Life In 19x19

Bojanic wrote:

[*] In online tournaments, if matches are not monitored or recorded by webcams, it will always be possible to cheat in a subtle way, like using a strong software just once or twice during the game. This will be undetected by statistical methods, just like when athletes are injected just enough EPO/growth hormone/testosterone/whatever so that their blood or urine analysis looks normal; these athletes should be considered innocent until material proof is found.

I don't think that camera is enough, esp if player has help. We have seen player in China who cheated in live game, which is substantially more difficult.
Player could open window, or on his phone, and watch online game analysis.
Or receive moves via message system.

One solution would be to have those matches organized at the local Go club. So there would be up to three computers available for such event, and an internet connection would be necessary. There would be a club representative (club president if he is not part of the team) that would act as a scrutineer, in charge to monitor the event. Ideally, all matches would be played at the same time, so one scrutineer can oversee all of them in one shot. If players from the same country are not from the same club/city, then that means organizing having two events and two scrutineers.

John Fairbairn wrote:There is an ambivalent attitude towards cheating. I think there may be a biological explanation. Cheating among primates seems to be a way that (say) physically weaker individuals can overcome physical stronger ones. So we recognise the need to cheat and can even applaud especially clever scams. Against that we have notions of fair play that seem to be imposed by moralists from outside and which can vary from culture to culture. All this, plus the threat of legal action, seems to put a severe brake, in various sports, on how harsh sanctions should be.

To illustrate how the questions of cheating and fairness are biologically or culturally "hardwired" in our brains, there is a cognitive test, the Wason Task (see https://www.socialpsychology.org/teach/wason.htm ) which, when presented fairly abstractly using cards, numbers, and letters, is not so easy, a lot of people get it wrong. However, when presented as a problem of detecting cheating, almost 100% of people get it right.

pnprog wrote: One solution would be to have those matches organized at the local Go club. So there would be up to three computers available for such event, and an internet connection would be necessary. There would be a club representative (club president if he is not part of the team) that would act as a scrutineer, in charge to monitor the event. Ideally, all matches would be played at the same time, so one scrutineer can oversee all of them in one shot. If players from the same country are not from the same club/city, then that means organizing having two events and two scrutineers.

This sounds like a nice idea. This should drastically reduce the chance that cheating is attempted. Even if cheating occurs and is tolerated, at least there are multiple people involved, which is an important hindrance.

On the other hand, not everyone lives near a Go club that also has a laptop and internet. I, for instance, have for a long time lived an hour apart from the nearest club.

maf wrote:

pnprog wrote: One solution would be to have those matches organized at the local Go club. So there would be up to three computers available for such event, and an internet connection would be necessary. There would be a club representative (club president if he is not part of the team) that would act as a scrutineer, in charge to monitor the event. Ideally, all matches would be played at the same time, so one scrutineer can oversee all of them in one shot. If players from the same country are not from the same club/city, then that means organizing having two events and two scrutineers.

This sounds like a nice idea. This should drastically reduce the chance that cheating is attempted. Even if cheating occurs and is tolerated, at least there are multiple people involved, which is an important hindrance.

On the other hand, not everyone lives near a Go club that also has a laptop and internet. I, for instance, have for a long time lived an hour apart from the nearest club.

This is true. I think that requesting the local go club to have an internet connection is difficult. Of the two go clubs I visited in France, one would meet in a café, the other in a local provided by the city... An alternative would be to play at the scrutineer place, or some place he choose.
On the other hand, if one consider than the players in PGETC are amongst the best in their country, and that they are all affiliated to their national Go federation, then it is very likely they are located near one Go club. Otherwise, there are still the possibility to have a second scrutineer move to that player location and overview the match on the behalf of the first scrutineer.

But ideally, one scrutineer, one place, one time, club's computers, and eventually additional club members that happen to be there to watch the game would make cheating quite difficult.

Bill Spight wrote:To illustrate how the questions of cheating and fairness are biologically or culturally "hardwired" in our brains, there is a cognitive test, the Wason Task (see https://www.socialpsychology.org/teach/wason.htm ) which, when presented fairly abstractly using cards, numbers, and letters, is not so easy, a lot of people get it wrong. However, when presented as a problem of detecting cheating, almost 100% of people get it right.

Thank you for the reference! I knew about the Wason task but not about experiments with variants of it. There is some very interesting discussion of the various experiments here and it sounds like people still have different ideas about how to interpret the results.

I believe the ideal of no false positives is really unhealthy. It almost guarantees there will be a lot of false negatives (i.e. only a few cheaters are caught). Since so few are caught, they receive the whole ire of the community and likely exemplary (harsh) treatment. Since the treatment is so harsh, the pressure that false positives are avoided increases further. The whole control regime also ends up expensive and invasive. But since everyone knows most get away with cheating this completely undermines trust, increases pressure to cheat yourself (think cycling) despite having an expensive and invasive control system in place.

I would much prefer unsound evidence (not too unsound, but the 98% in the case), occasional false positives and lenient punishment, that keeps the control systems simple and overall trust (we catch most cheaters + few innocents) higher.

dfan wrote:

Bill Spight wrote:To illustrate how the questions of cheating and fairness are biologically or culturally "hardwired" in our brains, there is a cognitive test, the Wason Task (see https://www.socialpsychology.org/teach/wason.htm ) which, when presented fairly abstractly using cards, numbers, and letters, is not so easy, a lot of people get it wrong. However, when presented as a problem of detecting cheating, almost 100% of people get it right.

Thank you for the reference! I knew about the Wason task but not about experiments with variants of it. There is some very interesting discussion of the various experiments here and it sounds like people still have different ideas about how to interpret the results.

Thanks for the reference. I note that the supposed cheat detection task which was discussed still involves turning over cards, which is still to some degree abstract. Perhaps that is why only ¾ of people get it right. The cheating detection research I heard of years ago where something like 98% of people got it right involved actual social simulations in which they played the role of cheat detector. We need to remember our Piaget. Most people have not mastered abstract operations.

The Wason Task is also pertinent to our current discussion because it shows how people look for confirmatory evidence, believing it to be relevant.

tapir wrote:I would much prefer unsound evidence (not too unsound, but the 98% in the case), occasional false positives and lenient punishment, that keeps the control systems simple and overall trust (we catch most cheaters + few innocents) higher.

I understand your sentiment, but this will never be accepted. It is literally against a core value of Western society.

Also, realize that 98% means we would 'catch' 100 (one hundred) honest people at the EGC. That's around 500 people playing around 10 games each, and 2% of these would be 'detected'. That sounds impractical, at least unless being marked as a cheater loses all (rightfully) associated stigma and consequences.

tapir wrote:I would much prefer unsound evidence (not too unsound, but the 98% in the case)

IMO, this 98% number should never have been published. Many people seem to interpret it along the lines of "there's a 98% chance the player was cheating in this game", which is not at all what it means.

HermanHiddema wrote:

tapir wrote:I would much prefer unsound evidence (not too unsound, but the 98% in the case)

IMO, this 98% number should never have been published. Many people seem to interpret it along the lines of "there's a 98% chance the player was cheating in this game", which is not at all what it means.

At least it should have been accompanied by the even more scandalous fact that Carlo played 100% of his moves on the intersections of the board, just like Leela did too!

P.S. I got the Wason test right, I suppose I'm an analytical/logical type, plus it was as a break from refactoring a giant mess of if-then-else spaghetti code which probably counts as priming.

tapir wrote:I believe the ideal of no false positives is really unhealthy. It almost guarantees there will be a lot of false negatives (i.e. only a few cheaters are caught). Since so few are caught, they receive the whole ire of the community and likely exemplary (harsh) treatment. Since the treatment is so harsh, the pressure that false positives are avoided increases further. The whole control regime also ends up expensive and invasive. But since everyone knows most get away with cheating this completely undermines trust, increases pressure to cheat yourself (think cycling) despite having an expensive and invasive control system in place.

If people cheat on an online server in casual play, even if the games are rated, I am afraid that not much will be done, either in cheat detection or punishment, because nobody cares that much. Chess players, please correct me if I am wrong, but my impression is that that is pretty much the case now with casual online chess games, even though cheating detection using engines is quite well developed in chess.

However, I doubt if lessening the punishment for online cheating in tournaments, such as simply by forfeiting a single game, would be acceptable by the community. True, if we accepted a number of false positive results, so that people were punished because they were probably cheating, then that punishment would probably lessen. But that could also lead to long standing enmity between those who believed that a person cheated in a particular game and those who believed that he was unjustly convicted. Also, having, say, three or four games per tournament forfeited because of suspicions of cheating would be disruptive of tournament play itself. Honest players might fear playing too well, lest they lose a game because they are accused of cheating.

I would much prefer unsound evidence (not too unsound, but the 98% in the case), occasional false positives and lenient punishment, that keeps the control systems simple and overall trust (we catch most cheaters + few innocents) higher.

IMX, accusations of wrongdoing (cheating included, OC) do not improve overall trust. To the contrary, they erode it. Accepting more false positives would lead to more accusations, simply because they would be more likely to work.

As for the 98% matching evidence, you must understand that matching one of a bot's top three choices was chosen in order to generate impressive matching numbers, not through any theory of how a player might have cheated. (This motive may have been unconscious.) And restricting the possible matches to the fifty moves between moves 51 - 100 is also suspicious. In addition, it is confirmatory evidence instead of disconfirmatory evidence. IOW, it is not just unsound, it is crap.

Uberdude wrote:P.S. I got the Wason test right, I suppose I'm an analytical/logical type, plus it was as a break from refactoring a giant mess of if-then-else spaghetti code which probably counts as priming.

Even professors of logic have gotten it wrong.

Uberdude wrote:At least it should have been accompanied by the even more scandalous fact that Carlo played 100% of his moves on the intersections of the board, just like Leela did too!

Actually, as Italian statistical analysis showed, it is few percents less.
And if we have even 2% uncertainty, 100 players would be wrongfully accused and sanctioned on EGC, as another statistical analysis showed.

PS I would be very interested to know if there is some statistical analysis if 4d player could play entire game like 7d?
Not like, few moves well spotted in one game, but every move. In two games.

BTW
When I was first time on WAGC, Kato Masao analyzed my game against player from Taipei.
Since I am crap at remembering moves, I wrote them on paper, and it was difficult to find where next move would be. (ah those tenukis...)
And while I was looking for move, Kato would point out where move should be played. My opponent's moves were usually there, mine nowhere near.
I realized that my opponent might be somewhat stronger than me.
If you wonder about result, opponent made mistake in late middle game, that even I saw, I attacked him and managed to narrow difference, and lost only by komi (Kato said I maybe had some chance of winning).

Bill Spight wrote: As for the 98% matching evidence, you must understand that matching one of a bot's top three choices was chosen in order to generate impressive matching numbers, not through any theory of how a player might have cheated. (This motive may have been unconscious.) Also, restricting the possible matches to the fifty moves between moves 51 - 100 is also suspicious. In addition, it is confirmatory evidence instead of disconfirmatory evidence. IOW, it is not just unsound, it is crap.

So you think we should also include in analysis opening phase, which could easily show 100% similarities to Leela in many games?
BTW, opening usually lasts up to move 30, not 50 as you claim.
And why include endgame in analysis? (which in analyzed games started around move 120).
So it is not 50 moves, rather around 90.
And I am curious, what method would not show that Metta's two online games are very different to his two live games?

Theory how player might cheated is very simple:
- players plays his moves into program (Leela) at the same time as they are played.
- there is no need to use it in opening, so fuseki can be played on own,
- middle game is played with help of program, since it is most important part of the game.
Move is selected among top suggestions.
- after securing lead in middle game, endgame is played more leisurely, with more own moves, while lead is monitored and maintained by watching computer play.
Good enough theory?

One unjust sentence of murder is too many. One unjust sentence of doping is too many. One unjust sentence of cheating is too many. The former means loss of freedom (and somewhere death). The latter two can mean a lifelong destroyed reputation.