Stockfish: ZH : opperwezen stronger again : 6 wins ! (15+0 and 1+0)

[Vinvin20]

https://lichess.org/viRHaJOq
https://lichess.org/aEcWuW4h
https://lichess.org/h71C7U7I#37

https://lichess.org/5f4iK1wL#51
https://lichess.org/EZCK3E23#54
https://lichess.org/zwHZjZsc#73

Crazy guy ! :-)
May be there are a lot of opening preparation ...

[ppigazzini]

@Vinvin20 It would be useful if opperwezen (or other strong ZH players) could give his point of view about the weaknesses in SF's play

[ddugovic]

Simply unreal. I'll ask around & see if the Lichess cluster is using the latest Stockfish...

Just taking one example it seems impossible that Stockfish played 12... N@a4??:

$ cat test.txt
setoption name UCI_Variant value crazyhouse
position fen r2q1rk1/ppp1bppp/4p3/3pPbn1/5N2/P1P3PP/2P1PPB1/R1BQ1RK1/Nn b - - 23 12
go depth 22 movetime 1000

$ cat test.txt - | ./stockfish-x86_64-bmi2 
Stockfish 030717 64 BMI2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
info string variant crazyhouse startpos rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR[] w KQkq - 0 1
info depth 1 seldepth 1 multipv 1 score cp 359 nodes 77 nps 38500 tbhits 0 time 2 pv e7c5
info depth 2 seldepth 2 multipv 1 score cp 293 nodes 246 nps 123000 tbhits 0 time 2 pv g5e4 g2e4 f5e4
info depth 3 seldepth 3 multipv 1 score cp 293 nodes 339 nps 169500 tbhits 0 time 2 pv g5e4 g2e4 f5e4
info depth 4 seldepth 4 multipv 1 score cp 258 nodes 805 nps 268333 tbhits 0 time 3 pv g5e4 N@d4 e7c5 c1b2
info depth 5 seldepth 5 multipv 1 score cp 275 nodes 1712 nps 428000 tbhits 0 time 4 pv g5e4 c1b2 N@c4 g2e4 f5e4
info depth 6 seldepth 6 multipv 1 score cp 336 nodes 4288 nps 612571 tbhits 0 time 7 pv N@c4 N@d4 f5e4 f2f3 c4e5 f3e4 g5e4
info depth 7 seldepth 7 multipv 1 score cp 263 nodes 12203 nps 871642 tbhits 0 time 14 pv f5e4 g2e4 g5e4 B@f3 e4c3 d1e1 c3e4
info depth 8 seldepth 8 multipv 1 score cp 270 nodes 14514 nps 907125 tbhits 0 time 16 pv N@c4 N@e3 f5e4 g2e4 g5e4 e3c4 d5c4 d1d8 a8d8
info depth 9 seldepth 11 multipv 1 score cp 299 nodes 22750 nps 947916 tbhits 0 time 24 pv N@c4 N@e3 c4e3 c1e3 N@c4 a1b1 c4e3 f2e3 B@e4
info depth 10 seldepth 14 multipv 1 score cp 227 nodes 103504 nps 1024792 tbhits 0 time 101 pv N@c4 N@d4 c7c6 d4f5 e6f5 B@d4 c6c5 f4d5 g5h3 g2h3 d8d5 P@f6
info depth 11 seldepth 17 multipv 1 score cp 239 nodes 273500 nps 1028195 tbhits 0 time 266 pv N@e4 N@d4 e4c3 d4f5 c3d1 f5e7 d8e7 f1d1 Q@b6 B@e3 P@b2 e3b6 b2a1q
info depth 12 seldepth 19 multipv 1 score cp 224 nodes 589785 nps 1023932 tbhits 0 time 576 pv N@c4 N@d4 c7c6 d4f5 e6f5 B@d4 N@e6 f4e6 g5e6 a1b1 d8c7 h3h4 c4a3
info depth 13 seldepth 19 multipv 1 score cp 214 nodes 913963 nps 1025772 tbhits 0 time 891 pv N@e4 N@d4 e4c3 d4f5 e6f5 d1d3 N@e4 B@a5 d5d4 h3h4 P@e3 f2e3 e4g3 h4g5 g3f1 e3d4
info depth 14 seldepth 25 multipv 1 score cp 214 nodes 1007023 nps 1002012 hashfull 491 tbhits 0 time 1005 pv N@e4 N@d4 e4c3 d4f5 e6f5 d1d3 N@e4 B@a5 d5d4 h3h4 P@e3 f2e3 e4g3 h4g5 g3f1 e3d4
bestmove N@e4 ponder N@d4

$ cat test.txt
setoption name UCI_Variant value crazyhouse
setoption name MultiPV value 2
position fen r2q1rk1/ppp1bppp/4p3/3pPbn1/5N2/P1P3PP/2P1PPB1/R1BQ1RK1/Nn b - - 23 12
go depth 22 movetime 1000

$ cat test.txt - | ./stockfish-x86_64-bmi2 
Stockfish 030717 64 BMI2 by T. Romstad, M. Costalba, J. Kiiski, G. Linscott
info string variant crazyhouse startpos rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR[] w KQkq - 0 1
info depth 1 seldepth 1 multipv 1 score cp 359 nodes 148 nps 148000 tbhits 0 time 1 pv e7c5
info depth 1 seldepth 1 multipv 2 score cp 248 nodes 148 nps 148000 tbhits 0 time 1 pv g5e4
info depth 2 seldepth 2 multipv 1 score cp 293 nodes 404 nps 202000 tbhits 0 time 2 pv g5e4 g2e4 f5e4
info depth 2 seldepth 2 multipv 2 score cp 256 nodes 404 nps 202000 tbhits 0 time 2 pv N@a4 h3h4 a4c3
info depth 3 seldepth 3 multipv 1 score cp 274 nodes 741 nps 247000 tbhits 0 time 3 pv g5e4 N@e3 e7c5
info depth 3 seldepth 3 multipv 2 score cp 274 nodes 741 nps 247000 tbhits 0 time 3 pv f5e4 g2e4 g5e4
info depth 4 seldepth 4 multipv 1 score cp 302 nodes 1469 nps 367250 tbhits 0 time 4 pv f5e4 g2e4 g5e4 h3h4
info depth 4 seldepth 4 multipv 2 score cp 287 nodes 1469 nps 367250 tbhits 0 time 4 pv N@a4 h3h4 a4c3 h4g5
info depth 5 seldepth 5 multipv 1 score cp 299 nodes 2548 nps 424666 tbhits 0 time 6 pv f5e4 g2e4 g5e4 h3h4 e4c3
info depth 5 seldepth 5 multipv 2 score cp 287 nodes 2548 nps 424666 tbhits 0 time 6 pv N@e4 h3h4 e4c3 h4g5 c3d1
info depth 6 seldepth 6 multipv 1 score cp 366 nodes 5718 nps 476500 tbhits 0 time 12 pv f5e4 g2e4 g5e4 c1d2 B@f5 e2e3
info depth 6 seldepth 6 multipv 2 score cp 325 nodes 5718 nps 476500 tbhits 0 time 12 pv e7c5 c1d2 f5e4 g2e4 d5e4 N@c4
info depth 7 seldepth 8 multipv 1 score cp 272 nodes 19132 nps 531444 tbhits 0 time 36 pv N@c4 N@e3 f5e4 g2e4 g5e4 e3c4 d5c4
info depth 7 seldepth 8 multipv 2 score cp 245 nodes 19132 nps 531444 tbhits 0 time 36 pv g5e4 N@d4 c7c6 c1e3 e4c3 d4f5 e6f5
info depth 8 seldepth 9 multipv 1 score cp 277 nodes 27067 nps 520519 tbhits 0 time 52 pv N@c4 N@e3 f5h3 f4h3 g5h3 g2h3 P@b2 e3c4 b2a1q
info depth 8 seldepth 9 multipv 2 score cp 270 nodes 27067 nps 520519 tbhits 0 time 52 pv g5e4 N@d4 c7c6 c1d2 N@c4 d4f5 e6f5 g2e4 f5e4
info depth 9 seldepth 11 multipv 1 score cp 228 nodes 126711 nps 517187 tbhits 0 time 245 pv g5e4 d1e1 e7c5 e2e3 c7c6 N@d3 c5b6 g3g4 d8g5
info depth 9 seldepth 11 multipv 2 score cp 218 nodes 126711 nps 517187 tbhits 0 time 245 pv N@c4 N@d4 f5e4 f2f3 c4e5 f3e4 g5e4 g2e4 d5e4 B@e3
info depth 10 seldepth 14 multipv 1 score cp 269 nodes 217803 nps 514900 tbhits 0 time 423 pv g5e4 g3g4 e7c5 N@d3 e4c3 d3c5 c3d1 f1d1 f5c2 c5b7
info depth 10 seldepth 14 multipv 2 score cp 237 nodes 217803 nps 514900 tbhits 0 time 423 pv e7c5 N@d4 f5e4 f2f3 e4f5 e2e3 N@g6 f4g6 f7g6 N@f4 c5d4 e3d4
info depth 11 seldepth 16 multipv 1 score cp 256 nodes 322204 nps 514702 tbhits 0 time 626 pv g5e4 N@e3 e4c3 e3f5 c3d1 f5e7 d8e7 f1d1 N@c3 d1e1 N@c4 B@d4 P@d2 d4c3 d2e1q c3e1
info depth 11 seldepth 16 multipv 2 score cp 236 nodes 322204 nps 514702 tbhits 0 time 626 pv e7c5 N@d4 f5e4 g2e4 g5e4 B@b4 c5b4 c3b4 c7c5 b4c5 B@c3 c1e3 c3a1 d1a1
info depth 12 seldepth 20 multipv 1 score cp 245 nodes 518795 nps 508622 hashfull 248 tbhits 0 time 1020 pv e7c5 e2e3
info depth 11 seldepth 20 multipv 2 score cp 256 nodes 518795 nps 508622 hashfull 248 tbhits 0 time 1020 pv g5e4 N@e3
bestmove e7c5 ponder e2e3

[Vinvin20]

Simply unreal. I'll ask around & see if the Lichess cluster is using the latest Stockfish...

That would explain the things :-)
Please let us know when you'll get the answer ...

[ianfab]

Considering opperwezen's playing strength, his score of something like 20% is not unexpected. Losing a game is not per se a bug, but some of the examples indeed look strange, since they are not reproducible.

[ddugovic]

Considering opperwezen's playing strength, his score of something like 20% is not unexpected. Losing a game is not per se a bug, but some of the examples indeed look strange, since they are not reproducible.

Exactly; sorry for my lack of clarity but that is what I found stunning.

Today Lichess is upgrading Stockfish on its hosted machines. Hopefully over time fishnet donors (machines not under Lichess control) will upgrade as well.

So... should I leave this issue open indefinitely?

[ianfab]

Thanks a lot for the info.

I think that a dedicated team or thread in the lichess forum would be more suitable to share and discuss such findings.

[ddugovic]

I too think a dedicated team, Lichess forum thread, etc. would be appropriate.

I am an active member in Lichess forums / teams / Q&A:

• Lichess Feedback http://lichess.org/forum/lichess-feedback
• Coders http://lichess.org/forum/team-coders
• Crazyhouse Engine Development and Game Analyses http://lichess.org/forum/team-crazyhouse-engine-development-and-game-analyses
• Q&A http://lichess.org/qa

and I respond to @Toadofsky in the Lichess forums

Additionally, I am an active member in the Lichess discord http://discord.gg/AWSze7n

I think any and/or all of the above are viable options for discussion of issues not reproducible in latest Stockfish. In fact, I have already created such a thread, and apparently nobody other than myself is using it (perhaps due to its negative connotation -- that coding cannot fix all real-world problems, especially those which are ill-defined and/or nonreproducible):
http://lichess.org/forum/team-coders/stockfish-xy-problems

I suggest someone pick one or more of the above options. Maybe I need to create a Lichess team devoted to tracking how/when Lichess deploys Stockfish updates?

[niklasf]

As an aside, I believe all playing instances of Stockfish have been updated. (Contributed instances are only used for analysis.)

[ddugovic]

Thanks, that is good news to hear.

I have created an AI level 8 tracker for people reporting "AI level 8 blundered but I don't know if this is reproducible in latest Stockfish":
http://lichess.org/forum/lichess-feedback/ai-level-8-tracker