Obsidian vs Stockfish:
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 081224 bmi2 : 3024 10 8 450 53.6 % 3000 92.4 %
2 Obsidian 14.14 avx2 : 3000 8 10 450 46.4 % 3024 92.4 %
Individual statistics:
1 Stockfish 081224 bmi2 : 3024 450 (+ 33,=416,- 1), 53.6 %
Obsidian 14.14 avx2 : 450 (+ 33,=416,- 1), 53.6 %
2 Obsidian 14.14 avx2 : 3000 450 (+ 1,=416,- 33), 46.4 %
Game Conditions: Arena Gui, Core-i7 12700h, 4 Cores, 1 min + 1 sec TC, Balsa Opening Book (5 moves), 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/61f4sjws ... 1.pgn/file
Obsidian Tests
Moderator: Ras
-
- Posts: 3621
- Joined: Wed Mar 08, 2006 8:15 pm
- Full name: Jouni Uski
-
- Posts: 190
- Joined: Sun Oct 30, 2022 5:26 pm
- Full name: Conor Anstey
Re: Obsidian Tests
>92% draw rate
cool book yes
meaningful results
cool book yes
meaningful results
-
- Posts: 19
- Joined: Tue Sep 10, 2019 9:58 am
- Full name: mehmet karaman
Re: Obsidian Tests
Obsidian vs Stockfish:
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 081224 bmi2 : 2431 36 36 300 58.8 % 2369 17.7 %
2 Obsidian 14.14 avx2 : 2369 36 36 300 41.2 % 2431 17.7 %
Individual statistics:
1 Stockfish 081224 bmi2 : 2431 300 (+150,= 53,- 97), 58.8 %
Obsidian 14.14 avx2 : 300 (+150,= 53,- 97), 58.8 %
2 Obsidian 14.14 avx2 : 2369 300 (+ 97,= 53,-150), 41.2 %
Stockfish 081224 bmi2 : 300 (+ 97,= 53,-150), 41.2 %
Game Conditions: Arena Gui, Core-i7 12700h, 8 Cores, 1 min + 1 sec TC, UHO 2022 6 moves +120_+129 openings, 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/6zn2ss1a ... 2.pgn/file
In this test ı used an unbalanced opening and the draw rate is only %18,
I'm in favor of using balanced and short move opening books in chess engine tests. I don't mind at all if there are many draws in the matches. In tournaments like TCEC, using unbalanced opening books might make sense to make the games more exciting to watch. For me, using unbalanced opening books is like FIFA deciding that, in football matches, to reduce draws and increase goals, the home team plays with 11 players while the away team plays with 10 players.
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 081224 bmi2 : 2431 36 36 300 58.8 % 2369 17.7 %
2 Obsidian 14.14 avx2 : 2369 36 36 300 41.2 % 2431 17.7 %
Individual statistics:
1 Stockfish 081224 bmi2 : 2431 300 (+150,= 53,- 97), 58.8 %
Obsidian 14.14 avx2 : 300 (+150,= 53,- 97), 58.8 %
2 Obsidian 14.14 avx2 : 2369 300 (+ 97,= 53,-150), 41.2 %
Stockfish 081224 bmi2 : 300 (+ 97,= 53,-150), 41.2 %
Game Conditions: Arena Gui, Core-i7 12700h, 8 Cores, 1 min + 1 sec TC, UHO 2022 6 moves +120_+129 openings, 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/6zn2ss1a ... 2.pgn/file
In this test ı used an unbalanced opening and the draw rate is only %18,
I'm in favor of using balanced and short move opening books in chess engine tests. I don't mind at all if there are many draws in the matches. In tournaments like TCEC, using unbalanced opening books might make sense to make the games more exciting to watch. For me, using unbalanced opening books is like FIFA deciding that, in football matches, to reduce draws and increase goals, the home team plays with 11 players while the away team plays with 10 players.
-
- Posts: 19
- Joined: Tue Sep 10, 2019 9:58 am
- Full name: mehmet karaman
Re: Obsidian Tests
Obsidian vs Integral:
Program Elo + - Games Score Av.Op. Draws
1 Obsidian 14.14 avx2 : 3044 15 13 300 56.3 % 3000 86.7 %
2 Integral v6 avx2 bmi2 : 3000 13 15 300 43.7 % 3044 86.7 %
Individual statistics:
1 Obsidian 14.14 avx2 : 3044 300 (+ 39,=260,- 1), 56.3 %
Integral v6 avx2 bmi2 : 300 (+ 39,=260,- 1), 56.3 %
2 Integral v6 avx2 bmi2 : 3000 300 (+ 1,=260,- 39), 43.7 %
Obsidian 14.14 avx2 : 300 (+ 1,=260,- 39), 43.7 %
Game Conditions: Arena Gui, Core-i7 12700h, 4 Cores, 1 min + 1 sec TC, Balsa Opening Book (5 moves), 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/24qlb5n5 ... 3.pgn/file
Program Elo + - Games Score Av.Op. Draws
1 Obsidian 14.14 avx2 : 3044 15 13 300 56.3 % 3000 86.7 %
2 Integral v6 avx2 bmi2 : 3000 13 15 300 43.7 % 3044 86.7 %
Individual statistics:
1 Obsidian 14.14 avx2 : 3044 300 (+ 39,=260,- 1), 56.3 %
Integral v6 avx2 bmi2 : 300 (+ 39,=260,- 1), 56.3 %
2 Integral v6 avx2 bmi2 : 3000 300 (+ 1,=260,- 39), 43.7 %
Obsidian 14.14 avx2 : 300 (+ 1,=260,- 39), 43.7 %
Game Conditions: Arena Gui, Core-i7 12700h, 4 Cores, 1 min + 1 sec TC, Balsa Opening Book (5 moves), 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/24qlb5n5 ... 3.pgn/file
-
- Posts: 19
- Joined: Tue Sep 10, 2019 9:58 am
- Full name: mehmet karaman
Re: Obsidian Tests
Obsidian vs Stockfish:
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 081224 bmi2 : 3016 7 5 510 52.3 % 3000 95.5 %
2 Obsidian 14.20 avx2 : 3000 5 7 510 47.7 % 3016 95.5 %
Individual statistics:
1 Stockfish 081224 bmi2 : 3016 510 (+ 23,=487,- 0), 52.3 %
Obsidian 14.20 avx2 : 510 (+ 23,=487,- 0), 52.3 %
2 Obsidian 14.20 avx2 : 3000 510 (+ 0,=487,- 23), 47.7 %
Stockfish 081224 bmi2 : 510 (+ 0,=487,- 23), 47.7 %
Game Conditions: Arena Gui, Core-i7 12700h, 8 Cores, 1 min + 1 sec TC, Balsa Opening Book (5 moves), 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/vzo7t5em ... 5.pgn/file
The Elo difference between Stockfish and Obsidian is decreasing every day. Obsidian's development speed is truly impressive. The talented developer behind Obsidian, Gabriele Lombardo, is only 16.5 years old. Gabriele Lombardo is younger than Luca-Mihnea Metehau, the 19-year-old developer of the Clover chess engine. Will Obsidian be able to take the title of the strongest chess engine from Stockfish in 2025? I'm very hopeful. We'll have to wait and see.
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 081224 bmi2 : 3016 7 5 510 52.3 % 3000 95.5 %
2 Obsidian 14.20 avx2 : 3000 5 7 510 47.7 % 3016 95.5 %
Individual statistics:
1 Stockfish 081224 bmi2 : 3016 510 (+ 23,=487,- 0), 52.3 %
Obsidian 14.20 avx2 : 510 (+ 23,=487,- 0), 52.3 %
2 Obsidian 14.20 avx2 : 3000 510 (+ 0,=487,- 23), 47.7 %
Stockfish 081224 bmi2 : 510 (+ 0,=487,- 23), 47.7 %
Game Conditions: Arena Gui, Core-i7 12700h, 8 Cores, 1 min + 1 sec TC, Balsa Opening Book (5 moves), 512 Mb Hash, Ponder Off
https://www.mediafire.com/file/vzo7t5em ... 5.pgn/file
The Elo difference between Stockfish and Obsidian is decreasing every day. Obsidian's development speed is truly impressive. The talented developer behind Obsidian, Gabriele Lombardo, is only 16.5 years old. Gabriele Lombardo is younger than Luca-Mihnea Metehau, the 19-year-old developer of the Clover chess engine. Will Obsidian be able to take the title of the strongest chess engine from Stockfish in 2025? I'm very hopeful. We'll have to wait and see.
-
- Posts: 19
- Joined: Tue Sep 10, 2019 9:58 am
- Full name: mehmet karaman
Re: Obsidian Tests
Obsidian vs Stockfish:
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 221224 bmi2 : 3008 11 8 200 51.2 % 3000 96.5 %
2 Obsidian 14.23 avx2 : 3000 8 11 200 48.8 % 3008 96.5 %
Individual statistics:
1 Stockfish 221224 bmi2 : 3008 200 (+ 6,=193,- 1), 51.2 %
Obsidian 14.23 avx2 : 200 (+ 6,=193,- 1), 51.2 %
2 Obsidian 14.23 avx2 : 3000 200 (+ 1,=193,- 6), 48.8 %
Stockfish 221224 bmi2 : 200 (+ 1,=193,- 6), 48.8 %
Game Conditions: Arena Gui, Core-i7 12700h, 12 Cores, 2 min + 1 sec TC, Balsa Opening Book (5 moves), 2048 Gb Hash, Ponder Off
https://www.mediafire.com/file/zbcqnpmr ... 9.pgn/file
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 221224 bmi2 : 3008 11 8 200 51.2 % 3000 96.5 %
2 Obsidian 14.23 avx2 : 3000 8 11 200 48.8 % 3008 96.5 %
Individual statistics:
1 Stockfish 221224 bmi2 : 3008 200 (+ 6,=193,- 1), 51.2 %
Obsidian 14.23 avx2 : 200 (+ 6,=193,- 1), 51.2 %
2 Obsidian 14.23 avx2 : 3000 200 (+ 1,=193,- 6), 48.8 %
Stockfish 221224 bmi2 : 200 (+ 1,=193,- 6), 48.8 %
Game Conditions: Arena Gui, Core-i7 12700h, 12 Cores, 2 min + 1 sec TC, Balsa Opening Book (5 moves), 2048 Gb Hash, Ponder Off
https://www.mediafire.com/file/zbcqnpmr ... 9.pgn/file
-
- Posts: 19
- Joined: Tue Sep 10, 2019 9:58 am
- Full name: mehmet karaman
Re: Obsidian Tests
Obsidian vs Stockfish:
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 130525 bmi2 : 3004 16 3 110 50.5 % 3000 99.1 %
2 Obsidian 15.22 bmi2 : 3000 3 16 110 49.5 % 3004 99.1 %
Individual statistics:
1 Stockfish 130525 bmi2 : 3004 110 (+ 1,=109,- 0), 50.5 %
Obsidian 15.22 bmi2 : 110 (+ 1,=109,- 0), 50.5 %
2 Obsidian 15.22 bmi2 :3000 110 (+ 0,=109,- 1), 49.5 %
Stockfish 130525 bmi2 : 110 (+ 0,=109,- 1), 49.5 %
Game Conditions: Arena Gui, Core-i7 12700h, 12 Cores, 2 min + 1 sec TC, Balsa Opening Book (5 moves), 2048 Gb Hash, Ponder Off
https://www.mediafire.com/file/09dpmxrj ... 0.pgn/file
Under these time conditions (12 Cores, 2 min + 1 sec) with a balanced opening book , the strength difference is very small. Out of 110 games, there was only one win. I will make another test with an unbalanced opening book.
Stockfish has been developed at full capacity by the community for 12–13 years, and its rate of improvement is gradually decreasing. Obsidian, on the other hand, has much more development potential. That's why I predict that Obsidian will surpass Stockfish before 2026
Program Elo + - Games Score Av.Op. Draws
1 Stockfish 130525 bmi2 : 3004 16 3 110 50.5 % 3000 99.1 %
2 Obsidian 15.22 bmi2 : 3000 3 16 110 49.5 % 3004 99.1 %
Individual statistics:
1 Stockfish 130525 bmi2 : 3004 110 (+ 1,=109,- 0), 50.5 %
Obsidian 15.22 bmi2 : 110 (+ 1,=109,- 0), 50.5 %
2 Obsidian 15.22 bmi2 :3000 110 (+ 0,=109,- 1), 49.5 %
Stockfish 130525 bmi2 : 110 (+ 0,=109,- 1), 49.5 %
Game Conditions: Arena Gui, Core-i7 12700h, 12 Cores, 2 min + 1 sec TC, Balsa Opening Book (5 moves), 2048 Gb Hash, Ponder Off
https://www.mediafire.com/file/09dpmxrj ... 0.pgn/file
Under these time conditions (12 Cores, 2 min + 1 sec) with a balanced opening book , the strength difference is very small. Out of 110 games, there was only one win. I will make another test with an unbalanced opening book.
Stockfish has been developed at full capacity by the community for 12–13 years, and its rate of improvement is gradually decreasing. Obsidian, on the other hand, has much more development potential. That's why I predict that Obsidian will surpass Stockfish before 2026
-
- Posts: 3621
- Joined: Wed Mar 08, 2006 8:15 pm
- Full name: Jouni Uski