Not logged inCSS-Forum
Forum CSS-Online Help Search Login
CSS-Shop Impressum Datenschutz
Up Topic Hauptforen / CSS-Stockfish / stockfish Fortschritt (regression tests)
- - By Guenter Stertenbrink Date 2018-04-08 09:48
hab die mal rausgesucht (aus den 518 Seiten fishtests)
seit Juni 2016 , normalerweise 40000 Partien

2016-jun10 , -43.3
2016-jun13 , -43.7
2016-jul26 , -30.3
2016-aug18 , -27.4
2016-sep09 , -13.5  ; +66.5
2016-oct11 , -03.4  ; +76.6
2016-oct15 , -02.4  ; +77.5
2016-nov04 , +00.0  ; +80.0
2016-dec31 , +05.2
2017-jan29 , +08.8
2017-mar09 , +10.8
2017-apr20 , +15.2
2017-may08 , +20.3
2017-Jun23 , +27.4
2017-Aug27 , +29.3
2017-Oct02 , +32.6
2017-Nov03 , +35.2
2017-Dec03 , +43.3
2018-Jan23 , +57.3  ; contemp=20 ; 2018/02/01 : sf9
2018-Feb09 , +59.4  ; +00.0    + 59.4 or + 66.3 for sf9 over sf 8 ?
2018-Feb29 , +64.6  ; +05.2
2018-Mar07 , +66.3  ; +06.9
2018-Mar14 , +72.4  ; +13.0
2018-Mar27 , +73.2  ; +13.8
2018-Apr05 , +76.1  ; +16.7

http://magictour.free.fr/1231G1.GIF  (wird spaeter upgedated)
Parent - By Guenter Stertenbrink Date 2018-10-26 13:01 Edited 2018-10-26 13:05 Upvotes 1
gute Uebersicht der "regression" Tests hier : https://github.com/glinscott/fishtest/wiki/Regression-Tests

+49 Elo gegenueber SF9 am 14.Okt.

aktuell : https://nextchessmove.com/dev-builds
> So I am pretty certain that the 2 NCM +201 elo tests are both with the fail-high patch
Up Topic Hauptforen / CSS-Stockfish / stockfish Fortschritt (regression tests)

Powered by mwForum 2.29.3 © 1999-2014 Markus Wichitill