Wednesday 13 March 2019

Horizon trial: Day 3 - Patny, Burke and Roll

Richard Roll outside court today
The most important information imparted today was by the judge, just before the end of proceedings at 4.25pm. He told the court he was on schedule to hand down his findings on the Common Issues trial at high noon this Friday. He added, as an aside, that it is 180,000 words long. Friday is going to be a loooooong day.

Patny, Burke and Roll

The three claimant witnesses being cross-examined today were Aakash Patny, son of Anup (who took the stand yesterday), Angela Burke, and Fujitsu whistleblower Richard Roll, pictured above.

Surprisingly, Aakash Patny was the most interesting of the three. He was involved in some of the same incidents his father had given evidence about and he was deeply unimpressed by the line of questioning the Post Office barrister, Simon Henderson, took with him.

Mr Patny says in his witness statement that back in 2016, £16,000 went "missing" when he declared the value of the stamps in his branch to Horizon. Mr Henderson pointed out, looking at the Horizon records it was nearer £18,000, but both agreed there was no way it could be that much.

Almost as soon as it happened some red flag went up at the Post Office and the branch got a call from a Mrs Debra Lambley from the Post Office's National Business Support Centre in Chesterfield. Mrs Lambley queried the declared value of the stamps and she and Aakash set about trying to get the correct value sorted and declared.

So far so good, but Mr Henderson appeared to think that Mr Patny only had one phone call with Mrs Lambley. Cue confusion all round.

   Q.  It has been picked up by Chesterfield and business as
       usual, they have said "You need to sort it out", you do
       sort it out and that's what happens.  Why do you say
       that that's a bug in Horizon?
   A.  Because after I've sorted it out, it's stayed there.
   Q.  I'm sorry, I didn't hear?
   A.  After I'd sorted it out, the issue has stayed there,
       hasn't it, it has gone back up to that many stamps up?
       So why would I change it down and then change it back
       again?  I'd understand you said that I'd made the
       mistake the first time, but I couldn't make the mistake
       again the second time, could I?
   Q.  I don't quite understand the point you're making?
   A.  I'm saying you're saying I potentially could have made
       a mistake the first time.  If I did make a mistake the
       first time, fair enough, but I couldn't make the same
       mistake again, could I?
   Q.  This just suggests that stamps are being declared at the
       same sort of level for that week until 25 May and it is
       then corrected, isn't it?
   A.  Yes, but the stamps issue came back.
   Q.  When?
   A.  The following week.  After I adjusted them they still
       went back up.  Debra had to phone me twice.
   MR JUSTICE FRASER:  It might help if you did some of this by
       reference to his witness statement, Mr Henderson.  We
       are having a lot of "I don't understand".
   MR HENDERSON:  I apologise, my Lord.
           Well, I don't understand and I apologise for that if
       it is my error, but are you saying that you carried out
       a declaration which took it down, took the value down
       and it went back up again?
   A.  Well, yes.  Debra -- sorry, I have forgotten her
   Q.  Lambley I think.
   A.  Debra Lambley phoned in the morning, spoke to my father.
       When I came in I spoke to her.  She made me readjust
       them down and what she made me do was very complicated,
       she made me go into areas of Horizon I could never dream
       of going into and then this issue of the extra stamps
       came back after I had done that, so then she's had to
       phone me again the following week and her words were
       "Why have you changed them back?" and I said
       "I haven't".
           So in answer to your question I could have made
       a mistake the first time, okay potentially I could have,
       but then why would I make the same mistake a second
   Q.  So what you have said -- if you look at your witness
       statement in paragraph 13, you say:
           "She told me it was showing the branch had
       overdeclared the amount of stamp stock that was held."
           And we have gone through that?
   A.  Yes.
   Q.  "Ms Lambley told me I had to adjust the stamp stock on
       the system."
           And you say that you did that.
   A.  Yes.
   Q.  All right, that's not shown in that declaration that we
       just looked at?
   A.  But she checked it on her side and said that everything
       was fine as well.
   Q.  And then you say you rang the helpline and said that
       they had given incorrect advice and you had to readjust
       the stamp stock?
   A.  No, when I rang the helpline that night I was £16,000
       short and they made me do various adjustments to make it
       okay again and what this has done is somehow the stamps
       have gone back up and Debra has had to ring me again the
       next week.  So then as I was answering your original
       question, if I made the mistake of entering extra stamps
       the first time, I haven't definitely done it the second
       time, have I?
   Q.  Well, I accept it is unclear, Mr Patny, but what appears
       to have been the case according to the stamp
       declarations is that there was a problem which persisted
       for a week and then was corrected.
   A.  But it hasn't corrected though.
   Q.  Well, it was corrected on 25 May.
   A.  On the day -- Debra rang me the first day, she has
       corrected it.  The issue has persisted, she has rang me
       again a couple of weeks -- well, a week and a half later
       and she has made me do the same thing again.
   Q.  And is it possible that you hadn't properly followed the
       instructions that you were given by her over the phone
       at the time?
   A.  No, because Debra checked it on her side and said that
       everything was okay.
   Q.  But she wouldn't know what stamps you had in the branch,
       would she?  You would have to count it properly?
   A.  Well, if she didn't know what we had in the branch then
       how did she know we were £18,000 up on stamps then?
   Q.  I think the point about that is it is just
       an extraordinarily high amount which everyone accepts
       would not be the number of stamps in a branch of your
   A.  Well, then you can't have it in -- you're talking
       hypothetically in her defence but talking factually in
       mine, aren't you?
   Q.  No, because I think it is agreed, isn't it, that you
       didn't have anything like £18,000 --
   A.  No.
   Q.  -- worth of stamps in your branch?
   A.  No, we didn't.
   Q.  You did have that number?
   A.  We didn't.
   Q.  You didn't, okay.  So what happens is a process whereby
       it is adjusted from a high figure that's been wrongly
       declared to the correct figure?
   A.  Yes.  And Debra has agreed at that point that everything
       is okay.
   Q.  Okay.  Well, I don't think I can take that any further
       for now.

I hope you're all following that. Mr Henderson then started off on a different issue Mr Patny had raised about a MoneyGram cancellation, but soon after asked for a short break.
On reconvening Mr Henderson went back to the stamps and having earlier told his witness "no one is saying it's the worst mistake to have made" promptly accused Mr Patny of false accounting:

Q.  you remember the £18,000
       declaration of stamps, was that an attempt by you to
       disguise a shortfall of cash?
   A.  No.
   Q.  I suggest it may have been.  I suggest it was, Mr Patny.
       You deny that?
   A.  I do, yes.

Mr Patny was implacable. In his mind he had neither made a mistake, nor suddenly decided to engage in potentially criminal activity. It was a Horizon bug.

Quite what advantage the Post Office sought to gain today by accusing Mr Patny of deliberately inflating the stamp value to hide a shortage of cash, I have no idea. It was all very odd. The stamp conversation had been put to bed and Mr Henderson had moved on to the MoneyGram problem. Then there is a sudden request for a break. After the break Mr Henderson comes back, parks the MoneyGram issue, returns to the stamp problem and without warning unleashes the Criminal Offence Baboon.

There must be some kind of legal tactic here, and if there is a lawyer reading this, I'd be grateful if they could get in touch. To my mind it was indicative of the Post Office's alleged historic willingness to assume criminality and prosecute so many Subpostmasters on the basis of flimsy computer evidence without bothering to investigate the actual issue.

Angela Burke

The Post Office fared rather better with the next two witnesses. Angela Burke had an entertaining tale of losing £150 during a computer crash. Thanks to her experience with Horizon, diligence at her own job, good fortune with CCTV and the happy co-operation of a customer, she was able to hunt down the source of her £150 discrepancy and provide the Post Office with the evidence she needed to successfully request a transaction correction from the Post Office.

Her natural suspicion of the Post Office and the failure of the Post Office helpline to reassure her that she wouldn't be out of pocket was part of the motivation for her Nancy Drew style detective work (read her witness statement here). Unfortunately for the claimants, the Post Office barrister was able to provide compelling evidence that even if she had done nothing, or made one call to the helpline flagging up her loss, she would have received a £150 transaction correction anyway. She wouldn't have known about it until it appeared because the Post Office don't seem to communicate these things, but from the evidence produced by the Post Office - it was in hand.

Richard Roll

The final witness was Richard Roll. Mr Roll got involved in this story seeing after seeing a film I fronted on the Post Office Horizon story for BBC Inside Out South on 7 Feb 2011. I approached Richard to see if he would be interviewed for Panorama in 2015. He very kindly agreed. We haven't really spoken since and couldn't speak today as he was being a witness, but he did, again very kindly, agree to post for some photographs outside court.

Richard's witness statements (read 'em here) are very similar to what he told me about the sorts of things he got up to whilst working on third tier support for Horizon at Fujitsu HQ in Bracknell between 2000 and 2004.

Today, the Post Office QC - Mr de Garr Robinson - using some extraordinarily detailed data provided by a Mr Parker from Fujitsu, did a very good job of convincing Mr Roll (and possibly the judge) that a lot of his witness statement was exaggerated and/or mistaken. This wasn't done by challenging the substance of the evidence Mr Roll provided, but by cleverly and methodically putting it into context.

As Mr Roll himself conceded:

"I think one of the comments Mr Parker made in his
 statement at some point was that some things stick in
 the mind more than others and over time one tends to forget the mundane aspects of the job and tends to remember more the aspects that were not mundane, so my recollection is that -- as I put in my statement --
 I spent sort of 70% of the time doing not mundane tasks,
 as it were, and only 30%, but that could be wrong, it
could be my perception of events from 15/17 years ago."

From that moment Mr de Garr Robinson started to narrow down the scope of Mr Roll's work, the importance and level and complexity of it and the amount of time he spent doing anything exciting. Whilst Mr de Garr Robinson was at pains to suggest to Mr Roll he was was not belittling his importance, he did so whilst progressively painting his witness as little more than a grunt. This is the QC quoting from Mr Parker's evidence:

Q.  He says:
           "Mr Roll would also have been regularly correcting
       the application environment after engineers had replaced
       failed counter hardware and clearing temporary files to
       increase disc space."
           Is that true?
   A.  It's possible.  I don't remember it, but it's possible.
   Q.  "This could fairly be described as 2nd line work and it
       was done by the SSC because it required a higher level
       of access to the system than other support teams had."
           Would you accept that?
   A.  It could be considered that.

There then followed a detailed statistical analysis of how much complex work Mr Roll did:

Q.  Okay.  Well, this is Mr Parker's categorisation and you
       will see that he distinguishes between potential
       software error and known issue/work-arounds and if we go
       down to the bottom of the table, looking at the overall
       figures of the 27,002 [the number of calls which came through to Mr Roll's department while he was there], we have 2,252 which are treated
       as potential software errors, that's 8.3% and I think we
       have already discussed that figure and I believe you
       said you are not in a position to dispute that
       calculation, is that right?
   A.  That's correct.
   Q.  And then if we go over to "Known issue/work round" there
       are 9,545 of those which represents 35%, 35.3% of the
       work being undertaken by the SSC while you were there.
       Again, would I be right in thinking you are not in
       a position to dispute that?
   A.  Yes.
   Q.  Thank you.
           Now, staying with this document I would like to go
       to a different tab please.  Could we go to the
       "Final responses" tab.  Now, this, Mr Roll, is an
       analysis of the PEAKs for which you were the final
   A.  Right.
   Q.  And you will recall that this is addressed in
       Mr Parker's witness statement.  You will see that there
       are 915 PEAKs in which you are the final responder while
       you were there in your 44 months when you were at
       Fujitsu and there's a breakdown of which particular
       categories you had applied to the relevant PEAKs when
       you closed them.
   A.  Right.
   Q.  Do you see that?  And if we could go back please to "RRP
       live PEAK by category" which I think is to the right,
       you will see that as well as an overall analysis of the
       work done by the SSC there's also an analysis of the
       would, that you did, "RR", this is in the pale -- is it
       beige?  When I was a child I think I would have called
       it fawn but we have to say beige now.  There is a beige
       line and it has your 915 PEAKs and you will see
       "Potential software error", the total for the SSC is
       8.3% but do you see that your total comes out at 3.2%?
   A.  Mm-hm.
   Q.  Would you be in a position to challenge that percentage?
   A.  No. 

It gets even more detailed...

MR DE GARR ROBINSON:  So, Mr Roll, with that analysis can we
       go back to your first statement please.  It is {E1/7/2}.
       It is paragraph 11 that I'm interested in.  You say:
           "I would estimate that there were over 30
       individuals working on the same floor as me ... in the
       Fujitsu building at Bracknell.  My recollection is that
       many of those individuals were involved in similar work
       to myself, or were involved in other Horizon related IT
       work.  I would describe much of the work being carried
       out as 'fire-fighting' coding problems in the Horizon
           For "coding problems" I take it you're talking about
       software bugs.  On the basis of the analysis of the
       evidence that we have just been through at slightly
       painful length, would you accept that it is possible
       that you may have rather exaggerated the position, no
       doubt unintentionally, in the last sentence of
       paragraph 11?
   A.  Certainly the way I remember it doesn't agree with the
       figures, so yes, it is possible that --
   Q.  Much of your work was not fire-fighting coding problems
       at all, was it, coding problems represented a tiny
       amount of your work?
   A.  On the evidence you have just supplied, that's correct.
   Q.  Nor did it occupy much of your colleagues' work either,
       if we look at the overall figures, 8.3% for software
       errors, would you accept that?  Even that -- it would be
       unfair to describe the process that was undertaken
       during that four-year period as fire-fighting coding
       errors by the SSC.  It's just not accurate, is it,
       Mr Roll?
   A.  From volume perhaps not.  Time-wise avoidance action
       would obviously take much less time than trying to find
       a software problem.  But the figures certainly
       contradict my recollections of it.
   Q.  That's fair.
And then even more forensic...

   Q. "Receipts and payments mismatches happened more
       often during the early life of Horizon ..."
           And there is a reference to a tab which I will go to
       in a moment:
           "My analysis of that data shows that there were
       around 8.6 such incidents per month on average between
       1 January 2001 and 31 December 2014 (417 out of a total
       of 27,005 incidents into SSC, or 1.5% of SSC incidents
       during that period)."
           We can go to the table if you like but I apprehend
       that you would probably give me the same answer that
       you're going to give me now which is that you are not in
       a position to challenge that analysis?
   A.  I'm not.
   Q.  So what we're talking about, if this analysis is right,
       is 1.5% of incidents going to third level, going to the
       third line which consist of payments and receipts
       mismatches and that means that less than 1.5% of
       incidents going into SSC would have been discrepancies
       in branch accounts caused by software issues, by coding
       issues.  Do you accept --
   A.  Yes.
   Q.  -- the logic of what Mr Parker is saying?
   A.  I don't know if the receipts payments mismatch would
       include other balancing problems or not.
   Q.  But I think you have already agreed that in the vast
       majority of cases such an occurrence would cause
       a receipts and payments mismatch?
   A.  I believe so.
   Q.  I'm grateful.  On the basis of that evidence -- would
       you give me a moment, Mr Roll, I do apologise.  One of
       my questions is being corrected by a junior who is far
       too able for his own good.
           At page 172 I think I did what is conventionally
       called an American misspeaking.  My question was "1.5%
       of incidents going to third level which consist of
       payments and receipts mismatches and that means less
       than 1.5% of incidents going into SSC would have been
       discrepancies in branch accounts caused by software
       issues" and I'm corrected, I should have said:
       potentially caused by software issues.
   MR JUSTICE FRASER:  I think you meant potentially caused by
       coding issues.
   MR DE GARR ROBINSON:  I really do mean that.
   MR JUSTICE FRASER:  But you then did go on to say "by coding
       issues" after software issues.
   MR DE GARR ROBINSON:  But the key word is "potential".
       I wasn't seeking to suggest to you -- I was seeking to
       suggest to you there is a wider category.  You recognise
       of course there are potential coding issues and then
       there are actual coding issues and the former is
       a larger class than the latter.  We have lots of venn
       diagrams today and what I was suggesting to you there
       was I was talking about the larger class.
   A.  Yes.
   Q.  That is what must be less than 1.5% --
   A.  Yes.
   Q.  -- would you accept that logic?
   A.  Yes.
   Q.  Thank you, Mr Roll.
           Now, if you encountered a software error causing
       a branch -- well, actually, I'm so sorry, let's go back
       to paragraph 19 of your first statement which is at
       page 3 {E1/7/3}.
           Could I suggest a way of rewriting paragraph 19 that
       would make it if not more accurate, at least read more
       fairly.  You start by saying:
           "In summary, the issues with coding in the Horizon
       system were extensive."
           Would it be fairer to say that the issues with
       coding in the Horizon system represented a very, very
       small proportion of the issues that the SSC dealt with?
   A.  From the evidence you have shown me just then I would
       agree with that.
   Q.  And then you say:
           "Furthermore, the coding issues impacted on
       transaction data and caused financial discrepancies on
       the Horizon system at branch level."
           Would you accept that the proportion of coding
       issues that had that effect was even smaller?
   A.  Yes.
   Q.  And the next sentence:
           "It was those issues that I, and other colleagues at
       Fujitsu, were routinely working on daily."
           Well, that's not really right, is it, Mr Roll?  You
       and others were not working on these issues daily, were
       you, they were a very small proportion of the work that
       was being done?
   A.  From the bugs then yes.
   Q.  Thank you.
   MR JUSTICE FRASER:  By bugs you mean coding?
   MR JUSTICE FRASER:  I seem to be the only person who cares
       about what words are used.

And finally, the coup de grâce...
   Q.  ...did you ever encounter a coding issue which had
       a financial impact on the branch account?
   A.  I can't remember.  There were times when I suspected
       something like that and passed it on, from what I do
       remember, to one of the other --
   Q.  To seek more senior people --
   A.  Yes.
   Q.  -- who might be able to look at it in more greater
   A.  Yes.
   Q.  But you don't recall ever having encountered, in all the
       PEAKs that you worked on, a coding issue that definitely
       caused a financial impact?
   A.  I don't recall discovering one, no.

Don't forget that whatever sections of Mr Roll's witness statement is not challenged stands, and quite a bit of it remains unchallenged. Read the transcript for more - but remember Mr Roll is still being cross-examined so he'll start again in the witness box tomorrow.

After Mr Roll we'll have the seventh and final claimant witness, Ian Henderson, director of Second Sight - the forensic accountants acrimoniously sacked by the Post Office in the middle of an investigation into Horizon.