Mythbuster continued

Discussion of computer chess matches and engine tournaments.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
beram
Posts: 1187
Joined: Wed Jan 06, 2010 2:11 pm

Re: Mythbuster continued ...but Komodo comes back

Post by beram » Sat Mar 10, 2012 4:55 pm

TimoK wrote:Hello Bram,

thank you very much for your commentary on the games! That is highly appreciated because it makes the matches more interesting. For me it's not possible to watch all the games at the same time so it's good to have a second pair of eyes looking for interesting games to comment on.

Indeed Komodo is doing a good job so far especially against Stockfish (as well as Houdini also leads against Stockfish). I think all 5 engines will be very close together in the end if we look at the scores (maybe Houdini will manage to break away from the rest). But they are playing very differently, so it's kind of interesting that we have 5 different engines playing at about the same very high level. After this test I think more games will be necessary so maybe I'll add another 40 games all against each other.

Best regards
Timo

Hi Timo, thank you for playing and sharing. I think that more would appreciate when you continu your interesting Clash of Titans, up till 100 games.
As far as can be judged by observing the games and the results, there is no diminishing strength difference to be seen between Komodo and Houdini, as was the myth created by Don.
Game 18 was an example of good late middlegame/endgame play by Houdini. With black it played very strong as you can replay from move 80 till move 112.
[D] 8/5pk1/8/1p2P1K1/2b2P2/8/1p2r3/1B5R b - - 0 112
At that move Houdini played 112. ..Kf8 (movetime 1 second) allowing the white king to come in. After the winning move 112. ..Rg2+ black would have displaced the white king to the h-file and still having all the odds. The other reply 113. Kf5 – Be4 114. Re1 – Rg3 wins for example 115. Re2 - Bf3! 16. Bxb2 ? Rh3 ! threatening mate, so winning the white rook. As the game went on despite the huge evaluation for black for long time, Houdini ultimately couldn’t prevent exchanging the last pawns, resulting in a drawn RB-R endgame.

grts Bram


[Event "Houdini-Komodo"]
[Site "?"]
[Date "2012.03.10"]
[Round "18.1"]
[White "Komodo64 SSE Version 4"]
[Black "Houdini 2.0c Pro x64 [t=1]"]
[Result "1/2-1/2"]
[ECO "D56"]
[Annotator "0.25;0.07"]
[PlyCount "368"]
[EventDate "2012.03.05"]

{AMD Phenom(tm) II X4 905e Processor 2999 MHz W=25.2 plies; 924kN/s B=26.6
plies; 1.953kN/s; 1.002.814 TBAs} 1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. Nf3 Be7 5.
Bg5 h6 6. Bh4 O-O 7. e3 Ne4 8. Bxe7 Qxe7 9. Rc1 c6 10. Qc2 Nd7 11. Be2 {0.25/
24 186} Nxc3 {0.07/24 189} 12. Qxc3 {0.27/26 56} dxc4 {0.07/25 173} 13. Qxc4 {
0.20/26 54} e5 {0.05/25 121} 14. O-O {0.27/24 34} Re8 {0.05/25 479 (Rd8)} 15.
Qc2 {0.22/25 529} e4 {0.04/26 0} 16. Nd2 {0.22/22 11} Nf6 {0.05/24 224} 17.
Rfe1 {0.22/25 49 (a3)} Bf5 {0.03/22 170 (Bg4)} 18. a3 {0.20/22 75} Rad8 {0.03/
22 90 (Nd5)} 19. b4 {0.27/24 306 (Qa4)} a6 {0.05/23 178 (Nd5)} 20. Qc5 {0.25/
25 462} Bg4 {0.06/24 105} 21. Bf1 {0.26/26 6 (Bxg4)} Qc7 {0.04/24 188 (Be6)}
22. h3 {0.26/24 201 (Qa7)} Be6 {0.03/23 157} 23. Qc2 {0.26/26 153 (Qa7)} Qd6 {
0.03/23 222} 24. Rb1 {0.28/26 423 (Be2)} Bd5 {0.07/23 288} 25. Rec1 {0.27/25
400 (Nc4)} b5 {0.00/24 212 (Nd7)} 26. Be2 {0.30/25 118} Nd7 {0.00/24 51} 27.
Bg4 {0.26/25 395 (Nb3)} Nb6 {-0.07/24 248 (Nf8)} 28. Nb3 {0.25/24 199} Qg6 {-0.
03/24 72} 29. Be2 {0.28/23 39 (Qd1)} Rd6 {-0.01/24 1084 (Rf8)} 30. Nc5 {0.34/
19 21} Qg5 {0.07/23 245 (Qf5)} 31. Kh1 {0.32/23 179 (a4)} Rg6 {-0.05/22 221}
32. Rg1 {0.17/24 158} Rf6 {-0.10/23 168} 33. Rgf1 {0.16/25 34 (Rbf1)} Qh4 {-0.
06/24 150 (Rg6)} 34. Kg1 {0.15/26 184} h5 {-0.09/23 311 (Rg6)} 35. Rbc1 {0.25/
23 164} Rd8 {-0.13/23 285 (Ra8)} 36. Qc3 {0.07/23 265 (Qd1)} Nd7 {-0.21/25 455}
37. Nxd7 {0.01/25 252 (Qe1)} Rxd7 {-0.29/22 49} 38. Qe1 {-0.03/25 144 (Kh2)}
Qg5 {-0.27/23 81} 39. Kh1 {-0.05/24 3 (Kh2)} Rdd6 {-0.35/22 91 (h4)} 40. Rc5 {
-0.03/25 247} Rf5 {-0.29/24 0 (Rg6)} 41. a4 {0.01/25 154} Rg6 {-0.29/23 96} 42.
g3 {-0.05/26 62} Qd8 {-0.27/23 101 (h4)} 43. h4 {0.05/25 187} Qd6 {-0.26/24
104 (Rgf6)} 44. Kh2 {0.05/27 307 (Kg2)} Re6 {-0.26/23 150 (Qf6)} 45. Kg2 {0.05/
26 144} Kh8 {-0.26/24 50 (Rg6)} 46. Qd1 {0.05/25 230} Rh6 {-0.26/24 0} 47. axb5
{0.06/25 113} axb5 {-0.26/25 149} 48. Qe1 {0.05/27 15} Re6 {-0.26/25 152 (Rg6)}
49. Rc2 {0.05/27 139 (Qa1)} Re8 {-0.36/24 159 (Kg8)} 50. Qd2 {0.05/27 215} Ra8
{-0.33/24 0 (Re6)} 51. Qc3 {-0.02/26 266 (Rfc1)} g6 {-0.35/24 137 (Kg8)} 52.
Rb2 {-0.01/26 129 (Rb1)} Kg7 {-0.36/23 78} 53. Bd1 {0.00/25 47 (Rc1)} Ra1 {-0.
34/22 79} 54. Re1 {0.01/25 31 (Rc2)} Qe7 {-0.34/22 101 (Ra6)} 55. Be2 {-0.05/
26 499} Ra7 {-0.33/26 0 (Ra8)} 56. Rc1 {0.00/25 109 (Rbb1)} Ra8 {-0.33/23 90}
57. Qd2 {-0.01/24 14 (Qe1)} Qf6 {-0.34/23 113 (Qd6)} 58. Rf1 {-0.01/24 41} Ra3
{-0.34/23 35} 59. Rc2 {0.00/25 55 (Rbb1)} Qe6 {-0.35/23 76 (Qd6)} 60. Qc1 {-0.
01/25 124 (Rcc1)} Ra4 {-0.36/24 89 (Ra8)} 61. Rb2 {-0.03/24 40 (Qd2)} Rf6 {-0.
36/24 71 (Qd6)} 62. Qc3 {-0.03/25 64} Ra8 {-0.36/24 310 (Qd7)} 63. Rc2 {-0.03/
25 89 (Rbb1)} Qd6 {-0.36/23 92} 64. Rb1 {-0.04/27 33} Rf5 {-0.36/23 47 (Qe7)}
65. Rf1 {-0.05/24 218} Kh7 {-0.36/25 0 (Rf6)} 66. Bd1 {0.02/24 155 (Rb1)} Rf6 {
-0.36/23 88 (Kg8)} 67. Rb2 {-0.05/23 54 (Rd2)} Qe6 {-0.37/23 55} 68. Be2 {-0.
05/27 121 (Kh2)} Qf5 {-0.39/23 53 (Kg7)} 69. Rc2 {-0.05/24 70} Kg7 {-0.39/23 0
(Kg8)} 70. Qc5 {-0.05/25 91 (Rb2)} Ra3 {-0.39/24 52 (Qd7)} 71. Rc3 {-0.05/26 66
} Ra4 {-0.39/25 0 (Ra1)} 72. Rc2 {-0.05/26 66} Qd7 {-0.39/24 33 (Ra8)} 73. Rb1
{-0.05/25 59 (Qc3)} Ra8 {-0.39/23 60 (Qf5)} 74. Qc3 {-0.04/23 18 (Rf1)} Qf5 {
-0.39/23 57} 75. Rf1 {-0.05/26 18} Ra7 {-0.39/24 66 (Ra6)} 76. Rd2 {-0.05/23 9}
Ra4 {-0.39/24 90 (Ra6)} 77. Rb2 {-0.05/24 75 (Bd1)} Kg8 {-0.38/23 60 (Be6)} 78.
Rc2 {-0.05/24 82} Ra8 {-0.38/24 0} 79. Rd2 {-0.05/25 56 (Qc5)} Ra7 {-0.38/23
45 (Ra6)} 80. Qc5 {-0.05/25 56 (Rc2)} Ra6 {-0.38/23 44 (Ra3)} 81. Qe7 {-0.05/
23 18 (Qc3)} Ra3 {-0.39/23 50 (Kg7)} 82. Rb2 {-0.25/22 118 (Qe5)} Kg7 {-0.41/
22 55} 83. Kh2 {-0.30/23 13} Be6 {-0.40/22 36 (g5)} 84. Qc7 {-0.45/22 52} Rc3 {
-0.53/22 67} 85. Kg1 {-0.46/24 34 (Qb8)} g5 {-0.56/23 44} 86. hxg5 {-0.65/24
182} Qxg5 {-0.57/25 0} 87. Qe5 {-0.64/20 3 (Kg2)} Qxe5 {-0.62/25 56} 88. dxe5 {
-0.76/26 18} Rf5 {-0.54/26 45} 89. Kg2 {-0.75/27 63 (Rd1)} Rxe5 {-0.54/25 47}
90. Rd1 {-0.75/25 25} Rg5 {-0.55/25 31 (Bg4)} 91. Kh2 {-0.83/23 42} Rf5 {-0.55/
24 31 (Kg6)} 92. Kg2 {-0.76/20 4} Bd5 {-0.55/24 48 (Re5)} 93. Rd4 {-0.71/23 104
} Kf6 {-0.55/26 0 (Rg5)} 94. Rbd2 {-0.71/22 12} Ke6 {-0.55/26 55 (Be6)} 95. Bd1
{-0.77/24 82} Rg5 {-0.55/26 0} 96. Kh3 {-0.83/22 48} Kf6 {-0.63/24 12 (f6)} 97.
Rc2 {-0.81/22 13 (Kh2)} Rd3 {-0.66/24 42 (Rxc2)} 98. Rxd3 {-0.80/23 31} exd3 {
-0.56/24 0} 99. Rd2 {-0.77/22 10} c5 {-0.59/25 24} 100. f4 {-0.80/24 12 (bxc5)}
Rf5 {-0.74/24 33} 101. Rxd3 {-0.86/25 26} Be6 {-0.78/26 73 (cxb4)} 102. Kh4 {
-1.07/22 17} cxb4 {-0.82/26 205} 103. e4 {-1.07/26 0} Rc5 {-0.82/21 0} 104. g4
{-1.16/22 25} Rc3 {-0.84/25 27} 105. Rd2 {-1.11/24 19 (e5+)} hxg4 {-0.90/24 26}
106. Bxg4 {-1.07/25 19} Bc4 {-0.88/25 17 (Rc4)} 107. e5+ {-0.86/23 30} Kg7 {-0.
92/26 121} 108. Kg5 {-0.93/27 19} Re3 {-0.94/25 105} 109. Bf5 {-0.93/26 32
(Rb2)} b3 {-0.98/24 27} 110. Rd1 {-1.07/24 19 (Rh2)} b2 {-1.52/22 44 (Rg3+)}
111. Bb1 {-1.04/23 27 (Rg1)} Re2 {-1.36/20 0 (Rg3+)} 112. Rh1 {-1.39/21 23} Kf8
{-1.56/23 1 (Rg2+)} ({Houdini 1.5 w32: 1)} 112... Rg2+ 113. Kh4 (113. Kf5 Bd5
114. Re1 (114. Rf1 Bb3 (114... b4 115. Rh1 b3) 115. Ke4 Rc2 116. Rf3 Ba4 117.
Kf5 Rc1 118. Be4) 114... Rg3 115. Re2 (115. e6 Bxe6+ 116. Ke5 Rg2 117. Be4 Rd2
118. f5 Bc4 119. Rg1+ Kf8 120. f6 Ke8 121. Bc6+) 115... Bf3 116. Rxb2 (116. Re1
Bh5 117. e6) 116... Rh3) 113... Kf8 (113... b4) 114. Kh3 Re2 115. Bf5 Ke7 116.
Kg4 Rg2+ 117. Kh3 Rd2 118. Kg3 Bd5 119. Re1 f6 120. Bb1 Rg2+ 121. Kh3 b4 122.
Rd1 Ba8 123. Re1 b3 124. Bf5 Rc2 125. Rb1 Bb7 126. e6 Rg2 127. Kh4 Bd5 128. Kh3
Re2 129. Kg4 Rf2 130. Re1 Rg2+ 131. Kh3 Rc2 132. Rb1 Re2 133. Kg4 {-2.51/30}) (
{Houdini 1.5 w32: 2)} 112... Kf8 113. Kf6 Ke8 114. Rd1 Rf2 115. e6 Bxe6 116. f5
b4 117. Bd3 Bd7 118. Be4 Rf4 119. Bd3 Rd4 120. Re1+ Kd8 121. Bb1 Rf4 122. Bd3
Rf2 123. Rb1 Be6 124. Ke5 Ba2 125. Re1 Ke7 126. Bb1 Bxb1 127. Rxb1 Re2+ 128.
Kd4 Kf6 129. Kd3 Rh2 130. Ke4 Rc2 131. Kd3 Rf2 132. Ke4 Rg2 133. Kf3 Rc2 134.
Ke4 Rf2 135. Ke3 Rg2 136. Ke4 Rf2 137. Ke3 {-1.97/30}) ({Houdini 1.5 w32: 3)}
112... Rf2 113. f5 Kf8 114. Kf6 Ke8 115. Be4 Bb3 116. Re1 Kd7 117. Bd3 Bc4 118.
Be4 Rd2 119. Rb1 Ba2 120. Re1 Rf2 121. Rd1+ Kc7 122. Bb1 Bb3 123. Re1 Kd8 124.
Be4 Rd2 125. Rf1 Kd7 126. Rh1 Rf2 127. Rb1 Kc7 128. Re1 Kd8 129. Bb1 Kd7 130.
Be4 Rd2 131. Bb1 Bc4 132. Kg5 Bb3 133. Kf6 Bc4 {-1.45/30}) 113. Kf6 {-1.57/22
22} Ke8 {-1.71/24 0} 114. Rd1 {-1.69/23 37} Rf2 {-1.89/26 49} 115. f5 {-1.79/
25 18} Re2 {-1.97/27 111} 116. Rh1 {-1.90/28 3 (e6)} Kd7 {-2.05/25 18} 117. Rg1
{-1.90/26 2} Kc6 {-2.38/26 60} 118. Kg7 {-1.92/28 52 (Rd1)} Rxe5 {-2.98/23 19}
119. Rg2 {-1.92/26 4} b4 {-2.99/23 0} 120. Rxb2 {-1.93/25 14} b3 {-2.98/24 0}
121. Bd3 {-1.92/22 3 (f6)} Bxd3 {-3.01/22 2} 122. Rxb3 {-1.93/24 5} Bc4 {-2.98/
29 15} 123. Rf3 {-1.94/26 3 (Rc3)} Kc5 {-2.98/27 16 (Re1)} 124. Kf6 {-1.95/22 2
} Rd5 {-3.06/28 28} 125. Rg3 {-1.95/27 8 (Rh3)} Rd6+ {-3.06/30 25} 126. Kg7 {
-1.94/27 3} Rc6 {-3.06/30 14} 127. Re3 {-1.93/29 16} Kd4 {-3.06/32 4} 128. Re7
{-1.93/28 9} f6 {-3.06/34 11 (Bd5)} (128... Bd5 129. Re8 Rb6 130. Re2 Rb3 131.
Kf6 Rb7 132. Kg7) 129. Kh6 {-1.90/30 26 (Kg6)} Bd3 {-3.06/34 15 (Ra6)} 130. Kg6
{-1.89/28 9} Rb6 {-3.06/35 7 (Be4)} 131. Re8 {-1.85/28 15} Be4 {-3.06/36 5
(Ra6)} 132. Re7 {-1.86/27 9} Rd6 {-3.06/36 7} 133. Re8 {-1.86/27 2} Rc6 {-3.06/
36 13 (Bd3)} 134. Re7 {-1.86/27 11} Rb6 {-3.06/36 6 (Ra6)} 135. Re8 {-1.86/28
15} Bb1 {-3.06/36 8 (Rd6)} 136. Re7 {-1.89/29 18} Rc6 {-3.06/37 0} 137. Rb7 {
-1.86/29 15} Bc2 {-3.06/36 0} 138. Re7 {-1.94/29 19} Rb6 {-3.06/37 7} 139. Rc7
{-1.86/30 11 (Re8)} Bd3 {-3.06/36 19} 140. Re7 {-1.85/29 3} Rd6 {-3.06/36 12}
141. Re1 {-1.86/28 3 (Re8)} Ra6 {-3.06/35 14 (Kd5)} 142. Re7 {-1.83/27 12} Be4
{-3.06/36 7} 143. Re8 {-1.86/27 3} Rd6 {-3.06/36 13} 144. Re7 {-1.86/28 2} Rc6
{-3.06/36 12 (Kd5)} 145. Re8 {-1.85/28 16} Ra6 {-3.06/36 0 (Kd5)} 146. Re7 {-1.
84/29 24} Kd5 {-3.06/37 0} 147. Re8 {-1.86/30 13} Rc6 {-3.06/38 13 (Rd6)} 148.
Re7 {-1.85/29 15} Rb6 {-3.06/38 0 (Bd3)} 149. Re8 {-1.84/28 9} Rd6 {-3.06/37
15 (Ra6)} 150. Re7 {-1.83/27 3} Rc6 {-3.06/37 22} 151. Re8 {-1.83/29 6} Ra6 {
-3.06/37 15 (Bd3)} 152. Re7 {-1.84/29 17} Bd3 {-3.06/36 0 (Rd6)} 153. Re8 {-1.
86/29 15} Rc6 {-3.06/36 0 (Rb6)} 154. Re7 {-1.85/26 8} Rd6 {-3.06/36 18 (Rb6)}
155. Re8 {-1.86/26 3 (Re1)} Ra6 {-2.37/28 13 (Rb6)} 156. Re7 {-1.86/28 15} Rb6
{-2.37/31 0} 157. Re8 {-1.85/27 9} Be4 {-2.25/29 6} 158. Re7 {-1.83/29 9} Bb1 {
-2.25/32 10 (Bc2)} 159. Re8 {-1.81/28 11} Rc6 {-1.68/29 4 (Ra6)} 160. Re7 {-1.
85/29 16} Ra6 {-1.68/32 0 (Bd3)} 161. Rb7 {-1.86/28 11} Bc2 {-1.68/31 7} 162.
Re7 {-1.85/28 11 (Rc7)} Rb6 {-1.68/32 14 (Rc6)} 163. Re8 {-1.82/27 10 (Rc7)}
Ba4 {-1.43/28 13 (Rc6)} 164. Re7 {-1.86/27 12 (Re3)} Kd4 {-1.34/30 66 (Bb5)}
165. Re2 {-1.83/24 3 (Kh5)} Bb5 {-1.09/28 56} 166. Re1 {-1.18/27 10 (Re7)} Ra6
{-1.16/27 57 (Kd5)} 167. Re7 {-0.83/23 7} Rd6 {-0.55/24 40} 168. Re1 {-0.78/25
10} Rb6 {-0.33/23 1 (Bd3)} 169. Re7 {-0.80/23 8} Rc6 {-0.27/23 6 (Ra6)} 170.
Rb7 {-0.78/23 12 (Rf7)} Bc4 {-0.27/26 10} 171. Rb8 {-0.81/21 1 (Re7)} Ke4 {-0.
27/23 9 (Ke5)} 172. Re8+ {-0.75/21 13} Kf4 {-0.27/25 0} 173. Rd8 {-0.75/20 2
(Ra8)} Rc5 {-0.25/23 10 (Be2)} 174. Rd4+ {-0.75/24 12 (Rd2)} Ke5 {-0.25/27 8}
175. Rd1 {-0.75/26 9} Bb5 {-0.25/24 5} 176. Rd8 {-0.76/24 6 (Re1+)} Rc4 {-0.25/
24 8 (Rc3)} 177. Rd1 {-0.76/22 3} Be8+ {-0.25/23 6} 178. Kg7 {-0.76/25 4} Kxf5
{-0.25/25 6} 179. Rf1+ {-0.76/24 1} Ke5 {0.00/27 11} 180. Re1+ {-0.77/26 5} Kf4
{0.00/39 5} 181. Rf1+ {-0.77/24 3 (Kxf6)} Kg3 {0.00/37 9 (Ke3)} 182. Kxf6 {-0.
75/21 5} Re4 {0.00/0 0 (Bc6)} 183. Rg1+ {0.00/0 0} Kf2 {0.00/0 0} 184. Rg5 {0.
00/0 0} Bd7 {0.00/0 0} 1/2-1/2

Jouni
Posts: 2198
Joined: Wed Mar 08, 2006 7:15 pm

Re: Mythbuster continued ...but Komodo comes back

Post by Jouni » Sat Mar 10, 2012 5:35 pm

I only just noticed, this is all against all tournament :) So current score from best to worst:

Houdini 40/72 55,6%
Komodo 40,5/76 53,3%
Critter 39,5/81 48,8%
Stockfish 41/87 47,1%
Rybka 37/80 46,3%

Note, that in Critter - Stockfish there has been 21 draws in 26 games so far.
Jouni

TimoK
Posts: 97
Joined: Sun Jan 03, 2010 11:28 am
Location: Hamburg

Re: Mythbuster continued ...but Komodo comes back

Post by TimoK » Sun Mar 11, 2012 12:02 am

beram wrote: As far as can be judged by observing the games and the results, there is no diminishing strength difference to be seen between Komodo and Houdini, as was the myth created by Don.
Game 18 was an example of good late middlegame/endgame play by Houdini. With black it played very strong as you can replay from move 80 till move 112.
Hi Bram,

indeed it seems like there is no worse scaling of Houdini with long TCs. Ok, Don and Larry were talking about version 1.5a, maybe they were right concerning that version, but version 2.0c shows the same excellent performance like in shorter TCs.

The game you mentioned was very thrilling in the late middlegame / endgame, but in the end the time was too short for Houdini to find the best moves. With an even longer TC like 125/50+125/50+... (repeating) maybe it would score even better.

Best regards
Timo

TimoK
Posts: 97
Joined: Sun Jan 03, 2010 11:28 am
Location: Hamburg

Re: Mythbuster continued ...but Komodo comes back

Post by TimoK » Sun Mar 11, 2012 12:30 am

Jouni wrote:I only just noticed, this is all against all tournament :) So current score from best to worst:
Houdini 40/72 55,6%
Komodo 40,5/76 53,3%
Critter 39,5/81 48,8%
Stockfish 41/87 47,1%
Rybka 37/80 46,3%
Hi Jouni,

yes, all against all was the idea to determine a winner the fairest way.

Here are the IPON results of the Top 5 against each other:

Houdini - Komodo: 75.0-75.0
Houdini - Critter: 79.0-71.0
Houdini - Stockfish: 85.5-64.5
Houdini - Rybka: 87.5-62.5
Komodo - Critter: 72.5-77.5
Komodo - Stockfish: 74.5-75.5
Komodo - Rybka: 81.5-68.5
Critter - Stockfish: 79.5-70.5
Critter - Rybka: 76.5-73.5
Stockfish - Rybka: 78.5-71.5

Overall Scores:
Houdini: 327.0/600 (54.5%)
Critter: 304.5/600 (50.75%)
Komodo: 303.5/600 (50.58%)
Stockfish: 289.0/600 (48.17%)
Rybka: 276.0/600 (46.0%)

Best regards
Timo

pichy
Posts: 2564
Joined: Thu Mar 09, 2006 2:04 am

Re: Mythbuster continued ...but Komodo comes back

Post by pichy » Sun Mar 11, 2012 11:34 am

TimoK wrote:
beram wrote: As far as can be judged by observing the games and the results, there is no diminishing strength difference to be seen between Komodo and Houdini, as was the myth created by Don.
Game 18 was an example of good late middlegame/endgame play by Houdini. With black it played very strong as you can replay from move 80 till move 112.
Hi Bram,

indeed it seems like there is no worse scaling of Houdini with long TCs. Ok, Don and Larry were talking about version 1.5a, maybe they were right concerning that version, but version 2.0c shows the same excellent performance like in shorter TCs.

The game you mentioned was very thrilling in the late middlegame / endgame, but in the end the time was too short for Houdini to find the best moves. With an even longer TC like 125/50+125/50+... (repeating) maybe it would score even better.

Best regards
Timo

It is true, but with a longer time control of 125/50+125/50+... (repeating) Komodo might benefit more than the rest of the engines.

beram
Posts: 1187
Joined: Wed Jan 06, 2010 2:11 pm

Re: Mythbuster continued ...but Komodo comes back

Post by beram » Sun Mar 11, 2012 7:15 pm

TimoK wrote:
beram wrote: As far as can be judged by observing the games and the results, there is no diminishing strength difference to be seen between Komodo and Houdini, as was the myth created by Don.
Game 18 was an example of good late middlegame/endgame play by Houdini. With black it played very strong as you can replay from move 80 till move 112.


Hi Bram,

indeed it seems like there is no worse scaling of Houdini with long TCs. Ok, Don and Larry were talking about version 1.5a, maybe they were right concerning that version, but version 2.0c shows the same excellent performance like in shorter TCs.

The game you mentioned was very thrilling in the late middlegame / endgame, but in the end the time was too short for Houdini to find the best moves. With an even longer TC like 125/50+125/50+... (repeating) maybe it would score even better.

Best regards
Timo

Hi Timo, and it goes on with another tactic finale by Houdini in game 23
[D] 6k1/1p1rqp2/r1p1p1pb/P3P2p/1pNPQ2P/1P3RP1/5PK1/3R4 w - - 0 45
After some squeezing manoeuvring in middlegame, the position asks for an attack on the black king by opening up on the kingside with g4.
Houdini played 45. Kh3 ! (giving already +0.73) Komodo4 had expected g4 (which is not that good, for example: 45.g4 Qxh4 46.Rg3 Bg7 47.Rh1 Qd8 48.gxh5 Rxd4 49.Qf3 Qd5 50.hxg6 f5)
Giving the position to Komodo 3 it just wants to play Ra1 for very long time till depth 28. At depth 29 after 16 minutes it sees the light:
Analysis by Komodo64 3:
45.Ra1 Bg7 46.Ra4 Qd8 47.Qf4 Bf8 48.Nb6 Rc7 49.Ra1 Bg7 50.Qe4 Re7 51.Ra4 Rc7 52.Ra2 Bf8 53.Qf4 Ra7 54.Kg1 Bg7 55.Ra4 Bf8 56.Qd2 Bc5 57.Rd3 Bxb6 58.axb6 Rxa4 59.bxa4 Rc8 60.Qxb4 c5
+/= (0.41) Depth: 29 00:07:56 634mN
45.Kh3
+/= (0.58) Depth: 29 00:16:01 1261mN
45.Kh3
+/- (1.05) Depth: 29 01:04:14 4918mN

All other engines; Stockfish2.2.2, Rybka4.1, Critter1.4, I ve tried, find Kh3 within reasonable time 10-60 sec. With reversed colors in game 24, it became a draw.
So now 13,5 – 11,5 for Houdini = 56,25 %

Grts Bram


(23) Houdini 2.0c Pro x64 [t=1] - Komodo64 SSE Version 4 [E05]
Houdini-Komodo (23.1), 11.03.2012
[0.12;0.03]
AMD Phenom(tm) II X4 905e Processor 3000 MHz W=23.2 plies; 1.863kN/s B=23.9 plies; 863kN/s 1.d4 Nf6 2.c4 e6 3.g3 d5 4.Nf3 Be7 5.Bg2 0–0 6.0–0 dxc4 7.Qc2 a6 8.a4 Bd7 9.Qxc4 Bc6 10.Bf4 a5 11.Nc3 Na6 12.Ne5 Bxg2 13.Kxg2 Nd5 0.03/24 3:00 14.Rad1 (Nxd5) 0.12/22 3:58 14...c6 (Nab4) 0.01/23 2:51 15.Bc1 (Qb3) 0.12/23 2:33 15...Nb6 0.05/25 4:14 16.Qb3 0.11/23 2:12 Nb4 0.17/24 2:17 17.e4 0.16/23 49 Qc7 (Bd6) 0.13/23 2:49 18.Be3 (Bf4) 0.18/23 10:47 18...Rfe8 (Rfd8) 0.17/22 3:43 19.Rc1 (Nd3) 0.25/22 3:01 19...h6 0.17/21 4:50 20.Rfd1 (Nb5) 0.24/24 0 20...Ra6 (Bf8) 0.26/23 4:58 21.Bf4 0.31/23 2:50 Qd8 0.37/23 36 22.Na2 0.27/22 5:39 Ra8 (Bf8) 0.37/24 1:36 23.Nxb4 (Be3) 0.31/22 3:48 23...axb4 0.21/24 2:07 24.h3 (Ra1) 0.28/23 2:20 24...Bf8 (Rf8) 0.19/23 4:18 25.Ra1 0.26/23 2:31 Nd7 0.46/25 4:47 26.Nc4 0.27/22 27 Nb6 0.46/25 4:25 27.Nd2 0.29/25 0 h5 (Ra6) 0.41/24 5:04 28.h4 (a5) 0.44/24 7:22 28...Nd7 (Ra5) 0.47/23 3:11 29.Qe3 0.50/23 2:48 Nf6 0.59/24 50 30.Nc4 (f3) 0.48/25 2:10 30...Nd5 0.34/23 48 31.Qf3 0.48/26 4:59 Nxf4+ 0.34/28 4:31 32.Qxf4 0.49/26 0 g6 (Qf6) 0.35/26 3:00 33.a5 (b3) 0.50/24 4:05 33...Ra6 (Bg7) 0.23/23 50 34.Rd3 (Qf3) 0.51/24 3:23 34...Qe7 (Re7) 0.43/24 5:58 35.Rad1 0.51/25 3:21 Rd8 0.44/24 56 36.e5 (b3) 0.51/25 7:20 36...Bg7 0.35/24 4:04 37.Rb3 (Rf3) 0.51/25 2:25 37...Rd5 0.18/22 33 38.Qe4 (Rf3) 0.51/25 6:51 38...Bf8 (Bh6) 0.27/25 3:48 39.Rf3 0.51/24 2:28 Bh6 (Bg7) 0.26/24 44 40.Qe1 (Nb6) 0.51/25 3:05 40...Bg7 (Rd8) 0.24/25 2:10 41.b3 (Qe2) 0.51/25 3:05 41...Rd7 (Rd8) 0.30/25 3:19 42.Qd2 (Qe3) 0.51/26 2:53 42...Qf8 (Rd5) 0.31/26 2:51 43.Qf4 (Qe2) 0.51/26 3:00 43...Bh6 0.54/25 8:26 44.Qe4 0.51/29 0 Qe7 (Qd8) 0.48/26 4:01 45.Kh3 (g4) 0.73/24 6:08 [45.g4 Qxh4 46.Rg3 Bg7 47.Rh1 Qd8 48.gxh5 Rxd4 49.Qf3 Qd5 50.hxg6 f5 51.exf6 Bxf6 52.Rh5 Qxf3+ 53.Rxf3] 45...Bg7 (Qf8) 0.81/24 11:19 46.g4 1.55/21 4:23 hxg4+ 1.11/23 1:24 47.Qxg4 1.70/21 3:16 c5 (Qd8) 1.00/25 2:37 48.h5 2.08/21 4:08 cxd4 (Rxd4) 1.15/25 18 49.hxg6 3.33/22 4:45 f5 1.79/26 0 50.exf6 (Qf4) 3.42/21 1:19 50...Bxf6 1.75/22 20 51.Nb6 3.43/23 0 Rd8 (Rc7) 1.77/23 3:10 52.Qh5 4.59/19 39 Qg7 1.85/22 16 53.Kg2 (Qh7+) 4.78/19 0 53...Rf8 3.70/22 2:13 54.Rc1 7.51/22 0 Rxa5 (Re8) 6.11/23 5:57 55.Qxa5 10.47/18 26 d3 6.23/23 1:08 56.Qxb4 (Qh5) 10.05/21 0 1–0

TimoK
Posts: 97
Joined: Sun Jan 03, 2010 11:28 am
Location: Hamburg

Re: Mythbuster continued ...but Komodo comes back

Post by TimoK » Mon Mar 12, 2012 8:04 am

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.

[Event "Houdini-Komodo"]
[Site "?"]
[Date "2012.03.12"]
[Round "27.1"]
[White "Houdini 2.0c Pro x64 [t=1]"]
[Black "Komodo64 SSE Version 4"]
[Result "0-1"]
[ECO "E84"]
[Annotator "0.18;0.16"]
[PlyCount "134"]
[EventDate "2012.03.05"]

{AMD Phenom(tm) II X4 905e Processor 2999 MHz W=24.8 plies; 1.925kN/s; 34 TBAs
B=25.3 plies; 983kN/s} 1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. f3 O-O 6. Be3
a6 7. Qd2 Nc6 8. Nge2 Rb8 9. Rb1 b5 10. cxb5 axb5 11. b4 Bd7 {0.16/24 188} 12.
h4 {0.18/22 326 (Nc1)} Re8 {0.19/23 441 (e6)} 13. Kf2 {0.21/22 193 (g4)} h5 {
0.05/22 239 (e6)} 14. d5 {0.16/21 199 (g3)} Ne5 {-0.15/19 64} 15. Nd4 {0.05/22
92} e6 {0.06/20 104} 16. dxe6 {0.08/22 127 (Bg5)} fxe6 {-0.09/20 140} 17. Kg1 {
0.05/22 1027} c6 {-0.13/24 36 (Nc4)} 18. Rh3 {0.10/20 176} Nh7 {-0.06/21 9
(Qe7)} 19. Rd1 {0.06/20 200 (a4)} Nf8 {-0.08/21 178} 20. Rh1 {-0.05/20 45 (f4)}
c5 {-0.60/21 76 (Nc4)} 21. Bg5 {-0.10/23 129} cxd4 {-0.64/23 84 (Bf6)} 22. Bxd8
{0.17/21 40} dxc3 {-0.60/21 48} 23. Qxc3 {0.09/24 291} Rexd8 {-0.56/26 171} 24.
Qb3 {0.09/22 62 (f4)} Nc6 {-0.64/22 50} 25. g3 {0.03/23 143} Nd4 {-0.77/25 38}
26. Qe3 {-0.03/23 105} e5 {-0.79/25 64} 27. Rh2 {-0.06/23 82} Rdc8 {-0.73/24
151 (Nfe6)} 28. Rf2 {-0.01/22 151} Rc6 {-0.64/22 72 (Nh7)} 29. Rc1 {-0.08/21
275} Ra6 {-0.76/26 54} 30. a3 {-0.16/22 106 (Rc3)} Rba8 {-0.87/24 88} 31. Rc3 {
-0.14/22 151} Nh7 {-0.90/25 88} 32. Bd3 {-0.17/21 121 (Rd3)} Nf6 {-1.09/23 139}
33. Qc1 {-0.30/22 172 (Kg2)} Kh7 {-1.23/23 63} 34. Kh1 {-0.32/23 99 (Qb2)} Be6
{-1.22/23 106 (Bh6)} 35. Kg2 {-0.33/23 741 (Bc2)} R6a7 {-1.20/24 107} 36. Qb2 {
-0.48/23 378 (Kh2)} d5 {-1.24/23 39} 37. Kg1 {-0.54/22 254 (Qc1)} dxe4 {-1.72/
26 407 (Bh6)} 38. Bxe4 {-0.75/22 133} Rf8 {-1.78/25 78 (Bd5)} 39. Qb1 {-1.09/
21 378 (Rc5)} Nxe4 {-1.77/24 123} 40. fxe4 {-1.08/23 0} Bg4 {-1.75/28 335} 41.
Qa2 {-1.12/26 0 (Rxf8)} Rf3 {-1.80/28 294} 42. Rfxf3 {-1.21/25 72} Bxf3 {-1.81/
28 18} 43. Re3 {-1.09/25 125 (Rxf3)} Bg4 {-2.24/27 303 (Rc7)} 44. Kh1 {-1.53/
26 148 (Kg2)} Rc7 {-2.30/27 121} 45. Qd5 {-1.82/28 235} Rc1+ {-2.42/28 536} 46.
Kg2 {-1.64/31 0} Rc2+ {-2.35/27 287 (Be6)} 47. Kf1 {-1.64/28 160} Rc6 {-2.29/
27 318 (Be6)} 48. Qd8 {-1.77/28 238 (Kg2)} Rc1+ {-2.57/27 390 (Bh3+)} 49. Kg2 {
-2.41/28 246} Rc2+ {-2.52/26 126} 50. Kg1 {-2.35/27 0 (Kf1)} Nf3+ {-3.19/27 492
} 51. Rxf3 {-2.72/30 0} Bxf3 {-2.89/27 56} 52. Qd3 {-2.70/28 436} Rg2+ {-3.38/
32 208} 53. Kh1 {-2.96/30 0} Rxg3+ {-3.43/27 35} 54. Kh2 {-2.95/28 13} Rg2+ {
-3.53/27 29} 55. Kh1 {-2.96/29 61} Rf2+ {-3.38/29 160} 56. Kg1 {-2.97/31 0} Be2
{-3.31/27 31} 57. Qc2 {-3.19/26 44} Rf1+ {-3.46/27 15} 58. Kg2 {-3.30/28 63}
Bc4 {-3.55/28 79} 59. a4 {-3.35/27 0} Re1 {-3.44/28 130 (Rf4)} 60. a5 {-3.70/
25 127 (Kf3)} Bh6 {-4.78/29 172} 61. a6 {-4.66/26 16} Be3 {-4.65/28 66} 62. Qc3
{-4.67/28 12} Rg1+ {-4.75/27 96 (Re2+)} 63. Kh2 {-4.67/26 63} Bd4 {-4.94/28 13}
64. a7 {-4.69/27 51 (Qf3)} Bxc3 {-5.18/26 62} 65. a8=Q {-4.41/29 0} Rf1 {-5.10/
28 129} 66. Qa7+ {-5.15/29 28 (Qc6)} Rf7 {-4.98/27 19} 67. Qb8 {-5.39/30 83
(Qc5)} Rf2+ {-5.33/27 94} 0-1

beram
Posts: 1187
Joined: Wed Jan 06, 2010 2:11 pm

Re: Mythbuster continued ...but Komodo comes back

Post by beram » Mon Mar 12, 2012 6:15 pm

TimoK wrote: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, Komodo hit Houdini on his weak spot and that is his overrating of the queen in many positions. This has been observed already in matches and games of Houdini 1.5 against Rybka 4 in the past, where it lost some games with Q vs RR or was far too optimistic in his chances with the Queen.
Rather impressive, Komodo now is the first to win same opening with black and white.

pichy
Posts: 2564
Joined: Thu Mar 09, 2006 2:04 am

Re: Mythbuster continued ...but Komodo comes back

Post by pichy » Tue Mar 13, 2012 12:47 am

TimoK wrote: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:

You forgot to place [D] in front of your fen position like this:

[D]1r1qrnk1/3b2b1/3pp1p1/1pp1n1Bp/1P1NP2P/2N2P2/P2Q2P1/3R1BKR b - - 0 21[/code]
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.

[Event "Houdini-Komodo"]
[Site "?"]
[Date "2012.03.12"]
[Round "27.1"]
[White "Houdini 2.0c Pro x64 [t=1]"]
[Black "Komodo64 SSE Version 4"]
[Result "0-1"]
[ECO "E84"]
[Annotator "0.18;0.16"]
[PlyCount "134"]
[EventDate "2012.03.05"]

{AMD Phenom(tm) II X4 905e Processor 2999 MHz W=24.8 plies; 1.925kN/s; 34 TBAs
B=25.3 plies; 983kN/s} 1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. f3 O-O 6. Be3
a6 7. Qd2 Nc6 8. Nge2 Rb8 9. Rb1 b5 10. cxb5 axb5 11. b4 Bd7 {0.16/24 188} 12.
h4 {0.18/22 326 (Nc1)} Re8 {0.19/23 441 (e6)} 13. Kf2 {0.21/22 193 (g4)} h5 {
0.05/22 239 (e6)} 14. d5 {0.16/21 199 (g3)} Ne5 {-0.15/19 64} 15. Nd4 {0.05/22
92} e6 {0.06/20 104} 16. dxe6 {0.08/22 127 (Bg5)} fxe6 {-0.09/20 140} 17. Kg1 {
0.05/22 1027} c6 {-0.13/24 36 (Nc4)} 18. Rh3 {0.10/20 176} Nh7 {-0.06/21 9
(Qe7)} 19. Rd1 {0.06/20 200 (a4)} Nf8 {-0.08/21 178} 20. Rh1 {-0.05/20 45 (f4)}
c5 {-0.60/21 76 (Nc4)} 21. Bg5 {-0.10/23 129} cxd4 {-0.64/23 84 (Bf6)} 22. Bxd8
{0.17/21 40} dxc3 {-0.60/21 48} 23. Qxc3 {0.09/24 291} Rexd8 {-0.56/26 171} 24.
Qb3 {0.09/22 62 (f4)} Nc6 {-0.64/22 50} 25. g3 {0.03/23 143} Nd4 {-0.77/25 38}
26. Qe3 {-0.03/23 105} e5 {-0.79/25 64} 27. Rh2 {-0.06/23 82} Rdc8 {-0.73/24
151 (Nfe6)} 28. Rf2 {-0.01/22 151} Rc6 {-0.64/22 72 (Nh7)} 29. Rc1 {-0.08/21
275} Ra6 {-0.76/26 54} 30. a3 {-0.16/22 106 (Rc3)} Rba8 {-0.87/24 88} 31. Rc3 {
-0.14/22 151} Nh7 {-0.90/25 88} 32. Bd3 {-0.17/21 121 (Rd3)} Nf6 {-1.09/23 139}
33. Qc1 {-0.30/22 172 (Kg2)} Kh7 {-1.23/23 63} 34. Kh1 {-0.32/23 99 (Qb2)} Be6
{-1.22/23 106 (Bh6)} 35. Kg2 {-0.33/23 741 (Bc2)} R6a7 {-1.20/24 107} 36. Qb2 {
-0.48/23 378 (Kh2)} d5 {-1.24/23 39} 37. Kg1 {-0.54/22 254 (Qc1)} dxe4 {-1.72/
26 407 (Bh6)} 38. Bxe4 {-0.75/22 133} Rf8 {-1.78/25 78 (Bd5)} 39. Qb1 {-1.09/
21 378 (Rc5)} Nxe4 {-1.77/24 123} 40. fxe4 {-1.08/23 0} Bg4 {-1.75/28 335} 41.
Qa2 {-1.12/26 0 (Rxf8)} Rf3 {-1.80/28 294} 42. Rfxf3 {-1.21/25 72} Bxf3 {-1.81/
28 18} 43. Re3 {-1.09/25 125 (Rxf3)} Bg4 {-2.24/27 303 (Rc7)} 44. Kh1 {-1.53/
26 148 (Kg2)} Rc7 {-2.30/27 121} 45. Qd5 {-1.82/28 235} Rc1+ {-2.42/28 536} 46.
Kg2 {-1.64/31 0} Rc2+ {-2.35/27 287 (Be6)} 47. Kf1 {-1.64/28 160} Rc6 {-2.29/
27 318 (Be6)} 48. Qd8 {-1.77/28 238 (Kg2)} Rc1+ {-2.57/27 390 (Bh3+)} 49. Kg2 {
-2.41/28 246} Rc2+ {-2.52/26 126} 50. Kg1 {-2.35/27 0 (Kf1)} Nf3+ {-3.19/27 492
} 51. Rxf3 {-2.72/30 0} Bxf3 {-2.89/27 56} 52. Qd3 {-2.70/28 436} Rg2+ {-3.38/
32 208} 53. Kh1 {-2.96/30 0} Rxg3+ {-3.43/27 35} 54. Kh2 {-2.95/28 13} Rg2+ {
-3.53/27 29} 55. Kh1 {-2.96/29 61} Rf2+ {-3.38/29 160} 56. Kg1 {-2.97/31 0} Be2
{-3.31/27 31} 57. Qc2 {-3.19/26 44} Rf1+ {-3.46/27 15} 58. Kg2 {-3.30/28 63}
Bc4 {-3.55/28 79} 59. a4 {-3.35/27 0} Re1 {-3.44/28 130 (Rf4)} 60. a5 {-3.70/
25 127 (Kf3)} Bh6 {-4.78/29 172} 61. a6 {-4.66/26 16} Be3 {-4.65/28 66} 62. Qc3
{-4.67/28 12} Rg1+ {-4.75/27 96 (Re2+)} 63. Kh2 {-4.67/26 63} Bd4 {-4.94/28 13}
64. a7 {-4.69/27 51 (Qf3)} Bxc3 {-5.18/26 62} 65. a8=Q {-4.41/29 0} Rf1 {-5.10/
28 129} 66. Qa7+ {-5.15/29 28 (Qc6)} Rf7 {-4.98/27 19} 67. Qb8 {-5.39/30 83
(Qc5)} Rf2+ {-5.33/27 94} 0-1

pichy
Posts: 2564
Joined: Thu Mar 09, 2006 2:04 am

Re: Mythbuster continued ...but Komodo comes back

Post by pichy » Tue Mar 13, 2012 12:51 am

beram wrote:
TimoK wrote: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:

[D]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.




Hi Timo, Komodo hit Houdini on his weak spot and that is his overrating of the queen in many positions. This has been observed already in matches and games of Houdini 1.5 against Rybka 4 in the past, where it lost some games with Q vs RR or was far too optimistic in his chances with the Queen.
Rather impressive, Komodo now is the first to win same opening with black and white.
I did not know that Houdini had a weak spot :lol:

Post Reply