Winboard Forum

Hi all

Here is a game played yesterday in a Skype tournament for the blind :

1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Qb6 6. Be2 cxd4 7. cxd4 Nge7 8. Na3
Nf5 9. Nc2 Be7 10. O-O Bd7 11. b3 a5 12. Rb1 Nb4 13. Nxb4 axb4 14. Ra1 O-O 15.
Bf4 Rfc8 16. Qd2 Rc3 17. Bd3 Rac8 18. Rfd1 Kf8 19. Bxf5 exf5 20. Rdc1 h6 21.
Qd1 Qa6 22. Bd2 Rxc1 23. Bxc1 Bb5 24. Ne1 Be2 25. Qd2 Kg8 26. Nc2 Bg5 27. Ne3
Bd3 28. Qxb4 Qxa2 29. Qxb7 Rxc1+ 30. Rxc1 Bxe3 31. Qc8+ Kh7 32. fxe3 Be4 33.
Kf1 Qxg2+ 34. Ke1 Bd3 35. Rc2 Bxc2 36. Qc3 Qg1+ 37. Kd2 Qd1# 0-1

It seems to me that moves played by black are very similar to the ones generated by Fritz (possibly Fritz 10 or 11) in the same positions. Please could you confirm or infirm the fact?

Time control was 1:45 for each player. Black's FIDE rating is around 1700.

Olivier

I don't have the engine you mention.

I ran a few other engines simultaneously on the game. I didn't pay particular attention to the time controls as one computer can be faster than anther and with the wrong engines, I only wanted to get a feel things. I think the ones I ran included Houdini 1.5, a version of Stockfish, Critter and Firebird, and a couple of others (Toga? Gaviota?). The moves played in the game seemed to fairly consistently match the moves suggested by the engines, though not always the same engine - the engines might disagree and typically propose a couple of moves from half a dozen engines. Sometimes the moves played would match one engine and sometimes another; they didn't consistently match one engine.

The exceptions were 11...a5 and 18... Kf8 which none of the engines I ran had predicted.

When I looked a few days ago, I was offline and unable to write here. Oddly enough it was the move Kf8 that I most suspected as being a computer style of move, despite none of the predicting it - which is why it aroused my attention. I couldn't understand black breaking off from what he was doing to defend against some threat that I couldn't see. But then maybe it is just that my chess isn't good enough. However if the threat was real, surely an engine would have defended against it too. Today I see it a little differently. I remember a game where I tried hard not to drop a pawn in the opening, only to suddenly notice that I had walled all my defending pieces in on the queenside and had little left to defend against a kingside attack. Maybe black got nevous about white's kingside majority and though Kf8 helped out.

I don't really know though how to detect cheating with an engine. I have learnt some opening lines from Stockfish versus Houdini game. I figure that when the weaker engine beats the stronger engine the opening line can't be too bad. So if you tried to detect whether I was cheating by using an engine it wouldn't be possible to tell from the opening line of a single game. Even at competitions there is suspect behaviour. At Gibralter a friend of mine had an opponent that went off to chat to friends in the public after every move. My friend was in the prize money and didn't want to take a risk and offered his opponent a draw at one point. Instead of the opponent studying the board position to decide, he went off to chat to his friends, came back and agreed the draw. Whether he chatted about football, the prize money or the board position only they know.

I don't think we should attach any importance to games where the players cannot be monitored.

Thanks for your time.

The case is over : the player has been interviewed by our comitee. About Kf8 he said it was meant to avoid tactics with Bh7 with check. He did not see the tactics behind Kg8 (avoiding Qxb4 with check), for him it was only a waiting move.

He was given the benefit of the doubt and was allowed to stay in the tournament.

Olivier