Originally posted by blowfish
Ok, so I've heard "instantaneous" for a good move, and "a long time" for a great move. I'm not looking for evidence that someone is using an engine, I'm trying to figure out how fast a game (how short a game) I should play to give a moderately good chance of knowing that if my opponent is going to use an engine, s/he is going to run out of time feeding it ...[text shortened]... e demo Fritz and it took ..."
Thanks very much for your time and understanding
I wonder how hard it would be to measure the length of time to make each move as a reason to suspect engine use.
For example, any human, even playing well would find themselves in some positions for which the move is pretty obvious, and others that could take longer. You would expect a human to have high variance in move times. Whereas using an engine would require overhead to setup the position, wait for an engine to generate a move, and then enter the move. If you watch all of a persons games, an engine user would likely be more consistent in how quickly he/she moves, and what order he or she moves through the list of games needing a move.
The average user could not tell this, but a site administrator could store the time between first viewing an opponents move and entering a response and perhaps detect suspicious patterns. I human would likely look at an opponents move and perhaps sleep on a response for a difficult move, but would often move instantly for obvious moves.
Of course, I don't understand why someone would use engines, but I fully expect that some engine users would use engines for some moves, but not others.