Hacker News Reader

Comment by xioxox | Hacker News Reader

In one case (Seema Misra) the judge asked the jury this in his/her summing up:

“There is no direct evidence of her taking any money... She adamantly denies stealing. There is no CCTV evidence. There are no fingerprints or marked bank notes or anything of that kind. There is no evidence of her accumulating cash anywhere else or spending large sums of money or paying off debts, no evidence about her bank accounts at all. Nothing incriminating was found when her home was searched...Do you accept the prosecution case that there is ample evidence before you to establish that Horizon is a tried and tested system in use at thousands of post offices for several years, fundamentally robust and reliable?”

The jury pronounced Seema Misra guilty.

There are many more details in this very comprehensive report by Private Eye: https://www.private-eye.co.uk/special-reports/justice-lost-i...

View on HN

This is a good example of why a jury trial is not always the best choice for a defendant. Juries are often overly impressed by authority figures, and assess the status of witnesses rather than the evidence itself.

Perfect example of the https://en.wikipedia.org/wiki/CSI_effect

Nevertheless a trial by our peers is the best option we have in the long run. Think of alternatives, such as an algorithm, or a single person.

It’s not great, sure, but people from the same circumstances judging a peer is as good as it gets.

It's not as good as it gets at all. It's just a good fallback to have available for defendants should the rest of the system fail.

It's a great option but it should not be the only option, in my view.

A single person (a judge) sounds better than a jury to me.

Personally I’d much rather be judged by 12 of my peers, where they must be unanimous in a conviction.

We've seen time and time again how juries become "unanimous" in their conviction of innocent people. So I'm not sure why you're putting so much weight into 12 people coming to an agreement in a closed room. It happens.

As a software person who has to write tests validating his systems all the time, who was also a juror in a trial that was a travesty... it strikes me as extremely odd that we don't run "mock trial tests" on this of all systems, where the jury, judge and lawyers don't know it's a test.

Let's just say I have SERIOUS doubts that this system satisfies https://en.wikipedia.org/wiki/Blackstone%27s_ratio

If it's too difficult to fake evidence, we could randomize say 1% of the cases where the convicted is exonerated and given immunity and asked to anonymously report their actual guilt to a system in the interest of supporting a better legal system... yeah, this too would be difficult... hmmm... there must be a way to do this

Yup. I knew someone who worked in the Ministry of Justice in France and he was ecstatic when they announced they would stop jury trials for criminal charges.

He said juries just aren't that great at the job. They're easily influenced by current events, susceptible to fallacies, etc. A lot more so than judges, according to my acquaintance (who was admittedly biased).

So they just use a single judge instead? That is (obviously) not always superior.

What about a panel of highly-trained rational-evidence evaluators? This would absolutely be superior to laypeople... Except for the corruption potential

The government should not have brought these cases in the first place. George Orwell and the American founders were hyper critical of the paternalistic and controlling nature of the British aristocracy precisely because it leads to this kind of thing happening.

What I find totally amazing about this whole scandal is that they ignored one of the most obvious and overwhelming examples of Ockham's razor I've ever seen.

On the one hand, you had the possibility of hundreds of sub-postmasters committing fraud in many different unrelated locations. There was zero evidence of them taking any money even after really deep scrutiny.

On the other hand, there was the possibility of software being buggy.

How could they happily conclude it was the first and not the second option is impossible to understand, and leaves me deeply worried about the UK legal system.

Extraordinary claims require extraordinary evidence. In order to support the existence of such a distributed fraud scheme, they should have brought convincing evidence. There was zero. Furthermore, they happily ignored an option that is several orders of magnitude more likely. Namely, that badly built software fails and management was deliberately hiding this to save their face.

> you had the possibility of hundreds of sub-postmasters committing fraud in many different unrelated locations

It is unfortunately fairly common. I have a friend who has been working for the USPS for over twenty years; way back when he got started, a co-worker was busted for pocketing cash from stamp sales. He was my roommate at the time, and was freaking out because she was trying to pin it on him, had to talk to the FBI, etc; the investigators figured out that she had been running the scam over 10+ years (something like $130k if I remember correctly (pocketing about 25 bucks a day, adds up!)). Retail scams just aren’t very unusual.

But you are correct in that evidence is, like, kind of important in a trial…

It's a "guilty until proven innocent" story, and I'm afraid there'll be many more in the future when the black-box software so prints out and whomever is in charge refuses to question the legitimacy of the claims for whatever reason they may have. This one time they made it out, countless more may not.

Yes, you face an uphill battle if one forensic “expert” has decided you were guilty. It is actually quite similar to non-falsifiable pseudo-science forensic techniques that are used to impress jurys. Being innocent is not enough.

Will be even worse when the black-box in question is a deep neural network that can't show its workings out

Thankfully stories like this help to reinforce positive outcomes in other similar trials.

If a jury concluded that the defendant is guilty, I don't think judicial process reform would help. The fix would require at least public education. Reforming judicial process is hard, but fixing public perception seems way harder.

I've heard part of the problem is that the Post Office were allowed to conduct their own prosecutions. If the public prosecutor (Crown Prosecution Service) would have been involved, they would have asked the Post Office for any material which would have benefited the defence (this would be passed as disclosure). Because the Post Office were "victim" and prosecutor they never did this and the defence didn't get this material.

I have said in the past, only half in jest, that if I were guilty of a crime, I would want a jury trial, but if I were innocent, I would want to be tried by a judge, alone.

Private Eye is a national treasure. Britain would be much worse without it.

I believe a US judge is capable of personally acquitting a defendant -- a jury is only needed to convict.

Is that not the case in Britain?

This is called a "judgement notwithstanding the verdict" in US federal courts.

It's also one of the few instances where the _prosecution_ may appeal without violating double jeopardy protections.

[1] https://en.m.wikipedia.org/wiki/Judgment_notwithstanding_ver...

a judge can indeed overrule a jury verdict in favor of the defendent if he thinks the jury is clearly off it's rocker or that the prosecutor did not produce evidence that meets the standard of the law. Sometimes it's best to go with a bench trial based on situation. This might have been a good one, if the judge is known to have a solid understanding of technology vs the general public who barely know how to work a mouse and keyboard and have little training in deciding things base on a logical premise like "all software has bugs" as opposed to appeals to emotion.

Never underestimate the lack of technical knowledge of judge, jury, Senate or Congress (I'm American, I can't speak to the British system - but it seems similar).

People think software is infallible, despite their direct evidence otherwise.

Reply on news.ycombinator.com