Originally posted by Toe
A little ramble to show just how difficult this is likely to be. Again, top show all who have signed up for possible inclusion.
Determine a policy of how accusations are submitted:
Where this accusation is made (not a public forum for a sta ...[text shortened]... same thing but related, as it requires the same analysis skills.
Detecting engines isnt that hard, other chess sites do it all the time. 95% match up to an engine isnt proof, in some chess games, those that are short, tactically forced will get a very high match up to any engine. However drag a game out for 40 moves and the chances of anyone matching something like Fritz8 for each move is slim, the chances of them matching up 95% to fritz8 for 20 games is something in the area of 1 in 10 to the power of 120 (10 with 120 zeros after it).
Stronger players will match up a lot more I wouldnt be surprsed to see some hitting 60% in some games but normally the match up is about 30%. This is because of the difference between fritz and humans, fritz will just try to crunch numbers and find good variations, the way humans select the candadate moves is firstly intuistion then the work the variations out. Theres many factors which effect the human chess play, the main two being style and rating. If you think of all the GM's in the past notice how diverse their style is, even the ones of simlar styles, their moves would and have never matched up 100% to any other GM,IM or FM game after game. Its this that allows us to say with certainty that people who are throwing out 95% match ups to any engine game after game are using that engine.
You are right about the openings, they shouldnt be taken into account when checking for engines, people can get openings from books, databases etc.
Engames is another special mater, I would totally expect a strong human player to be able to take a won endgame and win it against any engine since the game if often quite simple and given sufficent time any human can caculate the best lines. Yet there would still be a difference in play from the fritz caculation/tablebase and human play.