In dem neuen Discord Thread „tune-results“ fand ich folgenden Eintrag:

**Tune of CPuct/FpuValue/PolicyTemperature**
**LC0 version:** v0.26.3
**LC0 options:** Network: J92-190, Threads=1, MinibatchSize=30, MaxPrefetch=15, MaxCollisionEvents=15, MaxOutOfOrderEvalsFactor=2.0, Backend=cuda-fp16, MoveOverheadMs=0, ScoreType=centipawn_2019, MLH: TCEC-SuFi S18
**SF options:** SF-dev, Threads=1, Hash=2, Contempt=0, "Move Overhead"=0, SyzygyProbeDepth=10
**Tuning ranges:** CPuct: [0;5], FpuValue [-0.5;1.5], PolicyTemperature [0.5;3.0]
**Tuning configuration:** 4  separate tunings with different acq functions: mes/pvrs/ts/ei, rounds=5 and 1 tuning with CLOP: 10 games/iteration (equivalent to 5 rounds/iteration), 10000 games/tuning for each tuning session
**Hardware:** Ryzen 9 3950X (3.5GHz) + RTX 2060
**Time control:** 1.2s/game+0.02s/move (LC0/SF)
**Speed:** ≈200 nodes/move
**Book:** balanced 3-move book
**Tablebases:** 6-man
**Adjudication:** 6-man TBs + -resign movecount=3 score=500, -draw movenumber=20 movecount=5 score=10
**Software:** chess-tuning-tools / CLOP
**Optimum found:**
```                        CTT.EI   default  CTT.TS   CTT.MES  CTT.PVRS    CLOP
CPuct                      2.084    2.147    1.524    1.543     1.700    2.422
FpuValue                 0.296    0.443    0.261    0.360     0.190    0.612
PolicyTemperature    1.602    1.607    1.866    1.729     1.303    1.711```

**Comment:** Comparison between different optimiziation strategies in chess-tuning-tools and CLOP. Some of the setting configurations might be more suitable for specific tasks like pure exploration and are probably not designed to be used alone by itself. Nevertheless I wanted to see what works best in a 10000 games tuning.

**Test gauntlet:**
```   # ENGINE                      :  RATING  ERROR  CFS(%)   GAMES  DRAWS(%)  OppN
   1 lc0.net.J92-190.CTT.EI      :     0.4    7.0    54.3   10000      47.1     1
   2 lc0.net.J92-190.default     :     0.0   ----    79.6   10000      46.7     1
   3 lc0.net.J92-190.CTT.TS      :    -2.9    7.0    60.2   10000      47.0     1
   4 lc0.net.J92-190.CTT.MES     :    -3.9    7.0    66.0   10000      46.0     1
   5 lc0.net.J92-190.CTT.PVRS    :    -5.4    7.2    55.8   10000      47.9     1
   6 lc0.net.J92-190.CLOP        :    -5.9    7.0    77.4   10000      46.5     1
   7 stockfish-dev               :    -7.8    5.0     ---   60000      46.9     6
White advantage = 3.6 +/- 1.1
Draw rate (equal opponents) = 46.9 % +/- 0.2```

Lothar