..., das kann man aus den Tests von Mattblachess schließen:
**Match:** lc0.26.T60_64177 vs Stockfish 11 4CPU - 100 rapid games
**LC0-version:** v26.0 cuda
**LC0-options:** --backend=multiplexing --cpuct=2.147 --cpuct-factor=2.815 --cpuct-base=18368 --fpu-value=0.443 --policy-softmax-temp=1.607 --moves-left-max-effect=0.1 --moves-left-threshold=0.5 --moves-left-slope=0.007 --moves-left-quadratic-factor=0 --moves-left-scaled-factor=1 --moves-left-constant-factor=0
**Time control:** 15min + 5s (CCRL 40/15)
**Hardware:** CPU i7-8700 4cores vs 1070Ti GPU
**Book:** Custom short lines played from both sides (!sheet4 for opening list)
**Tablebase:** 6 piece syzygy (DTZ + WDL) for both engines
**Software:** Arena
**Speed:** Leela ratio ~0.85; Lc0 npm~200K, SF npm~200M (based on 24x256 nets, SV nets much slower)
**Context:** !sheet4 for test history
Ordo w/ comparison to previous T60 scores:
```diff
# PLAYER : RATING ERROR POINTS PLAYED (%) CFS(%) W D L D(%)
1 lc0.25.1.63677_MLH2 : 3529.3 31.7 54.5 100 55 56 15 79 6 79
2 lc0.25.63558 : 3525.7 30.0 54.0 100 54 68 14 80 6 80
+ 3 lc0.26.64177 : 3514.9 31.1 52.5 100 53 56 14 77 9 77
4 lc0.25.63450 : 3511.3 30.9 52.0 100 52 69 12 80 8 80
5 lc0.25.1.63730_MLH2 : 3500.6 30.9 50.5 100 51 59 12 77 11 77
- 6 Stockfish_11_x64_bmi2 : 3497.0 ---- 236.5 500 47 --- 40 393 67 79```
```lc0.26.64177 - Stockfish_11_x64_bmi2 : 52.5/100 14-9-77
(===0====10=====1=====1=0==1=1=======1=====1======0101======1=0======101==0=========11=======0=======) 53% -> 3515 ordo score```
Das ist bedauerlich, da auch die SV-Netze auf dem Trainingsdaten von T60 basieren.
Lothar