ShashChess

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

Moderators: hgm, Rebel, chrisw

Eduard
Posts: 1439
Joined: Sat Oct 27, 2018 12:58 am
Location: Germany
Full name: N.N.

Re: ShashChess

Post by Eduard »

Hello Andrea, could you implement the new network size in Brainlearn (it doesn't have to be in Shashchess)? It will be a long time until the next official SF release. I implemented Dimension 3072 for myself and did some testing. I also carried out tests with MCTS, first in analysis mode and then also on Lichess while playing. The MCTS search is better for analysis. The larger network scores better, especially against weaker engines on Lichess.

A test at Lichess with D3072 and MCTS (1 thread). Matchmaking mode, 6 cores used:
Image
amchess
Posts: 340
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

ShashChess 35.1
https://github.com/amchess/ShashChess/releases/tag/35.1
Nodestime option
corrected futility pruning: parent node
amchess
Posts: 340
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

Alexander 1.1
https://github.com/amchess/Alexander/releases/tag/1.1
Nodestime option
corrected futility pruning: parent node
Damir
Posts: 2809
Joined: Mon Feb 11, 2008 3:53 pm
Location: Denmark
Full name: Damir Desevac

Re: ShashChess

Post by Damir »

Very good Andrea. The Engine plays very strong with small network. Who needs bigger nnue, when small one are just as good..
The engine runs more than 50% faster on my machine than the current Stockfish and other Stockfish based engines. :) :)
Jouni
Posts: 3321
Joined: Wed Mar 08, 2006 8:15 pm

Re: ShashChess

Post by Jouni »

ShashChess seems to "scale" better than Crystal. In my testsuite was 21 positions unsolved in 1 minute. When I increase time to 3 min Shash solves 9 more, but Crystal only 5. One nice example:
[d]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1

Analysis by ShashChess 35.1:
...
1.Rxb4 Ka8 2.Ra4+ Kb7 3.Rb4 Nd3 4.Rxb3 Nc1 5.Rb4 Ne2 6.Ba5+ Ka8 7.Rh4 Nf2 8.Rxh5 g2 9.Rg5 g1Q+ 10.Rxg1 Nxg1 11.Kxa2 Kb7 12.Kb1 Ne4 13.Kc2 Nf3 14.Kd3 Nxe5+ 15.Ke3 Nc4+ 16.Kf4 Nxa5 17.Kxf5 Nb3 18.Kg6 Kb6 19.Kf5 Nf2 20.Ke5 d5
-+ (-4.50) Depth: 28/53 00:00:44 164mN, tb=171000
1.Rxb4 Ka8 2.Ra4+ Kb7 3.Rb4 Nd3 4.Rxb3 Nc1 5.Rb4 Ne2 6.Ba5+ Ka8 7.Rh4 Nf2 8.Rxh5 g2 9.Rg5 g1Q+ 10.Rxg1 Nxg1 11.Bb6 Ngh3 12.Kxa2 Kb7 13.Bd4 Kc6 14.Kb3 Kd5 15.Bb6 Nd3 16.Kc3 Nxe5 17.b4 Ng5 18.Be3 Ne4+ 19.Kb3 Nd6 20.Ka4 Ndc4 21.Bc5 f4 22.Bg1
-+ (-4.62) Depth: 29/54 00:01:01 225mN, tb=232579
1.Rxh1
-+ (-3.94 ++) Depth: 30/54 00:01:22 313mN, tb=311748
1.Rxh1 Nxh1 2.e6 dxe6 3.Bc7+ Kb7 4.Bxg3 f4 5.Bh4 e5 6.Bf2 Kc8 7.Be1 e4 8.Bh4 Kc7 9.Bg3 Kb7 10.Be1 Ka6 11.Bh4 Kb5 12.Be1 Kc4 13.Bf2 Kd3 14.Bh4 e3 15.Bg3 f3 16.Be1 Ke2 17.Bg3 Kf1 18.Be1 Kg2 19.Bg3 Kh3 20.Bf2 Kg4
-/+ (-1.50) Depth: 30/54 00:01:22 313mN, tb=311748
Jouni
fishpov
Posts: 103
Joined: Sat Mar 07, 2015 6:05 pm

Re: ShashChess

Post by fishpov »

Yes but with High Tal=True
amchess
Posts: 340
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

ShashChess 35.2
Fix montecarlo
Speedup WinProbability
amchess
Posts: 340
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

Alexander 1.2
Fix montecarlo
Speedup WinProbability
fishpov
Posts: 103
Joined: Sat Mar 07, 2015 6:05 pm

Re: ShashChess

Post by fishpov »

Hi amchess

Thanks for your reply but I don't see these options in Fritz gui.
I have MCTS not cheked by default with
MCTSThreads 1
MCTS Multi Strategy 20
MCTS Multi MinVisits 5
amchess
Posts: 340
Joined: Tue Dec 05, 2017 2:42 pm

Re: ShashChess

Post by amchess »

https://github.com/amchess/ShashChess/wiki/Matches
You can see why you get banned from the Stockfish discord channel...