SPCC: Testrun of Stockfish 190420 finished

[pohl4711]

Testrun of Stockfish 190420 finished. Another bad regression...

https://www.sp-cc.de


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

[JJJ]

There is no regression, probably more a false positive on your test. Any regression test on Stockfish are showing the same +16 elo since 2 or 3 month.

[pohl4711]

There is no regression, probably more a false positive on your test. Any regression test on Stockfish are showing the same +16 elo since 2 or 3 month.

All these regression-tests are Stockfish vs. another Stockfish (selfplay). My testruns are not - Stockfish plays vs. 5 other Engines. It is possible to have a regression vs. other engines and not in selfplay. And it is possible to have a regression in selfplay and not in tests vs. other engines.
That is, why I am doing these Stockfish-testruns.

[mehmet karaman]

Stockfish 10/03/2019 vs Stockfish 7: 222.77 elo
Stockfish 27/04/2019 vs Stockfish 7: 215.82 elo

https://nextchessmove.com/dev-builds

I think some new patches aren't as good as expected

[JJJ]

Indeed, somethings seems wrong with these patch.

[mehmet karaman]

Score of Stockfish 27042019 x64 bmi2 vs Stockfish 10032019 x64 bmi2: 173 - 194 - 633 [0.489]
1000 Games, ELO difference: -7.30 +/- 13.03

Cutechess Gui, 30''+ 0.5'' time control, 1 cpu, core i7-4712mq, 8 moves pgn, ponder off