Another cutechess-cli question

[Rebel]

I am planning to switch from Arena to cutechess-cli once and for all. The thing I don't like is that matches run approx 15-18% longer.

Arena has this nice option that you can configure games as a draw when (say) 5 consecutive moves have a (say) 0.00 score.

cutechess-cli however has:

-draw <n> <score>::
Adjudicate the game as a draw if the score of both engines is within <score> centipawns from zero after <n> full moves have been played.

The question is how to interpret full moves. As game-limit (max_moves) or as consecutive moves ?

IOW, can one configure cutechess-cli the same way as Arena?

[lucasart]

I am planning to switch from Arena to cutechess-cli once and for all. The thing I don't like is that matches run approx 15-18% longer.

Arena has this nice option that you can configure games as a draw when (say) 5 consecutive moves have a (say) 0.00 score.

cutechess-cli however has:

-draw <n> <score>::
Adjudicate the game as a draw if the score of both engines is within <score> centipawns from zero after <n> full moves have been played.

The question is how to interpret full moves. As game-limit (max_moves) or as consecutive moves ?

IOW, can one configure cutechess-cli the same way as Arena?

You should download and compile the latest and hottest cutechess-cli
https://github.com/cutechess/cutechess.git

As you can see the draw adjudication has been much refined now:

Code: Select all

  -draw movenumber=NUMBER movecount=COUNT score=SCORE
			Adjudicate the game as a draw if the score of both
			engines is within SCORE centipawns from zero for at
			least COUNT consecutive moves, and at least NUMBER full
			moves have been played.

In my testing, I use the following adjudication parameters:

Code: Select all

-draw movenumber=60 movecount=5 score=50
-resign movecount=3 score=700

Of course, this is a bit aggressive and will make mistakes, but not enough to measurably pollute the statistics (it's all about statistics in the end).

[ilari]

Like Lucas said, there's an unreleased version of cutechess-cli which has better draw adjudication options. But you don't have to compile it, there's a binary package for Windows: http://koti.mbnet.fi/~ilaripih/bin/cute ... -win32.zip

The Stockfish testing framework uses "-draw movenumber=34 movecount=8 score=20" for draw adjudication, but I guess Lucas' settings are fine as well.

[Rebel]

Thanks guys.

[Laskos]

Like Lucas said, there's an unreleased version of cutechess-cli which has better draw adjudication options. But you don't have to compile it, there's a binary package for Windows: http://koti.mbnet.fi/~ilaripih/bin/cute ... -win32.zip

The Stockfish testing framework uses "-draw movenumber=34 movecount=8 score=20" for draw adjudication, but I guess Lucas' settings are fine as well.

Isn't that a bit aggressive? Basically excluding late middle-games and endgames starting from fair positions? Lucas' settings seem aggressive too. For 10% time saving, one distorts results, especially endgame maneuvering, which seems like magic sometimes.