catching cheaters who use computers

As one who is new here and an old member of ICC, I don't know how effective this site is in rooting out players who cheat by using computers. I know ICC is pretty effective, but I don't know how they do it. So just for discussion sake, here is how I would do it.

Any player below expert level is highly unlikely to play a full game without at least one serious error - in my present tournament, although I have 13 wins and one loss, there is not a single game where post-game analysis did not show me making a losing move or throwing away a win.This ensures catching "smart" cheaters who use a computer, but never choose the best move proposed by the computer - computer moves yielding an advantage of say -0.30 will beat anyone below Master level.

So if a player rated say 1600 plays five games in a row without computer analysis showing any game-losing move or throwing away a won game, it is highly probable that he is using a computer. In a similar vein, if every move in a game is shown to always be among the 4 best computer moves, the player is playing above his head and is probably cheating.

However I can concede that it possible in a single game (especially if it is short or if the opponent made a game-losing blunder early in the opening) for an average player to play a game without such a move. That is why a decision  that a player is using a computer should usually be based on more than one game.

It is probably possible to find better critieria than the above by means of statistical analysis (what IS the probability of say a 1700-rated player to play an error-free game?), and I would hope that is what the chess sites are using. They could have a table showing the probability vs the OTB rating (corrected for the number of games).

What do you think?

-> http://www.chess.com/forum/view/community/chesscom-policy-on-cheating

 Likewise for 'points' fixation.

erik also wrote: