Vass wrote on 06/13/14 at 09:14:29:


I saw it and I completely agree with you!

Still cannot see it. When I am playing well I dont blunder and a lot of my moves match computer moves.

Will someone be along to arrest me now?

MartinC wrote on 06/12/14 at 12:41:00:


Condemn is perhaps the wrong word. Detect perhaps and what they are doing is legal for ICCF play but not for Over the Board. These are games where it is known that the probability of a chess engine being consulted during the game is reasonably high, but it isn't being done on the basis of just reproducing engine suggestions.

I wonder if you'd want to condemn them though? I doubt it. There's no way that anyone will be able to sit consoluting a computer in the same sort of way over the board.
(Until we get direct brain interfaces or something, but a while before that )

Regan did do a study of games played using Advanced Chess or centaur chess.

Here's a link:

http://www.cse.buffalo.edu/~regan/chess/fidelity/FreestyleStudy.html

Basically, the match-up rate is high, but not ridiculously high. What is incredibly low is the error rate. These can be seen by clicking on the links within the article.

You'd never, ever pick up just the odd computer move. Far too small a signal to find

The best feature of doing this is the negative impact - if someone does suddenly hit really good form then you know they're being checked so you can hopefully avoid worries that they're just parroting computer moves somehow.

Smyslov_Fan wrote on 06/11/14 at 14:55:55:


For many practical purposes that is a worthless test. It gives a possible means of making accusations against those who might come tooled up with a miniature camera and a shoe based computer. It doesn't tell you about someone who plays most games or even most moves legitimately but from time to time gets the odd hint or assessment from an engine during a game. So what is the purpose of monitoring every game and making potential accusations unsupported by any physical evidence?

I'm still curious as to what the statistics are for top correspondence players in events where computer use is allowed. Those who play at that level do not consider it adequate to simply take an engine's move or assessment at face value.

Don't think you'd have caught those French players this way either. Or some lower tech anologue. You're just relying on having 2/3 people in the loop making it all much more vunerable.