Cheating in chess

Discuss anything you like about chess related matters in this forum.
User avatar
JustinHorton
Posts: 10364
Joined: Mon Aug 04, 2008 10:06 am
Location: Somewhere you're not

Re: Cheating in chess

Post by JustinHorton » Wed Dec 23, 2020 12:34 pm

Am I right in thinking the individual concerned did not in fact start?
"Do you play chess?"
"Yes, but I prefer a game with a better chance of cheating."

lostontime.blogspot.com

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Cheating in chess

Post by Richard Bates » Wed Dec 23, 2020 12:53 pm

JustinHorton wrote:
Wed Dec 23, 2020 12:34 pm
Am I right in thinking the individual concerned did not in fact start?
I believe he tried to play his first game without turning Zoom on. And then following further claims about various technical issues was eventually defaulted. Was also paired in round 2 and opponent was subsequently given a walkover.

Matt Bridgeman
Posts: 1077
Joined: Wed Aug 03, 2016 9:21 pm

Re: Cheating in chess

Post by Matt Bridgeman » Wed Dec 23, 2020 12:59 pm

He did cause quite a bit of chaos in the qualifier, including beating England’s top (only?) current prodigy. I presume if he did now want to play in the 4NCL Congress or other such events he would be okay, because they’re different events?

Ian Thompson
Posts: 3551
Joined: Wed Jul 02, 2008 4:31 pm
Location: Awbridge, Hampshire

Re: Cheating in chess

Post by Ian Thompson » Fri Dec 25, 2020 6:51 pm

The player in the attached PGN file has had their lichess account marked as "This account violated the Lichess Terms of Service" in a web browser and "This player uses chess computer assistance" in the app. That had happened by 18 December, and may well have been a week or so earlier if it's correct that players with accounts marked in this way are paired against other players who lichess think have cheated.

Whilst I have anonymised the player and his opponents in the file, I do know who he is.

What do people think? Are the games reasonable grounds for concluding that he used a computer?
computer -assistance-player.pgn
(18.62 KiB) Downloaded 133 times
If it's correct that lichess don't assess casual games for cheating, his account was marked after 9 rated games at the most, and perhaps after no more than 5 or 6.

Joseph Conlon
Posts: 339
Joined: Thu Jun 06, 2019 4:18 pm

Re: Cheating in chess

Post by Joseph Conlon » Sat Dec 26, 2020 11:40 am

I'm not offering an opinion either way regarding the games and the player (I haven't looked at them).

But on casual games - my understanding is that a player can get a lichess ban for engine use in casual games. My understanding (which is based on reading a lichess discussion thread) is that lichess allows engine use in casual games provided both players consent to it and and are aware of it. So if you set up a casual challenge match with a specific player, that won't be flagged because you could have arranged with them to use an engine. But if you take casual games from the lichess pool and use engines in those, you can get an engine mark from that.

Ian Thompson
Posts: 3551
Joined: Wed Jul 02, 2008 4:31 pm
Location: Awbridge, Hampshire

Re: Cheating in chess

Post by Ian Thompson » Sat Dec 26, 2020 12:48 pm

Joseph Conlon wrote:
Sat Dec 26, 2020 11:40 am
So if you set up a casual challenge match with a specific player, that won't be flagged because you could have arranged with them to use an engine.
Most of this player's casual games, other than those against lichess's computer, were actually in tournaments lasting 7 or 8 weeks where players were paired with an opponent each week by the organiser and then given the rest of the week to arrange with their opponent when they wanted to play the game. I think most of the players would have been casual players who had no experience of tournament chess.

Although I wasn't aware of the detail you've provided I think that means I'm right in assuming the games wouldn't have been subject to computer assistance checks.

Chris Rice
Posts: 3417
Joined: Thu Nov 08, 2012 5:17 am

Re: Cheating in chess

Post by Chris Rice » Sat Dec 26, 2020 1:03 pm

A three minute tournament the Levitov Chess Christmas Cup V was held on Lichess recently. The winner, GM Grizzly Bear 79 rated 2753 won it with a performance rating of 2942. He has since been banned and his account closed, the player concerned was allegedly none other that Parham Maghsoodloo but was he actually cheating? Here is his last round game and here is another from the same tournament. See anything suspicious? The main problem of course is that Lichess won't release the basis for their decision so a guy who is already pretty much a super GM has been branded a cheat without apparently any supporting evidence. My guess is that this will only cause more confusion and lose the trust of the professional players who will be less inclined to play on that server.

Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: Cheating in chess

Post by Roger de Coverly » Sat Dec 26, 2020 1:14 pm

Chris Rice wrote:
Sat Dec 26, 2020 1:03 pm
Here is his last round game
If they detected engine assistance it could well have been in the opening and early middle game. Those Botvinnik English positions arise time and time again and are usually very similar. Plenty of ground for in depth investigation of middle game ideas and let's repeat, completely legal if carried our before the game.

Chris Rice
Posts: 3417
Joined: Thu Nov 08, 2012 5:17 am

Re: Cheating in chess

Post by Chris Rice » Sat Dec 26, 2020 2:04 pm

Roger de Coverly wrote:
Sat Dec 26, 2020 1:14 pm
Chris Rice wrote:
Sat Dec 26, 2020 1:03 pm
Here is his last round game
If they detected engine assistance it could well have been in the opening and early middle game. Those Botvinnik English positions arise time and time again and are usually very similar. Plenty of ground for in depth investigation of middle game ideas and let's repeat, completely slegal if carried our before the game.
My guess is that Lichess rely on the engine-matching software produced by Ken Regan without understanding it. Indeed who could? For example an article entitled Predicting Chess and Horses shows that his assumptions and statistical model have changed over time and are steeped in so much jargon very few could challenge it without the risk of looking like idiots. A particular extract from this article is especially relevant to this case:

"For the elite class—2600 to 2800—in the same kind of many-choice positions, the new model does even better. Much more data on elite players is available, so I have 49,793 such positions faced by them:

Whereas elite players found the best move in 30.85% of these difficult positions, my new model finds their move in 34.64% of them.

Over all positions, the average prediction gain ranges from about 1 percentage point for the lowest players to over 2% for masters. These gains may not sound like much, but for cheating tests they give prime value. The reasons are twofold:

The gained predictions are all against their finding the computer’s move, so the act of finding the best move is more exposed.

The standard deviation is typically 3–5 percentage points depending on the volume of moves. Thus the prediction gain can enhance the measured z-score by upwards of 0.5 or more.

Initial applications in recent cases seem to prove this out more often than not. Of course, the larger purpose is to have a better model of human chess play overall."

So it appears to be that what is being put forward hear is a model to predict the human thought patterns of 2600-2800 players and to guess their moves correctly which apparently the model gets right one out of every three times. So the model gets the predictions wrong two thirds of the time? Anyway based on this we get a further analysis that deviations from this can indicate cheating. Now I'm very impressed by the polynomials and the Aitkin curves not that I've got a clue what they are but what looks to be going on is that games are being put into some sort of "black box" which know one understands except the creator and we simply have to trust the results that come out of it are credible. This is fine if we all buy in to this but if we have doubts who can explain how the conclusions will have been reached? Perhaps only an actual Court case will progress the matter but unless a substantial amount of money is involved I can't see that happening any time soon so these cases will rumble on and on I guess.

Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: Cheating in chess

Post by Roger de Coverly » Sat Dec 26, 2020 2:19 pm

Chris Rice wrote:
Sat Dec 26, 2020 2:04 pm
So it appears to be that what is being put forward hear is a model to predict the human thought patterns of 2600-2800 players and to guess their moves correctly which apparently the model gets right one out of every three times. So the model gets the predictions wrong two thirds of the time?
Where humans match engine moves in openings and middle games, that may be evidence of engine consultation. Unless there's additional witness evidence as in over the board chess, that's not conclusive evidence that the consultation took place illegally during the game rather than legally before it.

The games that are more suspicious to my eyes are where one of the players saves or wins what looks a very dubious position. It used to be and perhaps still is that engines were very good at heroic defence. That is they could land themselves in what appeared suspect positions by lack of positional understanding but analyse their way out of trouble finding repeated "only" moves. Humans in similar circumstances might well try a double or quits style of defence just trying to set traps.

Joseph Conlon
Posts: 339
Joined: Thu Jun 06, 2019 4:18 pm

Re: Cheating in chess

Post by Joseph Conlon » Sat Dec 26, 2020 2:20 pm

I would assume that any decision to put an engine mark on a super-GM account would go right to the top of lichess and would not be done without a lot of thought and human input. I can imagine algorithmic bans of 1900-rated players being done very quickly, but its hard to imagine a site banning someone of that strength without there being a lot of human eyes on the decision.

Roger Lancaster
Posts: 1910
Joined: Tue Mar 17, 2015 2:44 pm

Re: Cheating in chess

Post by Roger Lancaster » Sat Dec 26, 2020 2:30 pm

Joseph Conlon wrote:
Sat Dec 26, 2020 2:20 pm
... its hard to imagine a site banning someone of that strength without there being a lot of human eyes on the decision.
Really?

User avatar
John Upham
Posts: 7179
Joined: Wed Apr 04, 2007 10:29 am
Location: Cove, Hampshire, England.
Contact:

Re: Cheating in chess

Post by John Upham » Sat Dec 26, 2020 3:31 pm

Roger de Coverly wrote:
Sat Dec 26, 2020 1:14 pm


and let's repeat, completely legal if carried our before the game.
Is there a suggestion that what you are describing is not legal?

I'm curious to know who might not think it was OK to do this?
British Chess News : britishchessnews.com
Twitter: @BritishChess
Facebook: facebook.com/groups/britishchess :D

Jonathan Rogers
Posts: 4640
Joined: Tue Nov 18, 2008 9:26 pm

Re: Cheating in chess

Post by Jonathan Rogers » Sat Dec 26, 2020 3:32 pm

Roger Lancaster wrote:
Sat Dec 26, 2020 2:30 pm
Joseph Conlon wrote:
Sat Dec 26, 2020 2:20 pm
... its hard to imagine a site banning someone of that strength without there being a lot of human eyes on the decision.
Really?
Yes. Not saying it is the right decision, but they are known to be extra careful with titled players, let alone with world class players, so plenty of human attention surely.

User avatar
JustinHorton
Posts: 10364
Joined: Mon Aug 04, 2008 10:06 am
Location: Somewhere you're not

Re: Cheating in chess

Post by JustinHorton » Sat Dec 26, 2020 3:51 pm

"This decision looks wrong on the available evidence, so let us assume there is better evidence not available to us" is not an approach I would like to recommend.
"Do you play chess?"
"Yes, but I prefer a game with a better chance of cheating."

lostontime.blogspot.com

Post Reply