NCM plays each Stockfish dev build 20,000 times against Stockfish 15. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.
| Host | Duration | Avg Base NPS | Games | WLD | Standard Elo | Ptnml(0-2) | Gamepair Elo |
|---|
| ID | Host | Base NPS | Games | WLD | Standard Elo | Ptnml(0-2) | Gamepair Elo | CLI | PGN |
|---|
| Commit ID | 6c8663341eab5c45e57c85a0b1d7f79f52fa1709 |
|---|---|
| Author | Marco Costalba |
| Date | 2012-10-13 08:40:38 UTC |
|
Scale contempt factor to zero at endgame
Contempt Factor is more effective at opening/middle game
and seems harmful at endgame phase. See:
http://chessprogramming.wikispaces.com/Contempt+Factor
http://web.archive.org/web/20070707023203/www.brucemo.com/compchess/programming/contempt.htm
Therefore we scale down the contempt factor while going
on with the game so to reach zero at endgame phase.
No functional change.
|
|