Source: https://github.com/official-stockfish/S ... 17807fdebaAuthor: Jerry Donald Watson
Date: Tue Apr 10 17:47:55 2018 +0200
Timestamp: 1523375275
Simplify condition in space definition
We remove an unnecessary condition in the definition of safe squares
in the space evaluation. Only the squares which are occupied by our
pawns or attacked by our opponent's pawns are now excluded.
STC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 21096 W: 4321 L: 4199 D: 12576
http://tests.stockfishchess.org/tests/v ... 547e537d4e
LTC:
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 23437 W: 3577 L: 3460 D: 16400
http://tests.stockfishchess.org/tests/v ... 547e537d6a
It may be possible to further refine the definition of such safe squares.
Bench: 5351765
LCZero progress
Moderators: hgm, Rebel, chrisw
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
BTW, Stockfish 130418 is downloaded from chess.ultimaiq.net and I think 130418 is a date of compilation, but original version has been out in 100418:
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
All games published in this topic are played without using GPU.FWCC wrote:What type of graphics card
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED (%)
1 Stockfish 130418 D30 : 1058.3 113.9 193.0 200 97
2 Stockfish 130418 D15 : 472.0 72.5 938.5 1272 74
3 LC0 173_101eeb41 N32k : 262.1 99.3 17.0 72 24
4 LC0 173_101eeb41 N16k : 223.0 85.7 39.5 200 20
5 LC0 173_101eeb41 N8k : 180.9 90.1 32.5 200 16
6 LC0 173_101eeb41 N4k : 105.8 91.7 22.5 200 11
7 LC0 173_101eeb41 N2k : 39.4 95.7 16.0 200 8
8 LC0 173_101eeb41 N1k : 0.0 ---- 13.0 200 7
White advantage = 43.93 +/- 11.31
Draw rate (equal opponents) = 50.00 % +/- 0.00
-
- Posts: 1339
- Joined: Fri Nov 02, 2012 9:43 am
- Location: New Delhi, India
Re: LCZero progress
Then what's the Point ?majortom wrote:All games published in this topic are played without using GPU.
i7 5960X @ 4.1 Ghz, 64 GB G.Skill RipJaws RAM, Twin Asus ROG Strix OC 11 GB Geforce 2080 Tis
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
What are some people doing here if they don't see the point?
There are many other topics where it's easy to find the point.
There are many other topics where it's easy to find the point.
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED (%)
1 Stockfish 130418 D30 : 1058.9 133.4 193.0 200 97
2 Stockfish 130418 D15 : 472.3 69.8 977.0 1324 74
3 LC0 173_101eeb41 N32k : 272.3 83.5 30.5 124 25
4 LC0 173_101eeb41 N16k : 223.1 85.1 39.5 200 20
5 LC0 173_101eeb41 N8k : 181.0 85.9 32.5 200 16
6 LC0 173_101eeb41 N4k : 105.8 81.9 22.5 200 11
7 LC0 173_101eeb41 N2k : 39.4 100.9 16.0 200 8
8 LC0 173_101eeb41 N1k : 0.0 ---- 13.0 200 7
White advantage = 45.24 +/- 10.59
Draw rate (equal opponents) = 50.00 % +/- 0.00
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
[pgn][Event "?"]
[Site "?"]
[Date "2018.04.30"]
[Round "32"]
[White "Stockfish 130418 D15"]
[Black "LC0 173_101eeb41 N32k"]
[Result "0-1"]
[BlackTimeControl "inf"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1"]
[GameEndTime "2018-04-30T23:39:12.075 RTZ 2 "]
[GameStartTime "2018-04-30T22:50:20.281 RTZ 2"]
[PlyCount "64"]
[SetUp "1"]
[WhiteTimeControl "inf"]
1. d4 {book} Nf6 {book} 2. Nc3 {book} d5 {book} 3. Bg5 {book} Nbd7 {book}
4. Qd3 {book} c5 {book} 5. dxc5 {-0.05/15 0.96s} Nxc5 {+0.33/23 439s}
6. Qb5+ {+0.44/15 0.32s} Nfd7 {+0.30/22 25s} 7. Nxd5 {0.00/15 0.23s}
a6 {+0.36/22 95s} 8. Qc4 {-0.37/15 0.33s} b5 {+0.34/22 99s}
9. Qh4 {-0.01/15 0.27s} Bb7 {+0.40/22 203s} 10. e4 {-0.17/15 0.28s}
h6 {+0.48/22 232s} 11. f3 {+0.01/15 0.47s} Bxd5 {+1.43/22 248s}
12. exd5 {-0.83/15 0.35s} hxg5 {+1.48/22 11s} 13. Qxh8 {-0.79/15 0.30s}
Qa5+ {+1.54/22 7.3s} 14. c3 {-1.27/15 0.24s} b4 {+1.73/22 120s}
15. Rd1 {-1.15/15 0.67s} bxc3 {+1.64/22 37s} 16. bxc3 {-2.10/15 0.62s}
Qxc3+ {+1.65/22 2.5s} 17. Kf2 {-1.64/15 0.13s} Rb8 {+1.62/22 41s}
18. Be2 {-2.21/15 0.10s} Rb2 {+1.62/23 223s} 19. Kf1 {-3.07/15 0.61s}
Rd2 {+1.52/22 232s} 20. Rxd2 {-2.90/15 0.68s} Qxd2 {+1.92/21 58s}
21. Qh3 {-3.86/15 0.86s} Nd3 {+2.36/21 171s} 22. Bxd3 {-4.20/15 0.22s}
Qxd3+ {+2.27/21 8.6s} 23. Kf2 {-4.32/15 0.17s} e6 {+2.90/22 202s}
24. Qh8 {-3.66/15 0.27s} Qd4+ {+3.69/21 34s} 25. Kg3 {-4.08/15 0.10s}
Qf4+ {+5.42/21 173s} 26. Kf2 {-5.57/15 0.29s} Nc5 {+5.67/21 88s}
27. Kf1 {-7.08/15 0.24s} Qe3 {+12.48/20 113s} 28. dxe6 {-6.68/15 0.23s}
Nd3 {+46.74/17 19s} 29. exf7+ {-M6/15 0.005s} Kxf7 {+72.13/17 13s}
30. Qh5+ {-M6/15 0.004s} Kg8 {+68.41/18 17s} 31. Qh4 {-M4/15 0.001s}
gxh4 {+78.10/16 8.5s} 32. a3 {-M2/15 0.001s} Qf2# {+120.04/15 4.3s, Black mates}
0-1
[/pgn]
[Site "?"]
[Date "2018.04.30"]
[Round "32"]
[White "Stockfish 130418 D15"]
[Black "LC0 173_101eeb41 N32k"]
[Result "0-1"]
[BlackTimeControl "inf"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1"]
[GameEndTime "2018-04-30T23:39:12.075 RTZ 2 "]
[GameStartTime "2018-04-30T22:50:20.281 RTZ 2"]
[PlyCount "64"]
[SetUp "1"]
[WhiteTimeControl "inf"]
1. d4 {book} Nf6 {book} 2. Nc3 {book} d5 {book} 3. Bg5 {book} Nbd7 {book}
4. Qd3 {book} c5 {book} 5. dxc5 {-0.05/15 0.96s} Nxc5 {+0.33/23 439s}
6. Qb5+ {+0.44/15 0.32s} Nfd7 {+0.30/22 25s} 7. Nxd5 {0.00/15 0.23s}
a6 {+0.36/22 95s} 8. Qc4 {-0.37/15 0.33s} b5 {+0.34/22 99s}
9. Qh4 {-0.01/15 0.27s} Bb7 {+0.40/22 203s} 10. e4 {-0.17/15 0.28s}
h6 {+0.48/22 232s} 11. f3 {+0.01/15 0.47s} Bxd5 {+1.43/22 248s}
12. exd5 {-0.83/15 0.35s} hxg5 {+1.48/22 11s} 13. Qxh8 {-0.79/15 0.30s}
Qa5+ {+1.54/22 7.3s} 14. c3 {-1.27/15 0.24s} b4 {+1.73/22 120s}
15. Rd1 {-1.15/15 0.67s} bxc3 {+1.64/22 37s} 16. bxc3 {-2.10/15 0.62s}
Qxc3+ {+1.65/22 2.5s} 17. Kf2 {-1.64/15 0.13s} Rb8 {+1.62/22 41s}
18. Be2 {-2.21/15 0.10s} Rb2 {+1.62/23 223s} 19. Kf1 {-3.07/15 0.61s}
Rd2 {+1.52/22 232s} 20. Rxd2 {-2.90/15 0.68s} Qxd2 {+1.92/21 58s}
21. Qh3 {-3.86/15 0.86s} Nd3 {+2.36/21 171s} 22. Bxd3 {-4.20/15 0.22s}
Qxd3+ {+2.27/21 8.6s} 23. Kf2 {-4.32/15 0.17s} e6 {+2.90/22 202s}
24. Qh8 {-3.66/15 0.27s} Qd4+ {+3.69/21 34s} 25. Kg3 {-4.08/15 0.10s}
Qf4+ {+5.42/21 173s} 26. Kf2 {-5.57/15 0.29s} Nc5 {+5.67/21 88s}
27. Kf1 {-7.08/15 0.24s} Qe3 {+12.48/20 113s} 28. dxe6 {-6.68/15 0.23s}
Nd3 {+46.74/17 19s} 29. exf7+ {-M6/15 0.005s} Kxf7 {+72.13/17 13s}
30. Qh5+ {-M6/15 0.004s} Kg8 {+68.41/18 17s} 31. Qh4 {-M4/15 0.001s}
gxh4 {+78.10/16 8.5s} 32. a3 {-M2/15 0.001s} Qf2# {+120.04/15 4.3s, Black mates}
0-1
[/pgn]
-
- Posts: 4190
- Joined: Wed Nov 25, 2009 1:47 am
Re: LCZero progress
Wow, if that is really reliable performance scaling against much stronger engine seems not so great. Roughly 60Elo per time/node doubling. At the level of play of LC0 that is much worse than SF.majortom wrote:Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED (%) 1 Stockfish 130418 D30 : 1057.2 123.1 193.0 200 97 2 Stockfish 130418 D15 : 471.5 68.4 918.5 1248 74 3 LC0 173_101eeb41 N32k : 295.7 99.2 13.0 48 27 4 LC0 173_101eeb41 N16k : 222.8 84.8 39.5 200 20 5 LC0 173_101eeb41 N8k : 180.8 79.1 32.5 200 16 6 LC0 173_101eeb41 N4k : 105.7 94.9 22.5 200 11 7 LC0 173_101eeb41 N2k : 39.4 97.9 16.0 200 8 8 LC0 173_101eeb41 N1k : 0.0 ---- 13.0 200 7 White advantage = 41.43 +/- 11.51 Draw rate (equal opponents) = 50.00 % +/- 0.00
-
- Posts: 669
- Joined: Mon Nov 04, 2013 10:19 pm
Re: LCZero progress
Code: Select all
# PLAYER : RATING ERROR POINTS PLAYED (%)
1 Stockfish 130418 D30 : 1058.4 137.5 193.0 200 97
2 Stockfish 130418 D15 : 472.0 75.0 992.0 1340 74
3 LC0 173_101eeb41 N32k : 252.2 85.8 31.5 140 23
4 LC0 173_101eeb41 N16k : 223.0 87.0 39.5 200 20
5 LC0 173_101eeb41 N8k : 180.9 89.9 32.5 200 16
6 LC0 173_101eeb41 N4k : 105.8 92.9 22.5 200 11
7 LC0 173_101eeb41 N2k : 39.4 100.5 16.0 200 8
8 LC0 173_101eeb41 N1k : 0.0 ---- 13.0 200 7
White advantage = 44.10 +/- 10.54
Draw rate (equal opponents) = 50.00 % +/- 0.00