Christopher Kreuzer wrote:Georgiev says: "It made a strong impression on me that at some point, when the position was repeated twice, he thought for about three seconds and diverted, avoiding a three-fold repetition!" Is that relevant at all?
It was a technique taught in the former Soviet Union, that you repeat twice even you don't intend to draw. The idea was to disrupt your opponent's concentration and perhaps humiliate them a bit. Equally it's a way of proposing a draw without taking the risk that your opponent might accept.
Christopher Kreuzer wrote:
Houdart (the author of the chess engine Houdini) has this to say "It's not very useful to focus on individual move choices that appear unlikely (like Qd1 in the blitz game), it's the combined evidence of ALL the moves that is compelling."
The combined evidence is that most of his moves match those of computer engines. But in the absence of a demonstrated method of consulting these engines during play, you can only fall back on the explanation that the consultations took place before play and that he's found a training method and positional eye that matches the top engines in many positions. The family of engines from Rybka onwards have supplemented brute force calculation with sophisticated positional fine tuning, so if you knew or reverse engineered what that was, your move choices would match many of those of the engines because you were essentially doing the same calculation.
If he is cheating, then he has to have a method of communicating the board position or the latest move played to either an engine on his person or in the hands of third party accomplice(s). He also has to receive back the move recommended. Using the evidence of the rapid play tournament, all this has to be done at ten seconds a move.
