. Let's take that out of the picture.
A. No, this is my duty, that I have to make a balanced
assumption of the probability of those events not
assuming that either side's case is right when looking
at Horizon.
Q. But Dr Worden, you have reached that assessment by
starting, I would suggest to you, from a demonstrably
false premise that there is no special factor to
distinguish claimants from SPMs generally?
A. No, I have started from the assumption that a bug in
Horizon does not pick on claimants more than other
people. That is a different assumption.
Q. But in making that assumption you have ignored the
obviously relevant point that the definition of
a claimant is somebody who believes they have been so
affected?
A. I have not ignored that role in the point. I'm finding
great difficulty in following how this analogy relates
to the assumption I have made about how bugs in Horizon
act. That is the assumption. Not how claimants act.
Q. I will only ask you one more question about this before
I move to a different aspect of your report. Why would
someone who felt they had not been affected join the
group action?
A. I think somebody who felt they were not affected would
not join the group action. I do not see that as
relevant to the technical assumptions I made in my
section 8.
Q. Let's look at your scaling factor and your maths because
that's not a matter of opinion, is it? The advantage of
maths is that it is right or wrong?
A. That's very much the advantage from the court's point of
view and the expert's point of view.
Q. Let's look at the paragraph of your report Mr Coyne was
taken to at {D3/6/30}.
Just go back to page {D3/6/29} to give you context.
This is in the section 5.1 "Size of Claimants'
branches".
A. Yes.
Q. So this is the second of the assumptions I put to you
earlier, yes? You have assumed even distribution which
we have dealt with.
A. Even distribution of what?
Q. Of bugs across all subpostmasters, we have dealt with
that?
A. Bugs per transaction in Horizon.
Q. Yes. And you are now turning to your scaling factor, as
you have called it?
A. Yes.
Q. "... which expressed the average size of a Claimants'
branch (defined in terms of customer transactions per
day) divided by the average size of any branch across
the Post Office network defined in the same way."
A. Yes.
Q. The relevance of this is that because the claimants'
branches are smaller, smaller by number of transactions,
they are less likely to be hit by bugs. That is the
relevance.
A. That is right, that's the relevance. By about this
factor too.
Q. You say 0.37 in your first report and change it to 0.45
in your second report?
A. That is right.
Q. If we look at paragraphs 113 to 114 over the page, these
are the paragraphs that Mr Coyne was taken to {D3/6/30}.
You say four lines down "In my first report" because
this is your second report?
A. Where are we? Yes, got it.
Q. Paragraph 113, four lines down:
"In my first report, I made a calculation for each
year of the size of Claimant branches versus all
branches, and then took an average across the three
years."
A. Yes.
Q. Is that based on the data of 2001, 2007 and 2018?
A. Pretty much, I think, yes.
Q. Let's go back and see what you in fact did. {D3/1/149},
624.1.
A. Yes.
Q. You are referring at 624 to the figures in the
spreadsheet Mrs Van Den Bogard had?
A. Yes.
Q. You say:
"If this spreadsheet is accepted, it implies the
following about Claimants' branches:
"624.1. From summing rows of the spreadsheet, the
561 Claimants' branches carried out 558,000 customer
transactions per week ..."
Yes?
Then what you do is you subsequently compare that,
yes?
A. Yes.
Q. To get the number of customer transactions per day?
A. Yes.
Q. Now, there you are referring to the spreadsheet in
Mrs Van Den Bogard's evidence, yes?
A. Yes.
Q. And you actually say at 624.1 that you have looked at
the customer transactions per week in 2007?
A. Yes.
Q. Not an average over three years?
A. I have made an assumption of, if you like, linear change
to get at that figure. I mean, this figure is very
approximate and only needs to be very approximate, and
it was very approximate in my first report. I tried to
improve it a bit in my second report, but 2007, if you
have a uniform rate of change, 2007 is the average of
those three years.
Q. What you have actually done is take the figures for
2007?
A. Yes, as representative of the three years.
Q. But what you didn't do, you didn't take an average of
the three years by actually adding up the three years
and --
A. No, there was an extra assumption in there that the rate
of change between those two periods was the same, as the
average of the three years was the same as the 2007
figures. And that was, if you like, a hidden assumption
that I probably should have been clear about.
But as I say, the level of precision -- one is
always concerned with the level of precision one
needs --
Q. We have got that point.
A. -- in an answer, and I felt that the level of precision
I needed in this figure, even if we changed it to 1 it
wouldn't make much difference, really.
Q. I understand, I have got your point. But on precision
about what you said, it's not right what you say in your
second report. What you actually did was take the
figures from 2007?
A. Well, 2007 I took as representative --
Q. I understand that. It is not what you said you did.
A. No, I should have explained more carefully.
Q. And the figures you had for the network were 2017 and
then you derived some for 2003. That is right, isn't
it?
A. Where are we?
Q. We are in 625 and 626. At 625 is the figure for 2017.
A. Yes. Again, there is an interpolation here that
I didn't have a middle point so I took the average from
the first and the last point.
Q. Okay. So that's what you in fact did in your first
report, and so the change of method that you referred to
in your second report as a subtle statistical change of
method, from one type of averaging to another, it
wasn't, shall we say, fully explained in your second
report, the change of approach, was it?
A. Well, the change in method was explained and I can
re-explain that if you like. It is a matter of --
Q. We have only got so much time, Dr Worden, so let's look
at the analysis.
You refer at paragraph 623 to the spreadsheet in
Mrs Van Den Bogard's witness statement which she refers
to at paragraph 179 of her witness statement. That
spreadsheet is at {F/1837/1}. Were you in court for
Mrs Van Den Bogard's evidence where the original version
had wrongly allocated the figures across branches and it
had been changed but she didn't know it had been
changed?
A. I wasn't in court for that, no.
Q. Did you realise that had happened?
A. I was told about some changes and I believed that within
the margins of precision that I needed, those changes
did not make any significant difference.
Q. Okay.
Now, the spreadsheet -- can I just pause. When you
were told there had been a mistake in the first
spreadsheet, in data that Post Office had extracted, and
they had to change it, did that worry you at all?
A. I thought about it, but as I say, the precision I needed
in this number is a rather low precision, and to that
precision I felt it didn't make a significant
difference.
Q. Did you do the calculations on the spreadsheet or did
Mr Emery?
A. I did it.
Q. Okay.
A. I produced another version of it which I did the sums --
Q. Can we please go down to row 74. Do you see that
there's a gap in column E?
A. Little Waltham, yes.
Q. If we go down to rows 84 and 85.
A. Yes.
Q. Gaps there?
A. I was aware of these.
Q. You were aware of the gaps?
A. I was aware the data was patchy, yes.
Q. And in fact, Mrs Van Den Bogard explained in her witness
statement that the gaps had been filled, didn't she?
A. Mm.
Q. Yes?
A. Yes.
Q. After further investigation the gaps had been filled?
A. Yes.
Q. And you had intended to take the number of the filled or
unfilled column.
A. I can't remember the precise details of my calculation
because, as I say, I was looking for a precision which
was rather low. And therefore in terms of priorities
and what I put my effort to, I knew that other parts of
my calculation had greater imprecision than that, and so
I didn't spend a lot of effort on trying to refine this
figure.
Q. Okay. So will you take it from me that the column E
figure is 558,260, which is the figure you have used?
A. Column E?
Q. Do you see E is the one with the missing ones in?
A. Yes.
Q. And F has got the gap filled?
A. Yes.
Q. Do you see that?
A. Yes.
Q. And the sum of column E is 558,260 and with gaps filled
the figure is 575,803?
A. Sorry, can I have those two figures again? Are they on
this spreadsheet? They are not. You have done the
sums.
Q. They're not but we could sum them at the bottom. I've
got one here.
A. What are the two figures?
Q. Column E with the gaps in it is 558,260, which is what
you have used.
A. Yes.
Q. So you have used the columns with the gaps in it.
A. Yes.
Q. And Mrs Van Den Bogard explained another column was
there which had the gaps filled, and you have not used
that and that comes to 575,803.
A. So the difference is 4%.
Q. I'm just saying you used the wrong column.
A. Well, the kind of calculation one makes is I'm doing
an average of three numbers, one of them is going to
change by 4%, so the impact on the average is 1%. So
does that matter?
Q. No, Dr Worden, sorry, you used the 2007 figure as
a spurious average proxy for three years' averages on
the basis of an assumption that you didn't tell anyone
until you were being cross-examined. That's what
happened.
A. No, it is not what happened.
Q. Right. You explain what happened here.
A. What happens --
Q. Because that sounds like --
A. -- is when you are making an engineering estimate you
have in mind two things: one is the precision of your
result that you need to assist the court, or whatever it
may be, and the other is the precision that you can
achieve in your calculation.
Now, I was aware when I made this calculation that
the precision I could achieve, firstly the precision
I needed was rather small. It might have been to
a factor of 3, or something like that. Secondly, the
precision I could achieve was rather small for reasons
completely unconnected with this; other parts of the
calculation were less precise.
Therefore, you say: when I'm looking at this
calculation, how precise do I need to make it? And if
there's going to be a 4% difference in one of the
numbers that is added up to 3, is that a significant
difference? And you say no, and so it is, you know, not
worth bothering about that. And that is what people do
in engineering calculations and they do not worry about
spurious precision.
Q. Okay. Would you regard it as spurious precision to have
found out for yourself whether the gaps made a material
difference before you decided not to care about them?
A. You can do it by eyeballing.
Q. So you did eyeball the difference? You are saying you
can do it by eyeballing. Are you telling his Lordship
you did eyeball the differences and form a view that it
wasn't worth bothering about?
A. I formed a view that the changes --
Q. No, did you eyeball it and form a view, or is that
an answer of convenience?
A. I believe I did.
MR JUSTICE FRASER: Right, you mustn't overtalk. It is not
you, Dr Worden.
Mr Green, you really mustn't because I would like to
hear what the witness says and then you can follow
it up.
A. But as I say, the whole basis of this calculation is how
much precision do I need for different parts of it. And
I don't spend time struggling to produce high precision
in parts of the calculation that will not produce high
precision in the result.
Q. So are you saying you knowingly chose the column with
gaps in it because you eyeballed and decided not to
bother with the correct column?
A. I can't remember the precise sequence, but I decided as
a judgment about precision required that it was not
necessary to worry about these details.
Q. Why did you divide by 561? Because that's the number of
claimants? Is that the reason?
A. That's basically the reason.
Q. Did you notice that there were only 496 listed on the
spreadsheet or was that not a level of detail you needed
to bother with?
A. Again, I can't remember how exactly I treated that,
but --
Q. You just didn't notice, did you?
A. No, I did notice. I was aware of that difference.
Q. So you knowingly presented to the court a figure which
represented 496 --
A. I knowingly presented to the court a set of estimates
which I believe are sound and have adequate precision in
each part of the estimates. And the ultimate precision
that's required in the assessment is not very high.
Q. Can we go to row 1, please. Do you see what the
heading is?
A. Sorry, what are we looking for?
Q. The heading of the entire table.
A. Yes. The header row or the heading?
Q. It says "Volume of Customer Sessions", doesn't it?
A. Yes.
Q. Are you are aware that that's what it was both from the
heading and from the fact that Mrs Van Den Bogard said
it was customer sessions?
A. Yes.
Q. And you are also aware that a session is not confined to
one transaction; it could be three?
A. The average I believe is 1.7, but again, in the sources
of data that I had, there were ambiguities.
Q. Hold on a second, you have just told the court that the
average session contains 1.7 transactions. You just
said that.
A. There are various different terms used in different
places and there are issues of comparability of data,
I agree with that entirely.
Q. Hold on, you have scaled down the claimant branches on
the basis of one variable and one variable alone, namely
the number of transactions?
A. I have --
Q. To do that you have used the number of sessions for
claimants and the number of transactions for all
branches, and you know that they are different and you
know that on your evidence the difference is a factor
of 1.7?
A. No. When I was consulting various different sources
about volumes of transactions, there are various
different sources as well as Ms Van Den Bogard's
evidence, there are early documents and so on and so
forth, and I was aware that there are uncertainties of
definitions about these things and I was doing the best
I could and there are uncertainties in the result.
There is no question of it.
Q. Dr Worden, this is a serious factor because you well
understood, as an expert who has looked into transaction
integrity in session data, that there could be more than
one transaction in a session. You knew that, didn't
you?
Take it in stages. Did you or did you not know that
there could be more than one --
A. Well, the definition of transaction is rather fluid.
For instance, when you have a customer session, that is
genuinely packaged in -- what the 1.7 refers to, it is
unclear in different spaces, but what the 1.7 refers to
is generally how many things a customer does in one
session. But that still doesn't alter the fact that
a session in some terminology is one transaction because
it is a success unit in the database.
So there are these ambiguities and I agree there are
big uncertainties in my calculations.
Q. So in fact, even if we take your figure of 1.7, we
might -- none of this is explained in your report, none
of the uncertainties?
A. Well, this is not the big uncertainty in the analysis.
The biggest uncertainty in the analysis comes from the
number of bugs and the scale of each bug, and that is by
far a bigger uncertainty. And so I felt in terms of
priorities and length of explanation and so on and so
forth, it was not -- you know, the engineering approach
is not to drill down on things that you do not think are
the main uncertainty in your result.
Q. But you might have had to double, or nearly double your
scaling factor on your own evidence?
A. I put in a conservative (inaudible) factor from 0.45
to 0.5, which I will acknowledge I could have done.
Q. But you have left out a massive point the court would
need to know and certainly came up in cross-examination.
A. There are all sorts of uncertainties in this analysis
and I drew attention to what I felt were the main
sources of uncertainty.
Q. So just to clarify, wrap up, you knowingly took the 2007
column with gaps?
A. Yes, that's, as we discussed, an effect of 1 or 2%..
Q. And you eyeballed that to guess what the effect was?
A. Yes.
Q. You say that you noticed that there were only 496 lines
in the spreadsheet but proceeded nonetheless?
A. I believe I corrected that. In fact I believe what
I did was a rough correction which said just put in that
factor 491 over 561 or whatever it is basically.
Q. We don't see it anywhere in your calculation --
A. No, absolutely. I was making decisions about what's
worth drilling down and what's worth exploring and
I felt that the imprecision in the calculation, the
subjects I had to talk about at length, were other
imprecisions and that this claimants' scaling factor, it
doesn't make much difference if the claimant branch size
on average is the same as "other", because a factor 2 is
less than the precision I need in my final result in
order to be useful to the court.
So I was focusing my attention on what I felt to be
the biggest areas of uncertainty in my analysis.
Q. And there's no hint of those uncertainties mentioned in
your report?
A. No, the biggest areas of uncertainty, there is a lot of
mention of them.
Q. And there are a whole load of other points. Let's just
see, for example, you refer to, taking it very quickly,
you may take it from me you refer to 48 million
transactions. Again, it is not a big difference but the
actual figure is not in Angela Van Den Bogard's second
witness statement, it is in the first one. It is
47 million, not 48 million. You say Mr Coyne also
referred to that and he correctly recites it as 47.
This --
A. 47/48, in the context of my calculation I certainly took
a difference of 47 to 48 as being insignificant.
Q. So all of those errors, as we say they are, you say they
are fine? They are all against the claimants, aren't
they? If they are errors, which is a matter for his
Lordship, all of them have an effect against the
claimants, don't they?
A. What I have tried to do is present a numerical
calculation --
Q. Can you just answer the question?
A. Yes.
Q. They are. And if you are an independent expert and the
chances of you making a mistake one way or the other,
say against the claimants, is 50/50, yes?
A. Mm.
Q. The chances of you making the first mistake against the
claimants is a half?
A. Sorry? Which mistakes against the claimants are we
talking about?
Q. The four that I have just identified.
A. Right and what I would say is those can have an effect
and that effect, in my opinion, is balanced by the
claimant favouring factors, that you can see in table
8.4, which have an effect of a factor of approximately
30.
Q. Dr Worden, we are just talking about the primary maths.
A. Absolutely, the primary maths, the maths folds in errors
and changes made from various sources. There are
uncertainties and uncertainties and there are deliberate
claimant favouring factors which have injected, which
amount to a factor of 30.
Q. We will come to those. Just on this analysis, making
four 50/50 mistakes against the claimants, the chance of
6.25% of that being random; 1/16.
A. You are talking two to the four, is that what we're
talking about?
Q. Half of 50 is 25. Half of 25 is 12.5.
A. I do not see the relevance of that figure.
Q. Okay. It is just that the claimants are just unlucky
that all your errors are against them?
A. In this respect yes, but there are large numbers of
errors which are in their favour in the calculation and
I think one should take a balanced view of the errors in
their favour versus the errors against them.
Q. Dr Worden, would you accept that there is a difference
between an error that is not apparent on the face of the
report and a prominently flagged bit of generosity to
the claimants when you adjust a figure that you have
derived in their favour slightly? You understand what
the difference is, don't you?
A. I have tried to present a balanced opinion on this
number.
Q. I'm not asking you that because you have given that
answer about six times. I'm asking you do you
understand the difference between a concealed error in
your report, ie one that we can't see on the face of the
report, and somewhere where you say: I'm now going to be
enormously generous to the claimants and increase the
number to be really generous and conservative. You know
the difference between those two things, don't you?
A. I would not characterise it that way, I would say I made
some engineering assessments about what were the largest
sources of error in my calculation and I focused my
attention on those, and one made the typical engineering
decision that you put your priorities and your attention
where the big errors are and that's what I did.
Q. Let's very briefly, in your second report, we will just
go back to that quickly {D3/6/30}. This is where we
were looking at it and you explain that you have looked
across three years. We have only seen two years of data
for all branches. You used only two years data for the
all branches figure, didn't you? Yes?
A. I think that's the way I did it, yes.
Q. Okay. Let's look at the smaller branches scaling factor
methodology now, how you used the scaling factor. Let's
look at paragraph 630 in your first report {D3/1/150}.
We are looking at paragraph 630.
A. This is in the first report?
Q. Yes, it is. On the basis of the session data
transaction data comparison you say:
"So the claimants' branches, being generally smaller
than the Post Office average, have fewer transactions
per month and so are less likely to be hit by Horizon
bug in a given month."
Then you give the scaling factor.
A. Yes.
Q. Now that analysis depends on three things, doesn't it?
First, it depends on there being one variable by which
that scaling is to be measured and that's transaction
numbers?
A. Yes.
Q. Yes?
A. Mm.
Q. And it also depends on that varying directly in
proportion to bugs --
A. There is a linear assumption in there, yes.
Q. So the assumptions behind it are one variable, varying
directly in proportion to bugs and therefore reaching
your view that the scaling factor is how to calculate
the likely incidents of bugs on those claimant branches.
That is a fair summary, isn't it?
A. I think so, yes.
Q. Now, what we do have is data about branches which were
affected by actual bugs that we do know about?
A. We do.
Q. Yes. And if we were to take the Dalmellington bug, the
88 branches affected by the Dalmellington bug were tiny
outreach branches, weren't they?
A. Well, an outreach branch is tiny but it is an outreach
from a branch which -- I haven't looked into the size of
branches which had outreach branches.
Q. Yes and so to do this, even if you assume it is one
variable, which we will come to and so forth, if you
wanted to get even a sense of whether this was right, it
would be sensible to calculate -- to look at the number
of transactions done by the branches that were suffering
the identified bugs because you don't know how they
compared to average branch sizes, do you?
A. No, I don't.
Q. And that's not a calculation you did?
A. No, this is what I would call the next level of detail
down that one could have gone to look at the branches
affected by the known bugs or the acknowledged bugs~--
Q. Would you -- sorry.
A. -- and one could have calculated the average size of
those and I could indeed have factored that in my
calculation, I --
Q. Would you accept that's basic statistics to identify the
characteristics of the control?
A. Identify the characteristics of the control? To
the extent you can, yes.
Q. And you didn't do it?
A. As I say, it is an extra level of sophistication in my
calculation which I didn't do. I acknowledge that.
Q. But you are a bit of a legend in advanced statistics in
court, aren't you?
A. I don't know what that means.
Q. Well, you have given evidence in very prominent cases,
expert evidence, of advance statistical techniques; you
have a PhD in theoretical particle physics; and you have
come to this court in an important case to give expert
evidence about statistics?
A. Yes.
Q. Right. And you say what about why it is acceptable for
you to leave out that step of the analysis?
A. Because one makes a calculation at certain levels of
approximation based on the precision you require in the
result, and the precision I required in the result was
probably to within a factor of 3 or a factor of 10 and
I made the judgment that to achieve that precision
I didn't need to go to that next level of detail that
you have described to me.
Q. So if the average branches, if you had regard to them
having, say, 1/10 of the average transactions of normal
branches, you might have had to have a scaling factor of
x10 instead of x0.37 or 0.45?
A. There is a contradiction there. If my central
assumption is that branches are affected by bugs in
proportion to the number of transactions that they do,
then it is highly unlikely that the branches affected by
the three known bugs should be small in terms of their
numbers of transactions.
Q. That's why it throws up whether your assumption that it
is a uni-variable analysis based on transaction numbers
is right, isn't it?
A. Sorry, I don't understand the question.
Q. You have pointed out that the small branches being hit
would have helped you to identify whether you were
likely to be right that the only relevant variable to
consider was transaction numbers?
A. Sorry, I'm not really following the question.
Q. You have just pointed out when I gave you the
Dalmellington example, you said if it is lots of tiny
branches being hit, you said, well, that would question
my assumption that it is proportionate to transactions.
A. But I think it is unlikely that the Dalmellington
branches were tiny for precisely the reason I have just
stated.
Q. And it is right that a bug that affects an SPM when they
roll over -- yes?
A. Yes.
Q. If you had a bug which just affected rolling over?
A. Yes.
Q. A branch, however large or small, they all have the same
number of --
A. Same number of rollovers. Absolutely. But the branch
that is bigger has some bigger figures in its accounts,
so the effects of one bug on one rollover is likely to
be larger. So that --
Q. In number terms but not in instance and probability
terms?
A. Sorry, I do not understand.
Q. I'm just saying if a bug affects rolling over, just the
process of rolling over, the chances are not affected by
the number of transactions that they are otherwise
doing?
A. No, but the financial impact, which is the main
calculation, is affected.
Q. Yes, but I'm just trying to point out to you that it is
not based on the number of transactions as the frequency
or likelihood of bugs happening in all cases, is it?
A. But my main calculation was financial impact and I knew
that for financial impact larger branches had larger
numbers of transactions and they also had bigger numbers
involved in roll over, so the financial impact, it was
equivalent.
Q. But you deal with that separately. Your scaling factor
is about probability, isn't it?
A. Which scaling factor?
Q. The scaling factor that we have been talking about this
afternoon.
A. You say it is about probability --
Q. The premise of the scaling factor is that they are
equally likely to be hit; an average branch is equally
likely to be hit by another average branch, but
a smaller branch is less likely to be hit because they
do fewer transactions. That is the premise of your
scaling factor?
A. Well, the scaling factor applies to financial impacts.
Q. No, can you just answer the question because you know
what I'm asking. The scaling factor is designed to
control for a lower chance of a branch suffering
an impact of a bug, the incidence not the value, the
incidence of a bug by reference to the number of
transactions that happen in the branch?
A. Yes, we can agree --
Q. That is correct?
A. I can agree with you that if a bug affects roll over,
then a branch that is small is just as likely to be
affected as a branch that's larger, yes. Absolutely.
Q. And if a bug affects remming in the pouch --
A. Yes.
Q. -- if the pouch contains £100,000 or £2,000 for a
smaller branch, it doesn't have any impact necessarily
or any relation to the number of transactions they are
carrying out, does it?
A. I think it does because -- again we are going to a very
fine level of resolution here, but generally one would
expect a larger branch to have more cash remming.
Q. But that might be -- you haven't got any information
about that?
A. We are going down to the next level of resolution in the
calculation here and, you know, one has to cut off
somewhere in the precision of the calculation and the
place one chooses to cut off depends on the precision
you require in the answer.
Q. Finally, in relation to mathematics, we saw with the
spreadsheet at {F/1837/1}, if we can get that up, about
the column without the gaps filled been taken. Yes? We
have been over this. I'm just suggesting to you that
that was actually a mistake, Dr Worden. You just got
the wrong one. You misread -- go up to row 1 please.
What actually happened is you looked at the column
headings, saw 2007 gaps and thought that was the one
with the gaps in it and you made a mistake and took
column E. That's what really happened, isn't it?
A. I'm trying to remember what these columns do mean. You
seem to get the same figure in those two columns, so for
the level of precision I need, you can take either
column.
Q. What that first page doesn't show is all the gaps that
are on the other pages. I'm just putting to you that
actually the evidence you gave earlier isn't right, you
just made a mistake?
A. Well, there are imprecisions in that calculation we have
agreed that, and I felt that those imprecisions were
insignificant in the context of my overall calculation.
Q. It wasn't a deliberate choice to take the wrong column,
was it?
A. I made no deliberate choice to take the wrong column,
no.
Q. It wasn't a deliberate choice to take the column without
the gaps filled, was it?
A. I repeat that I don't think the gaps make a difference
in the context of the accuracy that I'm looking for.
Q. I'm only going to ask it one last time because we are at
4.29. The question is, was it a deliberate choice for
you to take the column that still had gaps in it?
A. It was not a deliberate choice, no.
Q. Was it a mistake?
A. It was a judgment that that issue was not important
enough for me to worry about it.
MR GREEN: My Lord, is that a convenient moment?
MR JUSTICE FRASER: Yes, it is. I have just got a couple of
requests it is just for some documents or some
references. Mr de Garr Robinson, can I just have a hard
copy of Dr Worden's corrections sheet.
MR DE GARR ROBINSON: My Lord --
MR JUSTICE FRASER: I don't need it now.
MR DE GARR ROBINSON: You can have it now --
MR JUSTICE FRASER: Well, if you have a spare one.
MR DE GARR ROBINSON: I believe I have got --
MR JUSTICE FRASER: I was reading it on screen but it is
just helpful and the other thing and this is --
MR DE GARR ROBINSON: This is clear, if I can hand it. My
learned friend has one.
MR JUSTICE FRASER: That is a different thing I think.
MR DE GARR ROBINSON: This is the corrections.
MR JUSTICE FRASER: So that is the corrections given in
chief. I think Mr Green is proffering a copy of the
signed joint statement. Thank you very much.
Then the only other reference which is one for
Mr Green, Mr Green can you just give me, tomorrow is
fine or by email in the morning, just a reference for
the covering letter that came with the MSC document that
you were putting this afternoon at {F/1834.14/1}.
MR GREEN: My Lord, one of them is at {H/325/1} which is
where it is explained that the documents with the new
MSCs are self-evident and don't need to be explained and
there I think is a couple of others which we will find
for you.
MR JUSTICE FRASER: If you maybe just pop them in an email
and copy into Mr de Garr Robinson and in the morning is
fine.
Thank you all very much. Actually when I say in the
morning I don't mean in the morning, actually for you I
do mean in the morning.
MR DE GARR ROBINSON: Really?
MR JUSTICE FRASER: I know we are not sitting tomorrow, so
I'm not asking you to do anything tomorrow. The thing
I was asking Mr Green to do by in the morning, I go back
to my original position, please do send it to me in the
morning.
MR GREEN: I will send it. I'm most grateful.
MR JUSTICE FRASER: Dr Worden it is going to be a bit more
difficult for you because you have one day interval, but
you know you are not to talk to anyone in the case and
we are going to come back on Thursday.
A. Yes.
MR JUSTICE FRASER: Is there anything else for today?
MR DE GARR ROBINSON: My Lord, no.
MR JUSTICE FRASER: I hope tomorrow goes as well as can be
expected in all the circumstances.
MR DE GARR ROBINSON: My Lord, that is very kind. Thank
you.
(4.32 pm)
(The court adjourned until 10.30 am on
Thursday, 13th June 2019)