Best move, accuracy and cheating

On a different thread, some thought perhaps his opponent cheated, citing a high accuracy and number of best moves in the game.

Accuracy only equates to making a better move, or best move in some cases, with the given board position. That is, making a move that either improves your position, or best move, does not degrade your position if able, or plays a decent, but not necessarily the best move, if your already in an inferior position.

In fact, you could have a skew of best moves and high accuracy and blunder it all away on a mouse slip or mental distraction.

Let's talk about BEST MOVE: it only signifies you played the best move at the given board position... even if it's a losing move. Perhaps, hypothetically, the best move was still losing, but gave you 5 more moves before being mated: if it's a hard to see mate, it means you've bought yourself time for the opponent to not play the correct sequence of moves to mate you.

They usually lock topics like this really quickly.

A few quick notes:

1) High accuracy in one game isn't indicative of cheating. Especially when you play a lower rated opponent, it's easier to get a very high accuracy score. Instead of suspecting an opponent after 1 or a few games, it's better to report them because chess.com has some fairly sophisticated methods of detecting cheaters.

2) The best move can only maintain the theoretical evaluation. From a mathematical perspective, the best move can never improve your position.

3) Playing the same moves as an engine is not playing strong or best moves as much as it's playing like an engine. That's why it's a reliable means of cheat detection. In a real game the best moves (in a practical sense, not purely mathematical sense) are not always chosen by an engine.

4) The post game analysis is fast and runs in your browser or phone. It has limited resources so it's not finding the "best" moves anyway. You can download and install the same engine and run it for more time for a more accurate assessment of your moves.

The whole point if my post was that high accuracy and high number of best moves does NOT mean the opponent is cheating.

I thought I was pretty clear about that in my original.poat.

Lots of people have games that are highly accurate (as rated by chess.com's auto analysis thing). You're right, that alone doesn't mean the person cheated.

But you gave a bad example. For example postponing mate the engine way (highest number of moves) usually gets you checkmated faster because it involves sacrificing material. A human way takes into account practical considerations... some moves are harder to find than others. So they'll let you have a mate in 3 instead of 5, but it's a really hard to find mate in 3. That's one area (defense) where it's easier to tell a human apart from an engine.

I'm not a chess engine and don't think like one. I didn't engines were in the habit of sacrificing material to achieve mate.

Personally, I try to avoid chess engines unless playing one that set to about my skill level or slightly higher. 😊

Yeah, it's a pretty signature thing.

When they can see mate is imminent they start playing like idiots.

For example if there's a mate in 10, the human probably hasn't seen it right? But if the engine sees it, and sacrificing the queen right away makes it a mate in 10 (instead of a mate in 9) then the engine will suddenly sacrifice its queen during your attack.

Ahh.

Yes, we lock topics about cheating pretty quickly: if you'd really like to discuss how to identify engine use, you should join the Cheating Discussion club

Thanks,

David, moderator