I tested effect of 5 piece syzygy tables (on harddrive). 150 positions from EG_MSb.epd with average 1 sec for move. Result:
1 Stockfish 211013 64 syz +83/=161/-56 54.50% 163.5/300
2 Stockfish 211013 64 def +56/=161/-83 45.50% 136.5/300
+31 ELO - great work Ronald!
Stockfish 211013 syzy vs default
Moderators: hgm, Rebel, chrisw
-
- Posts: 5228
- Joined: Thu Mar 09, 2006 9:40 am
- Full name: Vincent Lejeune
Re: Stockfish 211013 syzy vs default
Great !
I think the effect is even bigger with more time.
Could you test with 4 seconds instead of 1 ?
I think the effect is even bigger with more time.
Could you test with 4 seconds instead of 1 ?
-
- Posts: 5228
- Joined: Thu Mar 09, 2006 9:40 am
- Full name: Vincent Lejeune
Re: Stockfish 211013 syzy vs default
Please, post the games.
-
- Posts: 2041
- Joined: Wed Mar 08, 2006 8:30 pm
Re: Stockfish 211013 syzy vs default
± 27.5 (2SD error bar)Jouni wrote:+31 ELO
-
- Posts: 98
- Joined: Thu Jul 23, 2009 5:40 am
Re: Stockfish 211013 syzy vs default
Now an interesting info would be to know the difference between official Stockfish 4.0 vs. Stockfish 211013
-
- Posts: 3286
- Joined: Wed Mar 08, 2006 8:15 pm
Re: Stockfish 211013 syzy vs default
I repeated test against Houdini 3 with surprise. SF score has been from start 52 % = +16 ELO. But old SF dev version (no bases) got 70-78 % in same test !! Really surprising. Yes I know You need 20k games.
Jouni
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: Stockfish 211013 syzy vs default
Use SPRT with cutechess-cli, it will say how many games you need, for large enough ELO differences, often no more than several hundred.Jouni wrote:I repeated test against Houdini 3 with surprise. SF score has been from start 52 % = +16 ELO. But old SF dev version (no bases) got 70-78 % in same test !! Really surprising. Yes I know You need 20k games.