ShashChess

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

Moderators: hgm, chrisw, Rebel

amchess
Posts: 349
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

amchess
Posts: 349
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

amchess
Posts: 349
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

amchess
Posts: 349
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

ShashChess GZ

https://github.com/amchess/ShashChess/releases/tag/GZ
Completely renewed chess match conditions (20+200 games (LTC))
https://github.com/amchess/ShashChess/wiki/Match
A great thanks to the italian chess expert Afro Ambanelli.
The match on 200 games can't be always done, but in this case we made it and we obtained the following result:
https://github.com/amchess/ShashChess/wiki/Matches
Hard Positions Test Suite
https://github.com/amchess/ShashChess/w ... te-results

Any true statistician can conclude the validity of Shashin's theory applied to Stockfish.
Renewing thanks to the related community,
the ShashChess derivative is confirmed to be stronger than the original, at least at LTC
Eduard
Posts: 1439
Joined: Sat Oct 27, 2018 12:58 am
Location: Germany
Full name: N.N.

Re: ShashChess

Post by Eduard »

Thank You, Andrea!

This engine is now one of the best analysis engines ever. In my test, the engine solves 100 out of 110, and it seems that with just a little bit more time than 60s, even all positions could be solved.

Here are two examples where other engines have problems:

Diagram Pos 66:
Image

4K1k1/8/1p5p/1Pp3b1/8/1P3P2/P1B2P2/8 w - - 0 1

f4! Solved in 27.58s

Diagram Pos 73:
Image



Solved in 4.98s

https://solistachess.jimdosite.com/testing/
Eduard
Posts: 1439
Joined: Sat Oct 27, 2018 12:58 am
Location: Germany
Full name: N.N.

Re: ShashChess

Post by Eduard »

Thank You, Andrea!

This engine is now one of the best analysis engines ever. In my test, the engine solves 100 out of 110, and it seems that with just a little bit more time than 60s, even all positions could be solved.

Here are two examples where other engines have problems:

Diagram Pos 66:
Image

4K1k1/8/1p5p/1Pp3b1/8/1P3P2/P1B2P2/8 w - - 0 1

f4! Solved in 27.58s

Diagram Pos 73:
Image

8/8/3k2p1/5ppb/3P4/1P6/1P3NP1/4K3 w - - 0 1

g4! Solved in 4.98s

https://solistachess.jimdosite.com/testing/
Werewolf
Posts: 1902
Joined: Thu Sep 18, 2008 10:24 pm

Re: ShashChess

Post by Werewolf »

At LTC are we genuinely saying ShashChess > SF Dev/Lc0?
CornfedForever
Posts: 645
Joined: Mon Jun 20, 2022 4:08 am
Full name: Brian D. Smith

Re: ShashChess

Post by CornfedForever »

Werewolf wrote: Mon May 01, 2023 10:32 am At LTC are we genuinely saying ShashChess > SF Dev/Lc0?
Perhaps it depends on if one is a "true statistician" (?) :roll:
Eduard
Posts: 1439
Joined: Sat Oct 27, 2018 12:58 am
Location: Germany
Full name: N.N.

Re: ShashChess

Post by Eduard »

ShashChess GZ on TOP on PlayChess!

Ranking today:
Image

Place 1 with 16 Cores.
Place2 with 64 Cores.

Example: The ability is to win good positions, it's not easy.

[Event "5 min, rated"]
[Site "Engine Room"]
[Date "2023.05.03"]
[Round "?"]
[White "Ultima, Dark SisTer 4.6"]
[Black "Detlef Uter, ShashChess GZ"]
[Result "0-1"]
[ECO "A00"]
[WhiteElo "3009"]
[BlackElo "3047"]
[Annotator "-0.29;-0.28"]
[PlyCount "160"]
[EventDate "2023.05.03"]
[EventType "blitz"]
[TimeControl "300"]

1. c3 {B 0} d5 {B 0} 2. d3 {B 0} Nf6 {-0.28/35 12} 3. Nf3 {B 0} c5 {-0.27/31 3} 4. d4 {B 0} g6 {B 0} 5. Bf4 {B 0} Nc6 {B 0} 6. e3 {B 0} Nh5 {B 0} 7. dxc5 {B 0} Nxf4 {B 0} 8. exf4 {B 0} e6 {B 0} 9. Nbd2 {B 0} Bxc5 {B 0} 10. Nb3 {B 0} Bb6 {B 0} 11. Be2 {B 0} O-O {-0.27/31 13} 12. O-O {B 0} f6 {-0.23/31 2} 13. Qd2 {B 0} e5 {-0.29/30 3} 14. Rad1 {-0.29/30 14 (a4)} e4 {B 0} 15. Ne1 {-0.21/26 5} Be6 {-0.22/30 6} 16. Nc2 {-0.22/25 1} Qe7 {-0.21/29 3 (De8)} 17. Nbd4 {-0.22/28 12} Bd7 {-0.18/35 2 (Tad8)} 18. Nb3 {-0.26/28 6} Be6 {-0.20/36 0} 19. a4 {-0.20/27 0 (bd4)} a5 {-0.20/35 13 (Tad8)} 20. Nbd4 {-0.23/24 4} Rfd8 {-0.26/28 0} 21. Qe3 {-0.20/28 23} Nxd4 {-0.18/36 11} 22. Nxd4 {-0.22/27 0} Kh8 {-0.19/33 1} 23. Rd2 {-0.25/27 28 (g3)} Rg8 {-0.30/25 4} 24. Rfd1 {-0.22/28 13 (g3)} Bf7 {-0.26/30 10 (Tad8)} 25. Qh3 {-0.26/29 25 (g3)} Rad8 {-0.38/27 3} 26. Nc2 {-0.32/24 0 (g3)} Qd6 {-0.26/35 30 (Ae6)} 27. Qh6 {-0.31/28 26 (g3)} Qc7 {-0.45/25 3} 28. g3 {-0.38/29 12} f5 {-0.55/31 1} 29. Qg5 {-0.36/28 9} Rd6 {-0.58/32 0} 30. Nd4 {-0.38/26 1 (b5)} Be8 {-0.69/25 3} 31. Bb5 {-0.38/26 1} Bc6 {-0.69/26 1} 32. c4 {-0.36/23 1 (c2)} Qg7 {-0.78/26 2 (Tgd8)} 33. cxd5 {-0.96/28 11} e3 {-0.89/41 0} 34. fxe3 {-0.77/27 2} Bxb5 {-0.97/34 1} 35. Nxb5 {-0.79/27 2} Bxe3+ {-1.01/34 1} 36. Kf1 {-0.67/28 1} Bxd2 {-0.95/39 6} 37. Nxd6 {-0.81/30 0} Bb4 {-1.09/43 27} 38. Nxf5 {-0.64/35 0} gxf5 {-1.12/33 2} 39. Qxf5 {-0.61/28 1} Qxb2 {-1.13/32 1} 40. d6 {-0.79/28 5} Bc3 {-1.15/35 0} 41. Qd3 {-0.87/28 4} Bf6 {-1.14/35 0} 42. Qe2 {-0.84/29 3} Qc3 {-1.21/34 2} 43. d7 {-0.81/27 1} Qc6 {-1.23/32 3} 44. Kf2 {-0.84/27 2} h6 {-1.25/31 1} 45. Rd2 {-0.83/30 9} Bd8 {-1.27/34 0} 46. Qe5+ {-0.87/31 1} Kh7 {-1.20/35 5} 47. Rd6 {-0.91/34 6} Qc2+ {-1.04/34 2} 48. Kf1 {-0.88/33 3} Qb1+ {-1.30/33 0} 49. Kf2 {-0.87/33 1} Rg6 {-0.64/37 16} 50. Rxg6 {-0.81/35 0} Qxg6 {-0.58/36 7} 51. Qe8 {-0.82/33 1} Qf6 {-0.50/33 2} 52. Ke3 {-0.82/28 4 (e2)} Kg7 {-0.46/31 3} 53. Kd3 {-1.03/29 5} Qd6+ {-0.44/32 1} 54. Ke3 {-1.14/28 10 (e4)} Qe7+ {-1.35/33 5} 55. Kd4 {-1.17/30 0} Bb6+ {-1.39/32 0} 56. Kd3 {-1.22/27 2 (d5)} Kf6 {-1.78/28 1} 57. Qh8+ {-1.25/26 0} Kf5 {-2.08/26 1} 58. Qc8 {-1.21/26 0} Bd8 {-2.23/29 1} 59. Qxb7 {-1.33/28 0} Qd6+ {-2.27/27 1} 60. Ke2 {-1.38/26 0} Ke6 {-2.46/26 2} 61. Qb5 {-1.70/24 6 (h4)} Qxd7 {-2.86/26 4} 62. Qe5+ {-1.89/25 0} Kf7 {-2.88/26 0} 63. Qh5+ {-1.92/23 0} Kg7 {-2.95/23 0} 64. Qe5+ {-1.92/25 3} Kg8 {-3.07/29 0} 65. Ke3 {-1.97/21 0 (f1)} Qe7 {-199.33/33 3 (Dc6)} 66. Qxe7 {-2.33/23 1 (d4)} Bxe7 {-199.53/48 2} 67. Kd4 {-2.94/25 2 (e4)} Kg7 {-199.57/53 1} 68. Ke5 {-3.63/28 3} Kg6 {-199.59/58 0 (Rf7)} 69. g4 {-3.63/26 1} Ba3 {-199.61/54 5} 70. Kd4 {-4.03/29 0 (e6)} Kf6 {-199.64/52 2 (Ab4)} 71. h3 {-4.35/26 1 (e3)} Ke6 {-199.69/53 7 (Ac1)} 72. Kc4 {-5.30/29 3 (d3)} Bc1 {-199.77/41 2 (Rf6)} 73. f5+ {-5.31/30 1 (b5)} Kf6 {-200.00/40 5} 74. Kd5 {-5.90/31 0 (b5)} Ba3 {-200.00/38 1 (Ad2)} 75. Kd4 {-6.05/23 1 (h4)} Kg5 {-200.00/34 1 (Ab4)} 76. Ke5 {-6.83/24 1 (e3)} h5 {-200.00/37 2 (Ab2+)} 77. gxh5 {-5.77/16 0} Bb2+ {-200.00/41 0 (Rxh5)} 78. Ke6 {-7.10/24 1 (e4)} Kxh5 {-200.00/37 1} 79. f6 {-10.32/28 1 (f7)} Kg6 {-200.00/41 2} 80. Kd6 {-12.08/25 0 (f7)} Bxf6 {-200.00/30 1 (Rxf6) Ultima,Dark SisTer 4.6 abbandona (Lag: Av=0.27s, max=0.8s)} 0-1
Eduard
Posts: 1439
Joined: Sat Oct 27, 2018 12:58 am
Location: Germany
Full name: N.N.

Re: ShashChess

Post by Eduard »

Hello Andrea!

I would like to play with Brainlearn/ShashChess on InfinityChess, and use an existing learning file. Unfortunately that is not possible. ShashChess does not create a learning file on InfinityChess. Instead, an "experience_new.bin" file is created. This file is always created when no normal learning file is possible. Even after a game on InfinityChess, this learning file only had 0 bytes!

Learning files are created on PlayChess and Shredder Classic, but not on InfinityChess. A possible reason could be the following:

InfinityChess GUI always needs the full path to the learning file.

This is what it looks like when I use an engine with the Eman learning file:
C:\Users\eduar\OneDrive\Documents\InfinityChess\Data\Engines\Solista.exp

Without the complete path, no learning file will be created there either, I have to enter the complete path in the engine options.

With ShashChess, however, it is not possible to enter the full path because ShashChess' UCI options do not offer this. Andrea, could you change that? I would like to try a private testsversion.

Regards
Eduard