Latest Stockfish...

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

another update - includes two additional pull requests that I am certain will be merge at some point in the future

Code: Select all

600 rounds and 1800 games completed...
time control: 3000+50
Date: 08/15/16 : 03:46:17
1800 game(s) loaded
Rank Name                 Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR 
------------------------------------------------------------------------------------------------------
   1 Stockfish 081416-2y   3127   0.0    8    8  1200  669.5  55.8  326  187  687  27.2  57.2  3087 
   2 Stockfish 081116-2y   3120   6.4    8    8  1200  650.0  54.2  302  202  696  25.2  58.0  3090 
   3 Komodo 10.1 64-bit    3053  67.3    8    8  1200  480.5  40.0  191  430  579  15.9  48.2  3123 
------------------------------------------------------------------------------------------------------
the first one is by VoyagerOne and is titled "Use predicted depth for history pruning" and the second one is by ElbertoOne and is the start of incorporating fortress coding in SF and will help play/analysis in positions like this:
.

[d]6k1/4B3/5P2/7K/4b2P/8/8/8 b - - 0 55

[d]6k1/5p1p/5PpP/2Bb4/8/8/1P6/6K1 b - - 0 39

Both positions came up in recent TCEC games

As always, source and Mac exe are available here:

https://github.com/MichaelB7/Stockfish/tree/master/src
User avatar
Nordlandia
Posts: 2843
Joined: Fri Sep 25, 2015 9:38 pm
Location: Sortland, Norway

Re: Latest Stockfish...

Post by Nordlandia »

H4 Pro | Contempt 0 vs SF Dev with odds

Conditions:
a7+h7 removed with compromise, white is not allowed to castle either sides. Black possesses castling rights, although it's very unlikely the black monarch will make use of his the special move.

[pgn][Event "EXPERT-PC, Rapid 30m+22s"]
[Site "EXPERT-PC"]
[Date "2016.08.15"]
[Round "1"]
[White "Houdini 4 Pro x64"]
[Black "Stockfish 150816 64 BMI2"]
[Result "1/2-1/2"]
[SetUp "1"]
[FEN "rnbqkbnr/1pppppp1/8/8/8/8/PPPPPPPP/RNBQKBNR w kq - 0 1"]
[PlyCount "190"]

1. d4 Nf6 2. c4 e5 3. dxe5 Ng4 4. Nc3 Nxe5 (4... Bc5 5. e3 {0.98/23 Stockfish
7 64}) 5. Nf3 Nbc6 (5... Nbc6 6. Nd5 {0.99/23 Stockfish 7 64 POPCNT}) 6. Nxe5 (
6. Nd5 Nxf3+ 7. gxf3 {0.91/23 Komodo 10.1 64-bit}) 6... Nxe5 (6... Nxe5 7. Qd4
d6 {1.45/21 Deep Fritz 13}) 7. Bf4 d6 (7... Ng6 8. Be3 Ne5 {0.87/21 Deep Fritz
14 x64}) 8. Bxe5 (8. Bxe5 dxe5 {0.77/27 Komodo 10 64-bit}) 8... dxe5 (8... dxe5
9. Qxd8+ Kxd8 {0.99/29 Stockfish 5 64 SSE4.2}) 9. Qxd8+ Kxd8 (9... Kxd8 10.
Rd1+ Bd7 {1.37/19 Deep Rybka 4 w32}) 10. Rd1+ Bd7 (10... Bd7 11. a3 c6 {
0.72/29 Komodo 10.1 64-bit}) 11. a3 c6 (11... c6 12. e3 Kc7 {1.02/23 Deep
Fritz 14 x64}) 12. e3 Be7 (12... Kc7 13. Kd2 Be7 {1.03/22 Deep Fritz 14 x64})
13. h3 Kc7 14. Kd2 f5 (14... g5 15. Be2 Be6 {-1.43/21 Fritz 13}) 15. Be2 Be6 (
15... Be6 16. Kc2 Rh6 {0.94/28 Stockfish 160716 64 POPCNT}) 16. Kc2 (16. Kc2 e4
17. Rhg1 {0.82/34 Komodo 10 64-bit}) 16... Bf6 (16... g6 17. a4 Rh7 {0.77/24
Stockfish 5 x64}) 17. Rhe1 Be7 (17... Be7 18. Nb1 Ra4 {0.84/29 Stockfish 7 64})
18. Rd2 (18. Rd2 Rad8 19. Rxd8 {1.23/22 Deep Fritz 14 x64}) 18... Rh6 (18...
Rad8 19. Rxd8 Bxd8 {1.20/23 Deep Fritz 14 x64}) 19. Red1 Rg6 (19... Bg5 20. c5
{0.88/29 Komodo 10.1 64-bit}) 20. Bf1 Rh6 (20... Rh6 21. Rc1 Kb6 {1.65/22
Fritz 13}) 21. Kb1 Bh4 22. Rc1 Kb6 (22... Kb8 23. g3 Be7 {0.83/28 Stockfish 5
64 SSE4.2}) 23. Bd3 (23. a4 Kc7 24. b3 {1.30/21 Deep Fritz 14 x64}) 23... Bg5 (
23... Be7 24. Bf1 Rhh8 {0.96/23 Houdini 4 x64}) 24. Kc2 (24. Bc2 Kc7 {0.91/32
Stockfish FASM}) 24... Kc7 (24... Be7 25. Kb1 Rd8 {0.91/25 Stockfish 7 64
POPCNT}) 25. Rcd1 Bf6 26. a4 Be7 (26... Be7 27. e4 g6 {0.59/22 Fire 4 x64}) 27.
b3 (27. b3 Bc5 28. Kb2 {0.64/26 Komodo 10.1 32-bit}) 27... Bb4 (27... Bc5 {
0.65/26 Stockfish 6 x64}) 28. Kb2 Bc5 29. Na2 Be7 (29... e4 30. Bf1 Be7 {
1.14/19 Fritz 11 SE}) 30. Rc2 e4 31. Bf1 b6 (31... b6 32. Nc3 Bc5 {0.89/25
Stockfish 7 64 POPCNT}) 32. Nc3 Bc5 (32... Bc5 33. g3 {0.67/28 Komodo 10.1
32-bit}) 33. Ne2 (33. g3 g5 34. Bg2 {0.83/31 Komodo 9.2 64-bit}) 33... Bb4 (
33... Bb4 34. Nc3 g5 {1.18/25 Fritz 15}) 34. Rcc1 g5 (34... g5 35. Nc3 Bc5 {
0.89/29 Stockfish 240716 64 POPCNT}) 35. Rc2 Rf8 (35... Rf8 36. Nc3 Bd6 {
0.81/28 Stockfish 5 64 SSE4.2}) 36. Nc3 (36. Nc3 Bd6 37. Rcd2 {1.42/23 Deep
Fritz 13}) 36... Bc5 (36... Bc5 37. g3 Bf7 {0.88/28 Stockfish 240716 64 POPCNT}
) 37. Rcd2 Bb4 (37... Rhh8 38. Be2 Bc8 {1.09/20 Deep Fritz 14}) 38. g3 Re8 (
38... Re8 39. Bg2 Rhh8 {0.72/26 Stockfish 7 64}) 39. Rc2 (39. Bg2 Bc8 40. Rc2 {
1.13/22 Deep Fritz 14 x64}) 39... Bf8 (39... Bf8 40. Ne2 Rh8 {1.15/19 Deep
Fritz 14 x64}) 40. Bg2 Bb4 (40... Bd7 41. Rcd2 Be6 {0.79/23 Komodo 8 64-bit})
41. Na2 Bd6 42. Rcd2 (42. Rcd2 Bf8 43. Kc2 {0.62/27 Komodo 10.1 64-bit}) 42...
Bf8 (42... Be5+ 43. Nc3 Reh8 {1.40/25 Deep Fritz 13}) 43. Nc3 Rh8 (43... Bb4
44. Rc2 Bf7 {0.77/29 Stockfish 7 64}) 44. Ne2 (44. Ne2 Bb4 45. Rc2 {0.86/30
Stockfish 5 64 SSE4.2}) 44... Bb4 (44... Bb4 45. Rc2 Bf7 {0.67/27 Komodo 10
64-bit}) 45. Rc2 (45. Nc3 Bc5 46. Kc2 {1.60/26 Fritz 13}) 45... Bf7 (45... Rh6
46. Rcc1 Be7 {0.79/26 Komodo 9.02 64-bit}) 46. Rg1 (46. Nc3 Rh6 47. c5 {
0.67/25 Stockfish 6 64}) 46... Bd6 (46... Bc5 47. Rd1 Bb4 {0.78/26 Houdini 4
Pro x64}) 47. Ka2 (47. Ka2 Be6 48. Rh1 {0.70/22 Deep Junior Yokohama x64})
47... Rhg8 (47... Rhg8 48. Rd1 Bh5 {0.59/28 Komodo 10.1 64-bit + PB 2016 +
Full syzygy TB on SSD}) 48. Bh1 Rh8 (48... Bc5 49. Bg2 Bh5 {0.56/20 Fire 4 x64}
) 49. Nd4 (49. Bg2 {0.75/27 Stockfish 7 64}) 49... Rxh3 50. Nxf5 Bc5 (50... Bb4
51. g4 Bg6 {0.83/19 Deep Rybka 4 SSE42 x64}) 51. Rd2 (51. Nd4 Bg6 52. Bg2 {
0.67/31 Stockfish 7 64}) 51... Bb4 (51... Bg6 52. Nd4 Bb4 {0.79/32 Stockfish 5
64 SSE4.2}) 52. Rb2 Bg6 (52... Bg6 53. Nd4 Kb7 {0.81/24 Deep Fritz 14 x64}) 53.
Nd4 (53. Nd4 Kb7 54. Re2 {0.58/29 Stockfish 7 64 POPCNT}) 53... Re7 (53... Re7
54. Re2 Rh2 {-0.64/27 Houdini 4 x64}) 54. Rc2 (54. Rc2 Be8 55. Nf5 {0.73/34
SugaR v5.4b 64-bit POPCNT}) 54... Bc5 (54... Kb7 55. Bg2 Rh2 {1.18/25 Fritz 13}
) 55. Rd2 (55. Rd2 Bb4 56. Re2 {0.84/23 Deep Fritz 14 x64}) 55... Bb4 56. Re2 (
56. Rc2 Bc5 57. Bg2 {1.37/25 Houdini 4 Pro w32}) 56... Rh2 (56... Rh2 57. Bg2
Re8 {0.68/31 Stockfish 160716 64 POPCNT}) 57. Bg2 (57. Bg2 Bh5 58. Rc2 {
0.83/25 Deep Fritz 14 x64}) 57... Bh5 (57... Bh5 58. Rc2 Bg4 {0.66/32
Stockfish 7 64}) 58. Rc2 (58. Rc2 Bg4 {0.43/33 Komodo 10 64-bit}) 58... Bg4 (
58... Bg4 59. Kb2 Bd7 {0.58/34 Stockfish 7 64}) 59. Bh1 Rh8 60. Kb2 (60. Bg2
Rh2 61. Kb1 {0.81/22 Deep Fritz 14 x64}) 60... Ree8 (60... Ree8 61. Rf1 Bh3 {
0.63/30 Stockfish 7 64}) 61. Rf1 (61. Bg2 Rh2 62. f4 {0.45/33 Komodo 10.1
64-bit}) 61... Bd7 62. Rd1 (62. Ne2 Rh6 63. Rd1 {0.81/23 Deep Fritz 14 x64})
62... Rh2 (62... Rh6 63. Bg2 Rh2 {0.56/28 Stockfish 7 64 POPCNT}) 63. Rg1 (63.
Rf1 Rhh8 64. Rfc1 {0.89/25 Deep Fritz 13}) 63... Rh3 (63... Rh7 64. Ne2 Rh2 {
0.53/30 Komodo 9.2 64-bit}) 64. Ka2 (64. Ne2 Bd6 65. Nc3 {0.66/33 Stockfish 5
64 SSE4.2}) 64... Bc5 (64... Bc5 65. Bg2 Rh2 {1.27/24 Houdini 4 Pro w32}) 65.
Rd2 (65. Rf1 Bd6 66. Bg2 {0.53/26 Komodo 8 64-bit}) 65... Bb4 (65... Rh2 66.
Nc2 Rf8 {0.75/24 Deep Fritz 14 x64}) 66. Re2 Bg4 (66... Kb7 67. Kb2 Rh7 {
1.22/23 Fritz 13}) 67. Rc2 (67. Rc2 Bd6 68. Kb2 {0.72/20 Fritz 11 SE}) 67...
Bc5 68. Kb2 (68. Kb2 Bf8 {0.51/29 Stockfish 6 x64}) 68... Bb4 (68... Bf8 69.
Ne2 Bb4 {0.81/21 Deep Fritz 14 x64}) 69. Ra1 Bd7 (69... Rh7 70. Rg1 Bd7 {
0.68/34 Stockfish 5 64 SSE4.2}) 70. Ka2 Bc5 (70... Rh6 71. Rg1 Bd6 {1.22/22
Fritz 13}) 71. Re1 Rh2 (71... Bf8 72. Rf1 Rh8 {0.79/20 Critter 1.5c 64-bit})
72. f4 (72. Kb2 Bb4 73. Rg1 {0.54/32 Stockfish 7 64 POPCNT}) 72... Rh5 (72...
Rh3 73. fxg5 Rxg3 {0.78/24 Deep Fritz 14}) 73. Bg2 gxf4 (73... Bxd4 74. exd4 e3
{0.38/22 Houdini 4 x64}) 74. gxf4 Bg4 75. Bh1 Rh4 (75... Rh4 76. f5 Bxd4 {
0.48/21 Fritz 11 SE}) 76. f5 (76. f5 Bxd4 77. exd4 {0.49/25 Komodo 10.1 32-bit}
) 76... Bxd4 (76... Bxd4 77. exd4 Bxf5 {0.82/34 Stockfish 5 64 SSE4.2}) 77.
exd4 (77. exd4 Bxf5 78. c5 {0.41/27 Komodo 10.1 64-bit}) 77... Bxf5 (77... Bxf5
78. c5 Rd8 {0.58/29 Stockfish 7 64 POPCNT}) 78. Rf2 (78. Rf2 Bd7 79. Rfe2 {
0.57/25 Fritz 13}) 78... Bd7 (78... Bd7 79. Rfe2 e3 {0.36/30 Stockfish 7 64})
79. Rfe2 (79. Rfe2 e3 80. Bf3 {0.03/29 Stockfish 240716 64 POPCNT}) 79... e3 (
79... e3 80. Bf3 Rxd4 {0.96/28 Fritz 15}) 80. Bf3 Rxd4 (80... Bg4 81. Bxg4 Rxg4
{0.62/24 Deep Fritz 14 x64}) 81. Rxe3 (81. Rxe3 Rxe3 {0.09/31 Stockfish 7 64
POPCNT}) 81... Rxe3 (81... Rxe3 82. Rxe3 Rh4 {0.36/24 Fire 4 x64}) 82. Rxe3 (
82. Rxe3 b5 83. axb5 {0.44/31 Stockfish 5 64 SSE4.2}) 82... b5 (82... b5 83.
axb5 cxb5 {0.27/36 Stockfish 7 64}) 83. cxb5 (83. cxb5 cxb5 84. a5 {0.00/34
Komodo 9.2 64-bit}) 83... cxb5 (83... cxb5 84. axb5 Bxb5 {0.13/32 Stockfish
240716 64 POPCNT}) 84. a5 (84. a5 b4 85. Re7 {1.11/21 Deep Rybka 4 w32}) 84...
b4 (84... b4 85. Re7 Kd6 {0.37/26 Deep Fritz 14 x64}) 85. Kb2 (85. Kb2 Rd2+ 86.
Kc1 {0.12/33 Stockfish 7 64}) 85... Rd2+ (85... Rd2+ 86. Ka1 Bc6 {0.00/44
Stockfish 5 64 SSE4.2}) 86. Ka1 (86. Kb1 Bf5+ 87. Be4 {0.33/25 Deep Fritz 14
x64}) 86... Bf5 (86... Rc2 {0.29/30 Komodo 9.02 64-bit}) 87. Re7+ (87. Re5 Bc2
88. Rc5+ {0.07/34 Stockfish 7 64}) 87... Kd6 (87... Kd6 88. Rf7 Bc2 {0.07/36
Stockfish 7 64}) 88. Re1 (88. Rf7 Bc2 89. Rb7 {0.00/34 Stockfish 7 64}) 88...
Rd3 (88... Rd3 89. Rd1 Kc5 {0.00/29 Deep Fritz 13}) 89. Rd1 (89. Rd1 Kc5 90.
Rxd3 {0.00/38 Stockfish 5 x64}) 89... Kc5 (89... Kc5 90. Rxd3 Bxd3 {0.00/39
Stockfish 7 64 POPCNT}) 90. Rxd3 (90. Rxd3 Bxd3 91. Bb7 {0.00/38 Houdini 4 x64}
) 90... Bxd3 (90... Bxd3 91. Bb7 Kb5 {0.00/54 Stockfish 7 64}) 91. Ka2 Kb5 (
91... Kb5 92. Kb2 Kxa5 {-0.07/49 Stockfish 7 64 POPCNT}) 92. Kb2 (92. Kb2 Kxa5
93. Kc1 {0.00/58 Komodo 10 64-bit}) 92... Kxa5 93. Kc1 Bf1 94. Kd2 Bh3 95. Ke3
Be6 1/2-1/2

[/pgn]Lichess Analysis - https://en.lichess.org/a8vlKkvI
Dann Corbit
Posts: 12870
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Latest Stockfish...

Post by Dann Corbit »

For those Windows users not satisfied with the bleeding edge stockfish from http://abrok.eu/stockfish/ here is a "bleeding-er edge" stockfish with Mike's updates:

https://www.dropbox.com/s/y7rddx58nsjhz ... MB.7z?dl=0

It also has my logging stuff and a few extra UCI parameters tossed in.
You will probably want to turn off logging, which is on by default if you decide to use my fiddled version.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

Dann Corbit wrote:For those Windows users not satisfied with the bleeding edge stockfish from http://abrok.eu/stockfish/ here is a "bleeding-er edge" stockfish with Mike's updates:

https://www.dropbox.com/s/y7rddx58nsjhz ... MB.7z?dl=0

It also has my logging stuff and a few extra UCI parameters tossed in.
You will probably want to turn off logging, which is on by default if you decide to use my fiddled version.
Dann if you want me to send me your files, with the changes - will be happy to add them to this "rogue" version of SF on Github.
User avatar
Nordlandia
Posts: 2843
Joined: Fri Sep 25, 2015 9:38 pm
Location: Sortland, Norway

Re: Latest Stockfish...

Post by Nordlandia »

Odds: two pawns removed with compromise, - white is not allowed to castle.

Hash: 4096 MB allocated per engine
TBs: 6-Men
GUI: Fritz 15
OS: Windows 10
CPU: i7-5960X 4.1GHz
[pgn][Event "EXPERT-PC, Slow 60m+30s"]
[Site "https://lichess.org/Qad362tH"]
[Date "2016.08.15"]
[White "Houdini 4 Pro x64"]
[Black "Stockfish 150816 64 BMI2"]
[Result "1/2-1/2"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "132"]
[Variant "From Position"]
[TimeControl "-"]
[ECO "?"]
[Opening "?"]
[Termination "Normal"]
[FEN "rnbqkbnr/1pppppp1/8/8/8/8/PPPPPPPP/RNBQKBNR w kq - 0 1"]
[SetUp "1"]
[Annotator "lichess.org"]

1. d4 e6 2. Nf3 Nf6 3. h4 d5 4. Nc3 c5 5. e3 Bd7 6. Be2 Ne4 7. Kf1 Nxc3 8. bxc3 Bc6 9. c4 dxc4 10. Bxc4 Nd7 11. Qd3 Qb6 12. Kg1?! { (1.55 → 0.72) Inaccuracy. Best move was Ng5. } (12. Ng5 Nf6 13. c3 Qc7 14. Rh3 Be7 15. Kg1 Nd5 16. Rb1 Bxg5 17. hxg5 O-O-O 18. Bd2 Nb6) 12... Be7 13. c3 Bxf3 14. gxf3 Bxh4 15. Bd2 O-O-O 16. Rb1 Qc7 17. Bb5 Be7 18. Kg2 Ne5 19. Qe4 Nd7 20. Rxh8 Rxh8 21. f4 g5 22. fxg5 Bxg5 23. Kf1 Nf6 24. Qg2 Qh2 25. Bd3 Qh1+ 26. Qxh1 Rxh1+ 27. Ke2 Rxb1 28. Bxb1 Ne8 29. a4 Nd6 30. f4 Bh6 31. Kf3 Nc4 32. Be1 Nb6 33. Bc2 cxd4 34. cxd4 f5 35. e4 fxe4+ 36. Kxe4 Nd5 37. Bb3 Nc7 38. Bb4 Kd7 39. a5 Bg7 40. Ba4+ Kc8 41. Bd6 Bf6 42. Bc5 Nd5 43. Bb3 Nc7 44. Bd1 Kd7 45. Bf3 Bh4 46. Bb4 Nd5 47. Bd2 Ne7 48. Be2 Bf6 49. Bb5+ Kd6 50. Bb4+ Kc7 51. Bc4 Nc6 52. Be1 Kd6 53. Bxe6 Kxe6 54. d5+ Kd6 55. dxc6 Kxc6 56. Kf5 Bd4 57. Ke6 Kb5 58. f5 Bb2 59. Kd6 Bf6 60. Kc7 Ka6 61. Bd2 Be5+ 62. Kd7 Kb5 63. Ke8 Ka6 64. Ke7 Kb5 65. Kd7 Bf6 66. Bf4? { (1.17 → 0.15) Mistake. Best move was Kc7. } (66. Kc7 Ka6 67. Kd7 Be5 68. Ke6 Bg7 69. Kf7 Bd4 70. Ke7 Bg7 71. Kd6 Kb5 72. Kd5 Bf6) 66... Kxa5 { Draw } 1/2-1/2[/pgn]

https://en.lichess.org/Qad362tH
Dann Corbit
Posts: 12870
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Latest Stockfish...

Post by Dann Corbit »

MikeB wrote:
Dann Corbit wrote:For those Windows users not satisfied with the bleeding edge stockfish from http://abrok.eu/stockfish/ here is a "bleeding-er edge" stockfish with Mike's updates:

https://www.dropbox.com/s/y7rddx58nsjhz ... MB.7z?dl=0

It also has my logging stuff and a few extra UCI parameters tossed in.
You will probably want to turn off logging, which is on by default if you decide to use my fiddled version.
Dann if you want me to send me your files, with the changes - will be happy to add them to this "rogue" version of SF on Github.
The files are in the link I provided.
SF is GPL, so if you modify the source, you must also provide the changed source.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

the latest experimental SF:

https://github.com/MichaelB7/Stockfish/commits/master

This one has several pulls that have not been committed to as of yet, includng Ronald's syzygy tb update, some fortress coding and others including VoyagerOne's Refutation Penalty Pull #761 which is looking excellent at longer time control.. Mac exe's can be found in the zip files.


I also included a current Mac exe of the the official SF latest developmental version, this is the true blue current master based on the source below:

https://github.com/official-stockfish/S ... 3568e48918

This is my final run still in process at game 2.5 min with 2.5 sec incerement, but I have already run enough games at fast micro tc 's to know that I haven't screwed anything up too bad at least ...

Code: Select all

120 of 600 games completed...
time control: 15000+250
Date: 08/22/16 : 22:18:23
120 game(s) loaded
Rank Name                      Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR 
-------------------------------------------------------------------------------------------------
   1 Stockfish 082216-3y        3121   0.0   27   27    60   33.0  55.0   11    5   44  18.3  73.3  3093 
   2 Stockfish 081416-2y        3111  10.4   27   27    61   32.0  52.5    9    6   46  14.8  75.4  3096 
   3 Stockfish 210816 64 POPC   3101  10.4   28   28    59   29.5  50.0    7    7   45  11.9  76.3  3100 
   4 Komodo 10.1 64-bit         3067  33.4   28   28    60   25.5  42.5    7   16   37  11.7  61.7  3111 
-------------------------------------------------------------------------------------------------
Stockfish 210816 64 POPC is the official current master - it is also the best version at micro time control of 10 sec game/0.1 sec increment, Stockfish 081416-2y was the previous experimental, Stockfish 082216-3y is the current experimental that may play better at longer tc.

Follow the last run here - it updates every 12 minutes , only 600 games are scheduled - about 4 hours to go.

https://www.dropbox.com/s/zoh9vltb5udks ... a.txt?dl=1
User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

Latest Stockfish fork Mac exe
https://github.com/MichaelB7/Stockfish/ ... _MacOS.zip with source https://github.com/MichaelB7/Stockfish
and includes the latest commit "Move CheckInfo under StateInfo" and all other recent functional commits.

Code: Select all

In addition, Stockfish-210816 differs from the official SF Master by the following : 

1.  Incorporates Ronald de Man syzygy updates https://github.com/syzygy1/tb
2.  It has the fortress coding by ElbertoOne (PR # #747) with no loss in strength ( due to other enhancements) _
3.  it has a speedup in play based on the idea by joergosaer (PR # 740 - but implemented differently) - it will play much quicker by several factors once it has mate score or tb hit at the root - you can turn this off, if so desire by setting comtempt to any value other than zero (default).
4.  And it has my own simplification and refinement of Depth R at line 765 in search.cpp going from this 

Depth R = ((823 + 67 * depth) / 256 + std::min((eval - beta) / PawnValueMg, 3)) * ONE_PLY;  ( the values used here are absurd, suggesting a level of precision that simply does not exist - but I do like how they factor the score in their null move reductions)

to this:

Depth R = ((12 +  depth) / 5 + std::min((eval - beta) / PawnValueMg, 4 )) * ONE_PLY;

some current testing still in process

624 of 1200 games completed...
time control: 15000+250
Date: 08/27/16 : 11:00:10
624 game(s) loaded
Rank Name                      Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR 
------------------------------------------------------------------------------------------------------
   1 Stockfish 082616-4y        3119   0.0   15   15   313  169.0  54.0   61   36  216  19.5  69.0  3094 
   2 Stockfish 082716-5y        3114   5.2   15   15   313  165.0  52.7   64   47  202  20.4  64.5  3096 
   3 Stockfish 250816 64 POPC   3113   0.8   15   15   312  164.5  52.7   62   45  205  19.9  65.7  3096 
   4 Komodo 10.1 64-bit         3053  60.1   15   15   310  125.5  40.5   44  103  163  14.2  52.6  3116 
------------------------------------------------------------------------------------------------------
624 game(s) loaded

4y and 5 y are exactly the same except 5y is faster due to  latest commit "Move CheckInfo under StateInfo" so the fact that it is currently doing better than 5y is just noise - this also tested well at SF standard tc testing of  10/.1 and 60/.6
User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

Sorry , I was rushing (was heading out with the wife) and provided the link to previous official stockfish Mac binary - here's the link to version with the changes described above

https://github.com/MichaelB7/Stockfish/ ... 716-5y.zip

There is also more current stockfish binary on github as well dated 250816, will get that updated for the latest commits made today when I get back home - about 10-12 hours from now.
User avatar
MikeB
Posts: 4889
Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania

Re: Latest Stockfish...

Post by MikeB »

As promised = latest mac exe's can be found here:
https://github.com/MichaelB7/Stockfish enjoy!

source for current dev standard SF can be found here
https://github.com/official-stockfish/S ... 99f48ad5a4


just fyi, as expected all three engines converged

Code: Select all

200 rounds and 1200 games completed...
time control: 15000+250
Date: 08/27/16 : 19:07:04
1200 game(s) loaded
Rank Name                      Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR 
--------------------------------------------------------------------------------------------------
   1 Stockfish 082716-5y        3116   0.0   11   11   600  318.5  53.1  125   88  387  20.8  64.5  3095 
   2 Stockfish 082616-4y        3115   0.8   11   11   600  317.5  52.9  118   83  399  19.7  66.5  3095 
   3 Stockfish 250816 64 POPC   3115   0.6   11   11   600  317.5  52.9  122   87  391  20.3  65.2  3095 
   4 Komodo 10.1 64-bit         3054  60.6   11   11   600  246.5  41.1   93  200  307  15.5  51.2  3115 
---------------------------------------------------------------------------------------------------

Code: Select all

Curent SF 082716-5y does very well against the the current development standard SF-270816

200 rounds and 600 games completed...
time control: 1000+10
Date: 08/27/16 : 23:51:19
600 game(s) loaded
Rank Name                      Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR 
--------------------------------------------------------------------------------------------------
   1 Stockfish 082716-5y        3132   0.0   15   15   400  227.0  56.8  125   71  204  31.2  51.0  3084 
   2 Stockfish 270816 64 POPC   3117  15.1   15   15   400  213.5  53.4  106   79  215  26.5  53.8  3091 
   3 Komodo 10.1 64-bit         3051  66.6   16   16   400  159.5  39.9   79  160  161  19.8  40.2  3125 
--------------------------------------------------------------------------------------------------