Cheating in chess

[Adam Raoof]

Wadih, great posts. Thanks for the info.

I like the combination of screenshare, webcam and pgn review. I think this would significantly reduce the level of opportunist cheating.

What does the arbiter see on the Zoom screen in this instance? I am currently running a congress with 40 players and it's already a lot to keep an eye on

[Wadih Khoury]

What does the arbiter see on the Zoom screen in this instance? I am currently running a congress with 40 players and it's already a lot to keep an eye on

I haven't asked, but as I understand it, they had 9 arbiters for the whole week, with about 200 children and therefore screens each day. Not sure if all 9 were simultaneously available or if they rotated. I strongly suspect that the "fair play breaches" that were detected were post the event. After, the high profile case was only disqualified post event.

Alternatively, if not enough manpower is available to watch live all screens, there is still the option to record everything in case the pgn review throws something suspicious.

[Jonathan Bryant]

The approach taken was as follow:

• All children had to be on camera at all times

Hello Wadih.

Is that first clause really true? Nobody was allowed to leave their computer? Not to use the toilet?

How long were the games?

[Wadih Khoury]

The approach taken was as follow:

• All children had to be on camera at all times

Hello Wadih.

Is that first clause really true? Nobody was allowed to leave their computer? Not to use the toilet?

How long were the games?

Well, seeing it only from the perspective of my son, I can't say how strictly it was enforced. I did ask the question before the start of the tournament to make sure he doesn't inadvertently break a rule, and I was told that he was free to do what he wants in between rounds, but had to stay during the game.

This was made easier by the fact that it was a 10+3 Swiss if I recall correctly, with 2 or 3 minutes between the end of the last game and the start of the next round, with one bigger 10 minute pause after round 5 of 9. Obviously for longer games, toilet breaks are unavoidable, and whether at home or OtB a den of temptation

According to the federation, they had reluctantly to forfeit the game of some juniors who were not answering the arbiters (presumably because they were not showing on camera).

[Jonathan Bryant]

Well, seeing it only from the perspective of my son, I can't say how strictly it was enforced. I did ask the question before the start of the tournament to make sure he doesn't inadvertently break a rule, and I was told that he was free to do what he wants in between rounds, but had to stay during the game.

This was made easier by the fact that it was a 10+3 Swiss if I recall correctly, with 2 or 3 minutes between the end of the last game and the start of the next round, with one bigger 10 minute pause after round 5 of 9. Obviously for longer games, toilet breaks are unavoidable, and whether at home or OtB a den of temptation

Thanks - that's a bit shorter than I'd imagined. Workable with that time control but not, as you say, replicable for longer time controls.


Toilet breaks in OTB chess are not really a den of temptation as you put it. I mean, I'm sure people have used them to cheat but it's not at all on the same level as happens online.

[Robin Nandi]

I compiled some engine matching statistics for the recent 4NCL Online Congress. These are based on the pgn download of the games from chess-results and engine analysis on lichess.

Engine matching stats for top 10 finishers:

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Performances of top 10 finishers ordered by % of moves matching Stockfish first choice:

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Ian - I'd be interested to know if these match what you are finding using pgnspy?

[NickFaulks]

.
Of note, is that the Federation's communication (here in French: http://echecs.asso.fr/Actu.aspx?Ref=12770 ) reaffirm their belief that " if a child breaks the rules, it is unlikely at their own initiative, but at the behest or pressure of some adults".

Not sure I agree, but what do other people think?

I tend to hold parents responsible for the behaviour of their children, but I think "unlikely" goes too far.

[Mick Norris]

I would imagine that those of us with children would think it was quite likely to be the the children's own initiative

If there's a correlation between a group of children cheating and a chess coach that would be disturbing

[NickFaulks]

Performances of top 10 finishers ordered by % of moves matching Stockfish first choice:

https://docs.google.com/spreadsheets/d/ ... sp=sharing

The Adams Jaden game at the bottom is impressive. Without doing the sums, I'd say there is convincing statistical evidence that the players were deliberately avoiding Stockfish moves.

[Pete Heaven]

....alternatively Adams-Jaden could be viewed as a 51 move game in which neither side made a mistake or blundered (4 inaccuracies by each, according to what lichess says).

[Ian Thompson]

Ian - I'd be interested to know if these match what you are finding using pgnspy?

I only ran the game quoted in this thread through PGNSpy and subsequently all games from the tournament of the winner of that game.

You (lichess) have 72% of moves matching Stockfish's 1st choice. PGNSpy has 58%.

For the individual games the percentages are:

Code: Select all

lichess                    PGNSpy
84% of 37 moves            63% of 16 moves
81% of 30 moves            56% of  9 moves
73% of 30 moves            58% of 26 moves
50% of 20 moves            50% of  8 moves

The most striking difference is the number of moves each program thinks worth evaluating, which probably makes the percentage differences meaningless. For PGNSpy the numbers are determined by parameters that are known and could be changed. For lichess I assume they are unknown.

[Kevin Thurlow]

"If there's a correlation between a group of children cheating and a chess coach that would be disturbing"

That has been known...

[NickFaulks]

....alternatively Adams-Jaden could be viewed as a 51 move game in which neither side made a mistake or blundered (4 inaccuracies by each, according to what lichess says).

Yes, you can make the statistics say that if you prefer. Your choice.

[Pete Heaven]

I didn't make them say it Nick. They do say it. And it was the 'worst' game, Stockfish-top-line-wise, that you picked on. And, given that all players must know by know by now that Prof Regan's software reports Stockfish's top line, it may be a more reliable guide to engine use, along with centipawn loss. That said, even if you add human analysis of games, which is sadly absent, it's all supposition on both sides.

[Roger de Coverly]

. And, given that all players must know by know by now that Prof Regan's software reports Stockfish's top line, it may be a more reliable guide to engine use, along with centipawn loss.

I think you would struggle to prove the assertion that because players avoided Stockfish's first choice, this means they knew what it was.