Discussion of computer chess matches and engine tournaments.
Moderator: Ras
-
Rebel
- Posts: 7539
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Post
by Rebel »
Code: Select all
Results from file Reckless-0.90-final.pgn: | Results from file Reckless-0.10-dev.pgn:
|
No. Name Score Games % | No. Name Score Games %
------------------------------------------ |---------------------------------------------
1 Reckless-0.90-final 6349.5 27000 60.6% | 1 Reckless-0.10.0-dev 6501.5 27000 61.1%
2 Stockfish-18 1565.5 3000 52.2% | 2 Stockfish-18 1532.5 3000 51.1%
3 PlentyChess-7.0.22 1362.5 3000 45.4% | 3 PlentyChess-7.0.22 1366.0 3000 45.5%
4 Alexandria-9.0 1252.0 3000 41.7% | 4 Obsidian-16 1223.0 3000 40.8%
5 Obsidian-16 1235.0 3000 41.2% | 5 Alexandria-9.0 1210.0 3000 40.3%
6 Stockfish-15 1195.5 3000 39.9% | 6 Stockfish-15 1170.5 3000 39.0%
7 Viridithas-19.0.1 1063.5 3000 35.5% | 7 Viridithas-19.0.1 1068.0 3000 35.6%
8 Caissa-1.24 1059.5 3000 35.3% | 8 Caissa-1.24 1042.5 3000 34.8%
9 Clover-9.1 1005.5 3000 33.5% | 9 Clover-9.1 978.5 3000 32.6%
10 Berserk-13 911.5 3000 30.4% | 10 Berserk-13 907.5 3000 30.2%
|
Total Games: 27000 | Total Games: 27000
White Wins: 7922 (29.3%) | White Wins: 7984 (29.6%)
Black Wins: 1713 (6.3%) | Black Wins: 1643 (6.1%)
Draws: 17365 (64.3%) | Draws: 17373 (64.3%)
27.000 games
TC 40m/10s
Overall gain
3-4 elo
Closing in on SF
7-8 elo
Top-10 rating list now :
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED W D L D(%)
1 Leela-0.32.1-BT4 : 3806.6 3.6 3154.5 6000 996 4317 687 72
2 Stockfish-18 : 3801.9 1.3 21833.5 36000 0177 23313 2510 65
3 Reckless-0.10.0-dev : 3788.8 2.0 16501.5 27000 7815 17373 1812 64
4 PlentyChess-7.0.37 : 3758.9 1.7 26308.5 44874 1691 29235 3948 65
5 Alexandria-9.0 : 3720.8 2.0 16424.0 31865 5679 21490 4696 67
6 Obsidian-16 : 3718.1 1.6 40501.5 80652 3384 54235 13033 67
7 Viridithas-19.0.1 : 3676.3 1.6 17338.5 38874 4851 24975 9048 64
8 Caissa-1.24 : 3666.8 1.9 27149.5 62652 6835 40629 15188 65
9 Clover-9.1 : 3651.7 1.7 33272.0 80649 7181 52182 21286 65
10 Berserk-13 : 3637.6 1.4 31782.5 80652 7293 48979 24380 61
90% of coding is debugging, the other 10% is writing bugs.
-
Jouni
- Posts: 3874
- Joined: Wed Mar 08, 2006 8:15 pm
- Full name: Jouni Uski
Post
by Jouni »
Overall gain 3-4 elo. But Stockfish has also gained 3-4 after version 18.
Jouni
-
Uri Blass
- Posts: 11201
- Joined: Thu Mar 09, 2006 12:37 am
- Location: Tel-Aviv Israel
Post
by Uri Blass »
Rebel wrote: ↑Thu Apr 02, 2026 11:36 am
Code: Select all
Results from file Reckless-0.90-final.pgn: | Results from file Reckless-0.10-dev.pgn:
|
No. Name Score Games % | No. Name Score Games %
------------------------------------------ |---------------------------------------------
1 Reckless-0.90-final 6349.5 27000 60.6% | 1 Reckless-0.10.0-dev 6501.5 27000 61.1%
2 Stockfish-18 1565.5 3000 52.2% | 2 Stockfish-18 1532.5 3000 51.1%
3 PlentyChess-7.0.22 1362.5 3000 45.4% | 3 PlentyChess-7.0.22 1366.0 3000 45.5%
4 Alexandria-9.0 1252.0 3000 41.7% | 4 Obsidian-16 1223.0 3000 40.8%
5 Obsidian-16 1235.0 3000 41.2% | 5 Alexandria-9.0 1210.0 3000 40.3%
6 Stockfish-15 1195.5 3000 39.9% | 6 Stockfish-15 1170.5 3000 39.0%
7 Viridithas-19.0.1 1063.5 3000 35.5% | 7 Viridithas-19.0.1 1068.0 3000 35.6%
8 Caissa-1.24 1059.5 3000 35.3% | 8 Caissa-1.24 1042.5 3000 34.8%
9 Clover-9.1 1005.5 3000 33.5% | 9 Clover-9.1 978.5 3000 32.6%
10 Berserk-13 911.5 3000 30.4% | 10 Berserk-13 907.5 3000 30.2%
|
Total Games: 27000 | Total Games: 27000
White Wins: 7922 (29.3%) | White Wins: 7984 (29.6%)
Black Wins: 1713 (6.3%) | Black Wins: 1643 (6.1%)
Draws: 17365 (64.3%) | Draws: 17373 (64.3%)
27.000 games
TC 40m/10s
Overall gain
3-4 elo
Closing in on SF
7-8 elo
Top-10 rating list now :
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED W D L D(%)
1 Leela-0.32.1-BT4 : 3806.6 3.6 3154.5 6000 996 4317 687 72
2 Stockfish-18 : 3801.9 1.3 21833.5 36000 0177 23313 2510 65
3 Reckless-0.10.0-dev : 3788.8 2.0 16501.5 27000 7815 17373 1812 64
4 PlentyChess-7.0.37 : 3758.9 1.7 26308.5 44874 1691 29235 3948 65
5 Alexandria-9.0 : 3720.8 2.0 16424.0 31865 5679 21490 4696 67
6 Obsidian-16 : 3718.1 1.6 40501.5 80652 3384 54235 13033 67
7 Viridithas-19.0.1 : 3676.3 1.6 17338.5 38874 4851 24975 9048 64
8 Caissa-1.24 : 3666.8 1.9 27149.5 62652 6835 40629 15188 65
9 Clover-9.1 : 3651.7 1.7 33272.0 80649 7181 52182 21286 65
10 Berserk-13 : 3637.6 1.4 31782.5 80652 7293 48979 24380 61
6501.5/27000 is illogical and not 61.1%
It seems that the score of reckless10 is not 6501.5 but 16501.5(same idea for reckless9)
-
Rebel
- Posts: 7539
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Post
by Rebel »
Uri Blass wrote: ↑Fri Apr 03, 2026 10:09 am
Rebel wrote: ↑Thu Apr 02, 2026 11:36 am
Code: Select all
Results from file Reckless-0.90-final.pgn: | Results from file Reckless-0.10-dev.pgn:
|
No. Name Score Games % | No. Name Score Games %
------------------------------------------ |---------------------------------------------
1 Reckless-0.90-final 16349.5 27000 60.6%| 1 Reckless-0.10.0-dev 16501.5 27000 61.1%
2 Stockfish-18 1565.5 3000 52.2% | 2 Stockfish-18 1532.5 3000 51.1%
3 PlentyChess-7.0.22 1362.5 3000 45.4% | 3 PlentyChess-7.0.22 1366.0 3000 45.5%
4 Alexandria-9.0 1252.0 3000 41.7% | 4 Obsidian-16 1223.0 3000 40.8%
5 Obsidian-16 1235.0 3000 41.2% | 5 Alexandria-9.0 1210.0 3000 40.3%
6 Stockfish-15 1195.5 3000 39.9% | 6 Stockfish-15 1170.5 3000 39.0%
7 Viridithas-19.0.1 1063.5 3000 35.5% | 7 Viridithas-19.0.1 1068.0 3000 35.6%
8 Caissa-1.24 1059.5 3000 35.3% | 8 Caissa-1.24 1042.5 3000 34.8%
9 Clover-9.1 1005.5 3000 33.5% | 9 Clover-9.1 978.5 3000 32.6%
10 Berserk-13 911.5 3000 30.4% | 10 Berserk-13 907.5 3000 30.2%
|
Total Games: 27000 | Total Games: 27000
White Wins: 7922 (29.3%) | White Wins: 7984 (29.6%)
Black Wins: 1713 (6.3%) | Black Wins: 1643 (6.1%)
Draws: 17365 (64.3%) | Draws: 17373 (64.3%)
27.000 games
TC 40m/10s
Overall gain
3-4 elo
Closing in on SF
7-8 elo
Top-10 rating list now :
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED W D L D(%)
1 Leela-0.32.1-BT4 : 3806.6 3.6 3154.5 6000 996 4317 687 72
2 Stockfish-18 : 3801.9 1.3 21833.5 36000 0177 23313 2510 65
3 Reckless-0.10.0-dev : 3788.8 2.0 16501.5 27000 7815 17373 1812 64
4 PlentyChess-7.0.37 : 3758.9 1.7 26308.5 44874 1691 29235 3948 65
5 Alexandria-9.0 : 3720.8 2.0 16424.0 31865 5679 21490 4696 67
6 Obsidian-16 : 3718.1 1.6 40501.5 80652 3384 54235 13033 67
7 Viridithas-19.0.1 : 3676.3 1.6 17338.5 38874 4851 24975 9048 64
8 Caissa-1.24 : 3666.8 1.9 27149.5 62652 6835 40629 15188 65
9 Clover-9.1 : 3651.7 1.7 33272.0 80649 7181 52182 21286 65
10 Berserk-13 : 3637.6 1.4 31782.5 80652 7293 48979 24380 61
6501.5/27000 is illogical and not 61.1%
It seems that the score of reckless10 is not 6501.5 but 16501.5(same idea for reckless9)
You are right, the "1" fell off during editing to get the 2 versions fit in code space, fixed it.
90% of coding is debugging, the other 10% is writing bugs.
-
Rebel
- Posts: 7539
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Post
by Rebel »
Jouni wrote: ↑Fri Apr 03, 2026 8:48 am
Overall gain 3-4 elo. But Stockfish has also gained 3-4 after version 18.
Look again, SF18
lost 3-4 elo.
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED W D L D(%)
2 Stockfish-18 : 3805.2 2.0 20301.0 33000 9727 21148 2125 64
See the not yet updated page :
https://rebel7775.wixsite.com/rebel/stc-rating-list90% of coding is debugging, the other 10% is writing bugs.
-
Jouni
- Posts: 3874
- Joined: Wed Mar 08, 2006 8:15 pm
- Full name: Jouni Uski
Post
by Jouni »
I hope You write exact test conditions to page. 1 core test. And book is?
Jouni
-
cc2150dx
- Posts: 461
- Joined: Sat Nov 30, 2013 9:51 am
- Full name: Jason Coombs
Post
by cc2150dx »
Hey Ed, Leela-0.32.1-BT4 What settings are you using ? and which BT4 net ?
-
Rebel
- Posts: 7539
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Post
by Rebel »
cc2150dx wrote: ↑Fri Apr 03, 2026 5:36 pm
Hey Ed,
Leela-0.32.1-BT4 What settings are you using ? and which BT4 net ?
1. Default settings
2. BT4-it332 net
90% of coding is debugging, the other 10% is writing bugs.
-
Rebel
- Posts: 7539
- Joined: Thu Aug 18, 2011 12:04 pm
- Full name: Ed Schröder
Post
by Rebel »
Jouni wrote: ↑Fri Apr 03, 2026 12:41 pm
I hope You write exact test conditions to page. 1 core test. And book is?
Good points
1. Single thread
2. 1500 openings
Updated the page accordingly.
https://rebel7775.wixsite.com/rebel/stc-rating-list90% of coding is debugging, the other 10% is writing bugs.
-
cc2150dx
- Posts: 461
- Joined: Sat Nov 30, 2013 9:51 am
- Full name: Jason Coombs
Post
by cc2150dx »
Rebel wrote: ↑Sun Apr 05, 2026 7:37 am
cc2150dx wrote: ↑Fri Apr 03, 2026 5:36 pm
Hey Ed,
Leela-0.32.1-BT4 What settings are you using ? and which BT4 net ?
1. Default settings
2. BT4-it332 net
Thanks
