Moderator: Ras
Code: Select all
Houdini 5 - SF 8, Blitz 3m+2s
1 Stockfish 8 64 POPCNT +20/=73/-7 56.50% 56.5/100
2 Houdini 5 x64-popc +7/=73/-20 43.50% 43.5/100Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
I think that we simply do not know about longer time control.APassionForCriminalJustic wrote:Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
Instructive black H5 win just to illustrate the strong side of Houdini 5Uri Blass wrote:I think that we simply do not know about longer time control.APassionForCriminalJustic wrote:Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
The only way to know is testing.
The gap between stockfish and houdini at short time control is not so big to be sure that stockfish is better at long time control.
Code: Select all
Analysis by Stockfish 8 64 POPCNT: 22...Qe8 23.Nb5 Rc8 24.Qe2 Qg6 25.Bc2
Rf7 26.Nd4 Ra8 27.Ra2 Rc8 28.Ra1 = (0.00) Depth: 35/44 00:00:21
141mN 22...Qe8 23.Nb5 Rc8 24.Qe2 Qg6 25.Bc2 Rf7 26.Nd4 Ra8 27.Ra2 Rc8 28.Ra1
= (0.00) Depth: 36/44 00:00:26 180mN 22...Qe8 23.Nb5 Rc8 24.Qe2 Qg6 25.
Bc2 Rf7 26.Nd4 Ra8 27.Ra2 Rc8 28.Ra1 = (0.00) Depth: 37/44 00:00:33
231mN 22...Rf7 = (-0.07 !) Depth: 38/44 00:01:07 469mN 22...Rf7
= (-0.14 !) Depth: 38/44 00:01:36 669mN 22...Rf7 = (-0.25 !)
Depth: 38/44 00:02:15 947mN 22...Kh8 =/+ (-0.40 !) Depth: 38/57
00:17:24 7272mN 22...Kh8 23.Bh5 Rg8 24.Kh2 Rff8 25.Bd1 Rg7 26.Ra2 Rfg8 27.
Qb1 Qe8 28.Kh1 Qg6 29.Kg1 gxf4 30.exf4 Bxa4 31.Bf3 Bd7 32.Kh1 Bc8 33.Re2 Re7
34.Qc2 Ba6 35.Bh4 Ree8 36.Nb5 Bxb5 37.cxb5 Ng3+ 38.Bxg3 Rxe2 39.Qxe2 Qxg3 40.
Qe3 Qh4 41.Ra1 Qf6 42.Qd4 Qxd4 43.cxd4 =/+ (-0.47) Depth: 38/57 00:26:
07 10885mN (Bram Mourik, AMD X6 1090T 3200 Mhz (Fritzmark 23,1 6x) 10.11.
2016)} Kh8 {[%eval -61,18] [%emt 0:00:04] (Qg7)} 23. Bf3 {[%eval 1,31] [%emt 0:
00:12] (Qe2) 2215: Stockfish 8 64 POPCNT - Houdini 5 x64-popc, Houdini 5 -
SF 8, Blitz 3m+2s r5k1/2pbq3/1p1p1r1p/p1nP1pp1/P1PNnP2/2P1P2P/2Q3P1/R2BBRK1 b
- - 0 1 Analysis by Stockfish 8 64 POPCNT: 22...Qe8 23.Nb5 Rc8 24.Qe2
Qg6 25.Bc2 Rf7 26.Nd4 Ra8 27.Ra2 Rc8 28.Ra1 = (0.00) Depth: 35/44 00:
00:21 141mN 22...Qe8 23.Nb5 Rc8 24.Qe2 Qg6 25.Bc2 Rf7 26.Nd4 Ra8 27.Ra2 Rc8
28.Ra1 = (0.00) Depth: 36/44 00:00:26 180mN 22...Qe8 23.Nb5 Rc8 24.
Qe2 Qg6 25.Bc2 Rf7 26.Nd4 Ra8 27.Ra2 Rc8 28.Ra1 = (0.00) Depth: 37/44
00:00:33 231mN 22...Rf7 = (-0.07 !) Depth: 38/44 00:01:07 469mN
22...Rf7 = (-0.14 !) Depth: 38/44 00:01:36 669mN 22...Rf7 =
(-0.25 !) Depth: 38/44 00:02:15 947mN 22...Kh8 =/+ (-0.40 !)
Depth: 38/57 00:17:24 7272mN20-7-73 is quite a deficit (45 Elo - I still think Houdini is over performing). Of course we need the superfinal to show us either way with respect to VLTC. I go back to Stage Three where Stockfish beat Houdini development twice. Both have gained significant Elo since then. Nonetheless, it still stands. Also Keck's tournament showed 15-1 versus Komodo 10.2. Stockfish gains with more time - and more cores. Stockfish only lost 55 games on IPON compared to 114 games for Houdini. Houdini will be hard-pressed to get even a few wins at VLTC versus an engine that hardly losses.Uri Blass wrote:I think that we simply do not know about longer time control.APassionForCriminalJustic wrote:Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
The only way to know is testing.
The gap between stockfish and houdini at short time control is not so big to be sure that stockfish is better at long time control.
With 100 games, that deficit could be pure noise.APassionForCriminalJustic wrote:20-7-73 is quite a deficit (45 Elo - I still think Houdini is over performing). Of course we need the superfinal to show us either way with respect to VLTC. I go back to Stage Three where Stockfish beat Houdini development twice. Both have gained significant Elo since then. Nonetheless, it still stands. Also Keck's tournament showed 15-1 versus Komodo 10.2. Stockfish gains with more time - and more cores. Stockfish only lost 55 games on IPON compared to 114 games for Houdini. Houdini will be hard-pressed to get even a few wins at VLTC versus an engine that hardly losses.Uri Blass wrote:I think that we simply do not know about longer time control.APassionForCriminalJustic wrote:Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
The only way to know is testing.
The gap between stockfish and houdini at short time control is not so big to be sure that stockfish is better at long time control.
Houdini will gain absolutely nothing over Stockfish with huge time. There would be nothing to suggest that it would. The games will speak for themselves.
You were already betting about Houdini being less good than Komodo. So plz, stop your arrogance.APassionForCriminalJustic wrote:Won't be better than Stockfish. You can bet on that.JJJ wrote:Yes, that's indicate Houdini between stockfish 8 and Komodo 10.2
We don't know yet if Houdini will be better at longer time control.
He got you there, Adam !JJJ wrote:You were already betting about Houdini being less good than Komodo
