TalkChess.com

Testrun of Stockfish 180801 finished.


http://www.sp-cc.de

(Perhaps you have to clear your browsercache or reload the website)

Thanks for running this one Stefan! There was some talk that the 08-01 (August 1st) could have been a regression even though no change in bench number and algorithmically it should be all the same. Once more the testing elsewhere did not seem reliable. It is possible here are also the error margins staying hidden in the actual results but from the trend you would conclude that maybe the july version was just tested a bit low. The 7-31 and 8-01 version are both "nice" which is totally my subjective feel. Meaning they respond well to little changes in the code I made.

+31 over SF9. Latest version got +38 in framework and about +37 in NCM (selftest both). Again the progress is heading to +70 ELO in year, stunning!

Team is on fire:

ELO: 41.11 +-1.9 (95%) LOS: 100.0%
Total: 39479 W: 8648 L: 3998 D: 26833

and this is waiting:

LTC:
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 19683 W: 3409 L: 3206 D: 13068 Elo +3.58

When the testrun of Booot 6.3.1 is done, I will start the next Stockfish-testrun, of course. Seems, that some nice progress was made!