Mythbuster continued

[Jouni]

Now all matches are 30 games or more and scores:

Komodo 54,6%
Houdini 54,0%
Critter 49,6%
Rybka 46,7%
Stockfish 45,9%

So maybe Don is right - Komodo scales well! But Stockfish not even if that is also claimed..

[beram]

Hi Bram,

Komodo strikes back! In a very interesting game with unbalanced material it managed to win with black against Houdini.

In the beginning, Houdini made some strange moves (loss of tempo like Rh3 and shortly after Rh1 again). In this position, Komodo decides to give his queen and a pawn for two knights and a bishop:

Code: Select all

1r1qrnk1/3b2b1/3pp1p1/1pp1n1Bp/1P1NP2P/2N2P2/P2Q2P1/3R1BKR b - - 0 21

Black plays 21. cxd4. That was the right decision because white's queen doesn't manage to make an impact in this closed position. However blacks pieces do and therefore black is able to strengthen his position.

[/size]

Hi Timo,

After 38 games, and two nice last wins with French opening, the LTC results between Komodo4 and Houdini2.0C in your Clash of the Titans are pretty the same as in the 100 games match between both, on Blitz level (5min 5 sec, of ATOMICC chess by Brent M http://atomicc-testing.blogspot.com/201 ... sults.html)
So no further closing of the ELO gap (just 15-20 ELO) between them at very LTC, as has been suggested by the author of Komodo, can be observed. The difference stabilizes.
Same discussions we have had here in the past, where the Rybka fans stubbornly repeated for long time the same claims for Rybka 4 being much better than Houdini at LTC. After Houdini 1.5a winning the TCEC matches in a clear way and the rise of Komodo and the rumour over Rybka, these unproven claims finally became less heard.
The same counts for the claim for Stockfish being a program that also would scale better, can clearly not been found in the current results against all of his opponents. The only true thing that can be said is that Komodo doesn’t scale that good at bullet/blitz. Having said that, nevertheless, I would like to see and buy Komodo 5 MP because Komodo is a very good allround positionally playing program.
Lets see if Komodo can strike back again

grts Bram

[TimoK]

Hi Timo,

After 38 games, and two nice last wins with French opening, the LTC results between Komodo4 and Houdini2.0C in your Clash of the Titans are pretty the same as in the 100 games match between both, on Blitz level (5min 5 sec, of ATOMICC chess by Brent M http://atomicc-testing.blogspot.com/201 ... sults.html)
So no further closing of the ELO gap (just 15-20 ELO) between them at very LTC, as has been suggested by the author of Komodo, can be observed. The difference stabilizes.
Same discussions we have had here in the past, where the Rybka fans stubbornly repeated for long time the same claims for Rybka 4 being much better than Houdini at LTC. After Houdini 1.5a winning the TCEC matches in a clear way and the rise of Komodo and the rumour over Rybka, these unproven claims finally became less heard.
The same counts for the claim for Stockfish being a program that also would scale better, can clearly not been found in the current results against all of his opponents. The only true thing that can be said is that Komodo doesn’t scale that good at bullet/blitz. Having said that, nevertheless, I would like to see and buy Komodo 5 MP because Komodo is a very good allround positionally playing program.
Lets see if Komodo can strike back again
grts Bram

Hi Bram,

I agree with all you said. Houdini played two very nice games against Komodo in this French opening variation (C15) and managed to win this position with white and black:

[D]rn1q1rk1/pbp2ppp/1p2pb2/8/3P4/P3B3/1PPQNPPP/2KR1B1R w - - 0 11

Especially the game with Houdini as black was really interesting. Houdini played a few extremly good moves:

[D]1r3rk1/6pp/2b1pb2/p1p4q/PpP4P/1P1BBP2/K1Q3P1/3R3R b - - 0 24
Houdini played 24. ..Bxa4!!

After Komodos reaction (Bf4), Houdini played another great move: 25. ..Bxb3+! Houdini already evaluates the position as better for black knowing that the black pawns on the queenside are very strong.

[Event "Houdini-Komodo"]
[Site "?"]
[Date "2012.03.14"]
[Round "38.1"]
[White "Komodo64 SSE Version 4"]
[Black "Houdini 2.0c Pro x64 [t=1]"]
[Result "0-1"]
[ECO "C15"]
[Annotator "0.15;0.06"]
[PlyCount "126"]
[EventDate "2012.03.05"]

{AMD Phenom(tm) II X4 905e Processor 2999 MHz W=26.3 plies; 1.025kN/s B=25.5
plies; 1.991kN/s; 3.509 TBAs} 1. e4 e6 2. d4 d5 3. Nc3 Bb4 4. Nge2 dxe4 5. a3
Be7 6. Nxe4 Nf6 7. Nxf6+ Bxf6 8. Be3 O-O 9. Qd2 b6 10. O-O-O Bb7 11. Nf4 {0.15/
21 59} Nd7 {0.06/22 194 (Qe7)} 12. d5 {0.17/23 196} Nc5 {0.00/23 0} 13. dxe6 {
0.30/24 168 (Bc4)} Nxe6 {0.16/23 200} 14. Nxe6 {0.27/25 74} fxe6 {0.12/25 191
(Qxd2+)} 15. f3 {0.30/24 113} Qd5 {0.11/24 61 (Qxd2+)} 16. Kb1 {0.28/24 156} c5
{0.13/24 103 (Rad8)} 17. Qc1 {0.19/24 226 (Qe1)} Qh5 {0.02/23 253 (Qe5)} 18. h4
{0.35/23 138} Bd5 {0.06/24 56 (Rad8)} 19. c4 {0.15/24 187 (Bg5)} Bc6 {0.11/24
284} 20. Qc2 {0.06/26 849 (Bd3)} Rab8 {-0.07/23 204} 21. Ka2 {0.15/23 0 (a4)}
b5 {-0.18/22 213 (e5)} 22. b3 {0.00/22 158 (Bf4)} b4 {-0.07/24 1179} 23. Bd3 {
-0.05/25 0} a5 {-0.02/23 192 (Rb6)} 24. a4 {0.36/23 131} Bxa4 {0.05/23 132
(Kh8)} 25. Bf4 {0.14/23 193 (Bxh7+)} Bxb3+ {0.09/21 158} 26. Qxb3 {0.25/26 929}
a4 {-0.48/23 50} 27. Qb1 {0.30/22 8} Rb6 {-0.11/24 1000 (b3+)} 28. g4 {0.16/21
102 (Be4)} Qf7 {-0.98/20 127} 29. Bxh7+ {-0.79/21 398} Kh8 {-1.33/23 0} 30. Qg6
{-0.78/23 116} a3 {-1.37/22 0} 31. Kb3 {-0.93/26 234} a2 {-1.74/23 0 (Bd4)} 32.
Qxf7 {-1.07/23 157} Rxf7 {-1.83/25 174} 33. Be4 {-1.60/25 172 (Bg6)} Ra7 {-1.
74/26 116} 34. Be5 {-1.78/27 22} Ra3+ {-1.74/25 3} 35. Kc2 {-1.82/27 34} b3+ {
-1.86/27 252} 36. Kb2 {-2.00/31 34} a1=Q+ {-2.11/27 199} 37. Rxa1 {-2.02/32 56}
Rxa1 {-1.99/27 0} 38. Kxa1 {-1.78/31 359 (Rxa1)} Bxe5+ {-1.66/27 85} 39. Kb1 {
-1.84/32 91} Ra6 {-2.07/29 288} 40. Kc1 {-2.15/30 155} Ra1+ {-2.01/29 32} 41.
Bb1 {-2.05/28 139} Ra2 {-2.01/31 164} 42. Kd1 {-1.92/29 24} Rg2 {-2.02/31 274}
43. Bg6 {-2.01/31 174 (Ke1)} b2 {-2.09/28 135} 44. Ke1 {-2.25/28 7} Bg3+ {-2.
18/29 217} 45. Kf1 {-2.13/29 24} Rf2+ {-2.27/30 123} 46. Kg1 {-2.34/30 120} e5
{-2.27/31 230 (Rd2)} 47. h5 {-2.26/28 265 (Bb1)} Rd2 {-2.45/29 248} 48. Kf1 {
-2.25/29 235} Rd1+ {-2.50/28 0 (Kg8)} 49. Kg2 {-2.28/25 19} Rc1 {-2.58/28 222}
50. Rf1 {-2.55/28 85} Kg8 {-2.60/28 54} 51. g5 {-2.60/27 80 (Be4)} Bf4 {-3.10/
25 104} 52. Rf2 {-3.16/28 312 (Be4)} b1=R {-3.82/26 137 (b1Q)} 53. Bxb1 {-3.16/
29 144} Rxb1 {-4.01/27 51} 54. g6 {-3.14/27 5} Kf8 {-3.95/27 85} 55. Ra2 {-3.
21/27 21} Rc1 {-4.09/27 140} 56. Kh3 {-3.18/26 87 (Ra7)} Rxc4 {-4.34/25 169}
57. Kg4 {-3.58/28 335 (Ra7)} Ke7 {-4.47/25 104} 58. Kf5 {-4.06/27 142 (Ra8)}
Kd6 {-4.69/23 21 (Rd4)} 59. Ra6+ {-3.93/23 9} Kd5 {-4.69/23 0} 60. Ra8 {-4.52/
28 87} Rc3 {-4.97/25 0 (Rd4)} 61. Kg4 {-4.55/30 100} Bh6 {-5.30/27 0 (Rd3)} 62.
Rh8 {-5.09/30 119} Rc4+ {-6.10/27 323 (Rb3)} 63. Kg3 {-5.52/28 59} Ke6 {-6.05/
25 0 (Rb4)} 0-1

I think this kind of opening position must be very comfortable for Houdini because it mangaged to win it also with white/black against Stockfish and with white against Critter and Rybka!

Best regards
Timo

[beram]

hi Jouni,
now after 38-39 games
Houdini2.0 59,8% and Komodo4 53,8 %
So Houdini scales not that bad

grts Bram

[Jouni]

Don and Larry have been quiet: either they have not found this thread or have started MP version testing

[Dr.Wael Deeb]

Don and Larry have been quiet: either they have not found this thread or have started MP version testing

Most probably they have serious problems with the MP version of Komodo
Dr.D

[beram]

Don and Larry have been quiet: either they have not found this thread or have started MP version testing

I think they are MP testing

Preliminary results by now:

Clash of the Titans by Timo Klaustermeyer Time controls 125min / 50 moves + (60min+5s) Ponder = ON
60 games Noomen testsuite 2011 30 pos Engines have 6 cores except in matches of Komodo4, 1 core
Fritz Benchmark for 1 thread: ~ 1900 kN/s

Code: Select all

Results after 388 of 480 games
HOUDINI:
Komodo4	Critter1.4	Stockfish2.2.2	Rybka 4.1	            total
25,5	        25,5	       30	            27,5	            108,5 points
48	          47	         48	            49	               192 games
53,1	        54,2	      62,5	          56,1                 56,5 %

KOMODO:
Houdini2.0C	Critter1.4	Stockfish2.2.2	Rybka 4.1	         total
22,5	        25,5	       25,5	             30	        103,5 points
48	          48	         49	               51	        196 games
46,9	        53,1	       52,0	             58,8	       52,8 %

[Houdini]

I think they are MP testing

No, Don is running his 6 sec/12 sec/24 sec/48 sec... test matches to prove scientifically that Komodo 4 overtakes Houdini at long time control.

Robert

[MM]

Don and Larry have been quiet: either they have not found this thread or have started MP version testing

Most probably they have serious problems with the MP version of Komodo
Dr.D

Hi Doc, I don't think so. I think they already have a relatively good MP version but i think they have serious problems to find some improvements that allow them to overtake (for real) Houdini and i don't think they will release anything untill they will reach this goal.

Best Regards

[TimoK]

I think they are MP testing

No, Don is running his 6 sec/12 sec/24 sec/48 sec... test matches to prove scientifically that Komodo 4 overtakes Houdini at long time control.

Robert

Ok, I know my test is statistically insufficiant, but after the games that have been played I'm strongly under the impression that Houdini has no scaling problem at all. So I wonder if Don will find any TC where Komodo 4 scores better than Houdini 2...

But ok, let's wait for Komodo MP, maybe Don and Larry have improved it a bit more so Houdini will have an opponent on an equal footing.

Best regards
Timo