Search found 181 matches

by Javier Ros
Fri Sep 28, 2018 5:41 pm
Forum: Computer Chess Club: General Topics
Topic: First Stockfish loss for some time ...
Replies: 7
Views: 1506

Re: First Stockfish loss for some time ...

[Event "Lc0server11261Stockfish9_25mas10HT6CPU1024syzygy"] [Site "J-PC"] [Date "2018.09.06"] [Round "23"] [White "Lc01711261syzygy"] [Black "Stockfish_9_x64_syzygy"] [Result "1-0"] [ECO "E17"] [Opening "Queen's Indian"] [Time "15:15:15"] [Variation "6.O-O O-O 7.Re1 d5"] [TimeControl "1500+10"] [Ter...
by Javier Ros
Fri Sep 28, 2018 5:38 pm
Forum: Computer Chess Club: General Topics
Topic: First Stockfish loss for some time ...
Replies: 7
Views: 1506

Re: First Stockfish loss for some time ...

Here you have some defeats of Stockfish 9 from my tests with Leela Chess Zero. Lc0 version 0.17 with net 11261 on Asus Geforce GTX 1060 3 Gb against Stockfish 9 with 6 threads on i7 4790S with 4 cores and 8 threads, ponder Off, Hash tables 1 Gb. for both under Arena gui. 5mas3 means 5 minutes + 3 se...
by Javier Ros
Thu Sep 27, 2018 7:01 am
Forum: Computer Chess Club: General Topics
Topic: LC0 0.18rc2 released
Replies: 24
Views: 4577

LC0 0.18rc2 released

Second LC0 release candidate of 0.18 released
Change:
Fixed a severe race condition bug when only happened when --out-of-order-eval is on (which is the default value now).
Could produce incorrect search results.

https://github.com/LeelaChessZero/lc0/releases
by Javier Ros
Wed Sep 26, 2018 4:58 pm
Forum: Computer Chess Club: General Topics
Topic: Leela doesn't obey well the Elo curve set by regular engines
Replies: 9
Views: 5557

Re: Leela doesn't obey well the Elo curve set by regular engines

In CCCC successive rounds, I noticed that Lc0 underperforms against weaker engines, often failing to convert won games, or even losing on tactical blunders to a weaker engine. I took the excellent FGRL rating list of Andreas Strangmüller at 60''+ 0.6'' time control, adapted to my conditions, and ha...
by Javier Ros
Wed Sep 26, 2018 3:50 pm
Forum: Computer Chess Club: General Topics
Topic: Beautiful game, highlights strengths of NN+MCTS
Replies: 17
Views: 3698

Re: Beautiful game, highlights strengths of NN+MCTS

Lc0 version 0.17 with net 11261 on Asus Geforce GTX 1060 3 Gb against Houdini 6 Pro with 6 threads on i7 4790S (4 cores and 8 threads), ponder Off, Hash tables 1 Gb. for both under Arena gui. Using the following set of 12 opening positions of AlphaZero-Stockfish8 Test with white and black without ta...
by Javier Ros
Wed Sep 26, 2018 3:44 pm
Forum: Computer Chess Club: General Topics
Topic: Chess.com 2018 computer chess championship
Replies: 279
Views: 57154

Re: Chess.com 2018 computer chess championship

This is the table after some cleaning up (not only the unterminated, but also the wrong game headers) Game 1 which was still testing and was played with no opening moves at all still counts... Andscacs instead of 1. e4 c5 2. Nf3 h6 3. c3 vs. SF played from the start position and the testing game wa...
by Javier Ros
Mon Sep 24, 2018 4:27 pm
Forum: Computer Chess Club: General Topics
Topic: leela is official(?) better than sf9
Replies: 81
Views: 14236

Re: leela is official(?) better than sf9

Our differences probably come from using different set of openings, as Lc0 and A0 seem sensitive to the choice. 12 openings for 1200 games match and 1 opening for 100 games match is bad testing methodology by A0 team. And probably on purpose bad, as it bolster A0 (or Lc0) by about 100 Elo points. P...
by Javier Ros
Mon Sep 24, 2018 1:42 pm
Forum: Computer Chess Club: General Topics
Topic: leela is official(?) better than sf9
Replies: 81
Views: 14236

Re: leela is official(?) better than sf9

Our differences probably come from using different set of openings, as Lc0 and A0 seem sensitive to the choice. 12 openings for 1200 games match and 1 opening for 100 games match is bad testing methodology by A0 team. And probably on purpose bad, as it bolster A0 (or Lc0) by about 100 Elo points. P...
by Javier Ros
Mon Sep 24, 2018 1:38 pm
Forum: Computer Chess Club: General Topics
Topic: leela is official(?) better than sf9
Replies: 81
Views: 14236

Re: leela is official(?) better than sf9

Our differences probably come from using different set of openings, as Lc0 and A0 seem sensitive to the choice. 12 openings for 1200 games match and 1 opening for 100 games match is bad testing methodology by A0 team. And probably on purpose bad, as it bolster A0 (or Lc0) by about 100 Elo points. P...
by Javier Ros
Sun Sep 23, 2018 8:59 pm
Forum: Computer Chess Club: General Topics
Topic: Beautiful game, highlights strengths of NN+MCTS
Replies: 17
Views: 3698

Re: Beautiful game, highlights strengths of NN+MCTS

Thank you for share these interesting analysis of the games! I have played more games with lc0 not published due to lack of time. This time I have managed to upload the games to the links below with the complete output of Arena, including eval, time, depth and PV. Lc0 version 0.17 with net 11261 on ...