[gwnn]

I just don't know whether or not to take this conversation seriously if you cannot even admit that the doubt you are raising is just a distant second-best alternative. If you were accepted as the ultimate judge and jury for this case (or similar ones), how would you expect to go from "yes. they did transmit information. but I am only certain to 1%..99% degree that they did it consciously." to a narrower range? I suppose you would ask for more information. But what other kind of information do you have in mind? In these cases, you saw FS take the board from the tray and put it in the right place, even when the opponents took the tray, and similarly you can see FN adjust the card to be vertical/horizontal precisely, and the same for BZ, moving the bidding cards carefully. These are (hopefully you accept) at least indications that they did not just do it unconsciously. I already know that the "they felt compelled by this OCD-like urge" is a possible retort to this evidence, but similarly I can make up a retort for just about any other evidence that purports to prove conscious cheating. Can you think of some kind of evidence that could sway you from "um it could be 1%, it could be 99%. I don't know." ? If you can't, that is rather telling.

 nullve, on 2015-November-05, 20:02, said:

Hmm. I thought "innocent until proven guilty" was the very foundation of our justice system. But I see what you mean: it's not always practical.

Innocent until proven guilty is OK to establish whether or not the act took place. It is also OK to see whether it was negligence or premeditated. It is not OK to use for "all of this was unconscious, I cannot explain why I pulled the trigger." Nobody will say "ah, I guess we cannot know for sure whether he did it on purpose, maybe it is 1%, maybe 99%. Not guilty."

Quote

We have natural language, which tends to be symmetric. And mirroring seems to be a well-documented psychological phenomenon. As for 'semi-independently', consider the fact that when a board is finished, every single detail about partner's hand is known in principle.

Natural language evolved over several thousands of years. Here, we are talking about two people who are separated by a large screen and sent 1000 signals per year to each other.

Quote

The B-Z case is different from the F-S and F-N case in that people haven't even bothered to look for evidence that signals were transmitted; they (and now you) are willing to hang B-Z solely on evidence that signals were sent (emitted). So horses (Clever Hans, in particular) play no role here. I'm happy to be able to use Brogeland here instead of pigeons, since their role was which to debunk the view that since Balicki sometimes acts so unnaturally, he must be cheating.

You are quite wrong about this (or I missed this chunk of the conversation). For FN, all they did was show which code they used (unseen honour or singleton=vertical, otherwise=horizontal) and showed that it fit 84/86 (if I recall correctly). For F-S, they showed that the desired lead matched the code, and showed that sometimes the partner did indeed make that lead, but stressed that this second part is a weak part of the case and mostly it was just anectodal (look at this hand! isn't it a weird lead?), not seriously based on statistics. Only for EW did I see Kit Woolsey actually look at the partner's lead choices, using an expert panel. For F-S and F-N as far as I know no such endeavor exists.

[kuhchung]

gwnn you are a legend

[billw55]

So, all these signals are finally being discovered, that can be seen through that little door in the screen. It makes me shudder to think what was going on before screens.

[nullve]

I almost let this slip through:

 gwnn, on 2015-November-05, 15:59, said:

Anyone could reply to any accusation, even if there is tape of them committing the crime, that it was just unconscious. It is simply not a feasible way of going about life to accept that proposition or even entertain accepting it without some sort of additional evidence.

So you think I would accept introspective evidence that a bank wasn't robbed intentionally? I'm sorry, but if this is a contest in Wittgensteinian behaviourism, I win.

 gwnn, on 2015-November-06, 05:30, said:

In these cases, you saw FS take the board from the tray and put it in the right place, even when the opponents took the tray, and similarly you can see FN adjust the card to be vertical/horizontal precisely, and the same for BZ, moving the bidding cards carefully. These are (hopefully you accept) at least indications that they did not just do it unconsciously.

Yes.

Quote

Can you think of some kind of evidence that could sway you from "um it could be 1%, it could be 99%. I don't know." ? If you can't, that is rather telling.

But I can. And I think you owe me a couple of probabilities.

Quote

Natural language evolved over several thousands of years. Here, we are talking about two people who are separated by a large screen and sent 1000 signals per year to each other.

Yes, but here we are talking about ridiculously simple codes like

small gap = good hand in context
big gap = bad hand in context,

Maybe playing 1000 boards a year for 30 years isn't enough, but I'm not convinced.

Quote

You are quite wrong about this (or I missed this chunk of the conversation). For FN, all they did was show which code they used (unseen honour or singleton=vertical, otherwise=horizontal) and showed that it fit 84/86 (if I recall correctly). For F-S, they showed that the desired lead matched the code, and showed that sometimes the partner did indeed make that lead, but stressed that this second part is a weak part of the case and mostly it was just anectodal (look at this hand! isn't it a weird lead?), not seriously based on statistics. Only for EW did I see Kit Woolsey actually look at the partner's lead choices, using an expert panel. For F-S and F-N as far as I know no such endeavor exists.

I agree with you about the facts. But in the case of F-N and F-S the feeling was that some actions were inexplicable unless they were taking advantage of UI.

[gwnn]

You can? Then why don't you? I asked you a very simple question and you managed to reply to my entire post except the direct question. You admit that the adjustments are indeed indications that the actions are conscious but you still say that the probability in your opinion is somewhere between 1 and 99%. What kind of evidence would you need to at lease move you to 75-99% guilty? And does it mean that without these indications you would have gone for 1-10%?

I intentionally did not provide exact probabilities because I think it is more or less irrelevant whether it is 1% or 10% or 0.1%. I do consider the possibility quite remote for the reasons I explained already. But anyway, we simply cannot take this line of defense seriously. You already more or less admitted that you can't falsify this so from now on every cheater could retort to it and you would be forced to accept it or at best shrug under the "1%-99%" principle. Without cherdano's "if players behave in a manner indistinguishable from cheaters, we must consider them cheaters" principle we will be paralyzed in the mire of logical possibilities.

[gwnn]

In the FN case the process took about 15 minutes. There were whispers about the pair but apart from the "give partner a ruff instead of cashing my other ace", which is not even covered by the hypothesis (only the upgraded but vaguer version of "oh yes and sometimes they do it later too") from a few months before the circus starterd, there was no open speculation (maybe because people were busy hunting other unconscious witches). For FS I know there was a feeling (as I said myself) but mostly based on anecdotal stuff.

[gwnn]

Double post.

[gwnn]

Quote

So you think I would accept introspective evidence that a bank wasn't robbed intentionally? I'm sorry, but if this is a contest in Wittgensteinian behaviourism, I win.

I hope, honest to God, that you wouldn't. But based on your assessment of probabilities in the bridge case I'm not sure your definition of 'reasonable doubt' matches mine.

Sorry, posting from phone, afraid longer posts will disappear.

[nullve]

 gwnn, on 2015-November-06, 13:15, said:

You can? Then why don't you? I asked you a very simple question and you managed to reply to my entire post except the direct question.

I did answer your question, which was of the form 'Can you ____?'.

Quote

You admit that the adjustments are indeed indications that the actions are conscious but you still say that the probability in your opinion is somewhere between 1 and 99%. What kind of evidence would you need to at lease move you to 75-99% guilty? And does it mean that without these indications you would have gone for 1-10%?

75-99 %: B-Z confess
1-10 %: evidence that B-Z didn't take advantage of the signals

Quote

You already more or less admitted that you can't falsify this so from now on every cheater could retort to it and you would be forced to accept it or at best shrug under the "1%-99%" principle. Without cherdano's "if players behave in a manner indistinguishable from cheaters, we must consider them cheaters" principle we will be paralyzed in the mire of logical possibilities.

I hope cherdano would agree with me that trying to peek into players' minds is pointless. And I've already alluded to one thing I think cheaters would do: take advantage of UI. If there's no evidence that B-Z were doing that, I see no reason to think they're guilty. Of course, they could still have signalled illegally as part of some practical joke/bizarre experiment a la Reese.

[gwnn]

I was thinking something other than a confession. Hopefully you won't take this as a "moving of the goalposts". I just meant the question as "how would you go on and prove about an *uncooperative* pair that they are consciously sending illegal signals as opposed to bleeding unconscious tells and perhaps using them?" of course if opponents are open about their inner lives, it is trivial to prove something (even though using confessions is far from foolproof, at least knowing that the accused are honestly representing their version of events removes a major obstacle).

[nullve]

 gwnn, on 2015-November-06, 14:36, said:

I was thinking something other than a confession.

Ok, sorry.

Quote

"how would you go on and prove about an *uncooperative* pair that they are consciously sending illegal signals as opposed to bleeding unconscious tells and perhaps using them?"

Maybe by

1) proving that UI has been taken advantage of
2) ruling out Clever Hans-like effects

2) may be the hardest by far and a task for psychologists/neuroscientists rather than bridge players like me. But a code would take time to evolve, and probably longer if it was symmetric, so there must be a limit, as a function of the numbers of boards played, to how complicated/symmetric it could be.

[gwnn]

Ok that's fair enough and I think both of those would indeed strengthen the case (although I obviously consider the case strong enough as it is:)). I do think that proving use of a subtle min/max modifiers of AI is a particularly daunting task. Even from normal AI situations (say, 1M-2M;?) you will have a wide range of styles from experts about pass vs game try vs game. It will be tough to get a decent expert consensus. If I understand correctly, Kit told them something like "you have hand X and you have previously made sequence A and now decided on call B. Do you consider B to be a max or a min (or at lower level, eager/loathe to compete) in the context of A?" that's easier to adjudicate than saying "partner has just made call B (and, for the non-control group: and partner is min/max/unclear). What will you do now?" and now you would have to have quite a large pool of experts both in the control group (people who only know B but not the sub zone). I think you'll have a fair amount of noise in both camps and you will need a lot of data to establish anything.

Actually, using your point above to my advantage, what if you noticed that BZ spaced their cards relatively evenly and uncorrelated to their strength, and then suddenly, between two tournaments, they switched to clearly bimodal spacings strongly correlated to their hands? Surely that would more or less totally rule out unconscious tells, since these latter ones ought to have evolved gradually. I guess you could say that there is some sort of Punctuated Equilibrium at play but I don't know why that would be. Of course, this is pure speculation on my part, as we don't have a lot of videos.

Finally, I'm really suspicious about all this unconscious tell/unconscious interpretation bit in practice. There is a clear rule telling us not to vary our mannerisms at the table and players should consciously try to abide it. I think someone who allows this subconscious universe of mathematical code to spontaneously (if gradually) evolve must not have paid a lot of attention to ignoring partner's tells and avoiding creating them too. At the same time, I feel a qualitative difference between allowing oneself to yield to some urge and actively deciding to do it (say, not calling the director when one overhears something vs sneaking up behind people to try to catch some indications). Anyway I guess we said most of what we had to say for now. I don't think your scenario is very realistic and you think it is. We would need a lot of data to see what kind of codes people inadvertently develop (while expressly trying not to develop one) whilst being separated by a wooden wall and stay more or less completely quiet. Unfortunately as far as I know this is totally undocumented so we can just guess.

[nullve]

 gwnn, on 2015-November-08, 09:18, said:

Actually, using your point above to my advantage, what if you noticed that BZ spaced their cards relatively evenly and uncorrelated to their strength, and then suddenly, between two tournaments, they switched to clearly bimodal spacings strongly correlated to their hands? Surely that would more or less totally rule out unconscious tells, since these latter ones ought to have evolved gradually.

Good point.

[gwnn]

On BridgeWinners there's a lot of cute little Vine-worthy moments popping up now. Like this one: https://www.youtube....1TDXA1lQ&t=2494

The sequence is:
1. far pass
2. look at cards again
3. move the pass a bit closer
4. look smooth

Of course, these little few-second videos prove nothing, and we are all biased observers, and they could all add up to nothing, and we would have to conduct a double-blind study (hiding the faces or use a neuralyzer from MiB), but they are still funny to watch. It doesn't look like an unconscious gesture to me, but all the above apply of course.