Stockfish_0308 vs Komodo5.1MP and Houdini 3

[beram]

At the moment +13 =19 -5 for Stockfish 110813 in LTC match against Komodo 5.1r2 as been played by Arriere Pensee to be followed on Rybka forum(http://rybkaforum.net/cgi-bin/rybkaforu ... 27448;pg=2)

The match is played with LTC 60min 30sec on a very fast system i7 3930 4700Mhz
Fritzmark 42(!) so regularly 25-27 ply depth for Komodo in middlegame

The first match with Noomen suite openingbook with SF 2707 ended in 51,5 -48,5 in favor for Komodo
But now after 37 games 60% for Stockfish 110813 (!)
Some wonderful attacking games by Stockfish have been played for example game 34
Judge by yourself

[pgn]
[Event "Komodo 5.1r2 64-bit vs Stockfish 110813 "]
[Site "NY"]
[Date "2013.08.17"]
[Round "34"]
[White "Stockfish 110813 64 SSE4.2"]
[Black "Komodo 5.1r2 64-bit"]
[Result "1-0"]
[ECO "D45"]
[Annotator "Noomen-Testsuite 2012"]
[PlyCount "141"]
[EventDate "2013.??.??"]
[TimeControl "3600+15"]

{Processor: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz (4700.MHz)^13 ^10 White:
Stockfish 110813 64 SSE4.2, hash size: 2048M, opening book: <none>^13 ^10
Black: Komodo 5.1r2 64-bit , hash size: 2048M, opening book: <none>^13 ^10 [%t
Long] White checkmates.} 1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 e6 5. e3 a6 6. b3
Bb4 7. Bd2 Bd6 8. Qc2 Nbd7 {White out of book} 9. e4 {+0.24|d29|110s (e5)
Black out of book} dxe4 {+0.24|d21|31s} 10. Nxe4 {+0.32|d31|72s} Nxe4 {+0.
27|d24|51s} 11. Qxe4 {+0.22|d32|121s} e5 {+0.06|d24|72s (dxe5)} 12. Nxe5 {+0.
22|d32|88s (Qe7)} O-O {+0.13|d24|61s} 13. O-O-O {+0.68|d28|116s} Re8 {+0.
13|d26|78s} 14. Bd3 {+0.54|d33|82s} Nf6 {+0.13|d27|80s} 15. Qf3 {+0.56|d34|130s
} c5 {+0.13|d27|78s} 16. Bg5 {+0.64|d34|94s} cxd4 {+0.26|d23|42s} 17. Bxf6 {+0.
64|d32|62s (Qxf6)} gxf6 {+0.13|d27|387s (Nxf7)} 18. Ng4 {+0.74|d31|119s} Kg7 {
+0.26|d24|129s} 19. c5 {+0.84|d32|135s (Bb8)} Bf8 {+0.26|d24|145s (Qg3)} 20.
Rhe1 {+1.09|d32|242s (f5)} Rxe1 {+0.58|d24|75s} 21. Rxe1 {+1.09|d31|27s} f5 {
+0.67|d25|75s} 22. Ne5 {+1.17|d31|21s} Bxc5 {+0.64|d24|63s (Kb1)} 23. Bc4 {+1.
03|d31|206s (Qh4)} Ba3+ {+0.59|d23|91s} (23... Qh4 24. Nd3 Ba3+ (24... Bd6 25.
Kb1) 25. Kb1 f4 26. g4) 24. Kb1 {+1.09|d32|214s} f4 {+0.54|d25|143s (Qxf4)} 25.
g4 {+1.19|d32|205s (Qf6)} b5 {+0.43|d24|146s (Bd3)} 26. Nxf7 {+2.86|d31|69s}
Qa5 {+1.63|d24|86s} 27. Qe4 {+3.27|d30|21s} bxc4 {+1.79|d26|67s} 28. Qxd4+ {+3.
31|d30|59s} Kxf7 {+1.78|d28|54s} 29. Re5 {+3.33|d31|70s} Qxe5 {+1.89|d27|56s}
30. Qxe5 {+3.41|d32|117s} Be6 {+1.88|d28|49s} 31. Qc7+ {+3.41|d30|7s (Kf8)} Kf6
{+1.89|d28|102s} 32. Qxf4+ {+3.81|d28|33s} Ke7 {+2.05|d26|40s} 33. Qc7+ {+3.
91|d30|189s} Kf6 {+2.07|d27|54s} 34. f4 {+3.95|d30|351s} cxb3 {+2.14|d27|57s}
35. axb3 {+4.28|d30|109s} Be7 {+2.15|d28|141s} 36. f5 {+4.36|d29|19s (Bd5)}
Bxf5+ {+2.25|d29|51s} 37. gxf5 {+4.70|d27|19s} Rf8 {+2.29|d28|48s (Qa7)} 38. b4
{+5.17|d30|66s} Rd8 {+2.57|d26|70s} 39. Qf4 {+5.51|d31|49s} Kg7 {+2.77|d27|62s}
40. Qe3 {+6.60|d31|70s (Rd1)} Rd7 {+2.96|d27|46s} 41. Qe2 {+6.76|d33|83s (Kf8)}
h5 {+3.25|d24|85s (Qxa6)} 42. Qe6 {+8.18|d30|27s} Rb7 {+3.39|d27|14s} 43. Qg6+
{+8.38|d31|26s} Kf8 {+4.16|d26|19s} 44. Qh6+ {+14.47|d31|58s} Kf7 {+5.
66|d26|65s (Qxa6)} 45. Qxh5+ {+21.37|d25|11s (Kg7)} Kf6 {+5.66|d26|22s} 46. Kc2
{+20.38|d28|16s} Bxb4 {+7.84|d26|49s} 47. Qh8+ {+89.04|d29|40s} Rg7 {+8.
63|d29|20s} 48. Qh6+ {+89.19|d32|36s} Kf7 {+317.92|d29|58s (Qe6)} 49. Qxa6 {
+89.19|d33|51s} Be7 {+32.80|d28|16s} 50. Qe6+ {+89.29|d32|46s} Kf8 {+34.
81|d31|13s} 51. f6 {+90.70|d32|41s} Bxf6 {+315.28|d33|33s} 52. Qxf6+ {+90.
70|d31|14s} Kg8 {+317.98|d27|28s (h4)} 53. Kd3 {+137.31|d31|13s (Ra7)} Rf7 {
+317.98|d24|31s} 54. Qe6 {+299.68|d28|44s (Kh8)} Kg7 {+299.74|d23|53s} 55. h4 {
+299.76|d28|19s} Rf3+ {+299.78|d22|10s} 56. Ke4 {+299.81|d30|58s (Rf6)} Rf1 {
+299.81|d20|6s (Qg4)} 57. Qb6 {+299.84|d30|59s (Rf7)} Rf6 {+299.80|d20|3s} 58.
Qd4 {+299.86|d31|19s} Kg6 {+299.86|d22|2s} 59. h5+ {+299.87|d30|10s} Kf7 {+299.
87|d23|4s} 60. Ke5 {+299.88|d32|33s} Re6+ {+299.88|d23|4s} 61. Kf5 {+299.
89|d36|36s} Re7 {+299.89|d22|3s (Qc4)} 62. Qc5 {+299.90|d38|28s} Re1 {+299.
90|d22|5s} 63. h6 {+299.91|d46|26s} Rf1+ {+299.91|d22|6s} 64. Kg5 {+299.
92|d62|26s (Rh1)} Rd1 {+299.92|d21|5s (Qf2)} 65. h7 {+299.93|d72|33s (Ke6)} Rh1
{+299.94|d20|3s} 66. Qd5+ {+299.94|d95|31s (Ke8)} Ke7 {+299.95|d19|2s} 67. Qxh1
{+299.95|d100} Kd6 {+299.96|d18|3s} 68. h8=Q {+299.96|d100 (Kc5)} Kc7 {+299.
97|d17|3s (Qe5)} 69. Qc3+ {+299.97|d100 (Kd7)} Kd8 {+299.98|d19|4s} 70. Qhc6 {
+299.98|d100} Ke7 {+299.99|d24|3s} 71. Q3f6# {+299.99|d100} 1-0
[/pgn]

[shrapnel]

Thank you for this new time control test.
I think you will find that Stockfish 100813 is even stronger!

youe were wright Paul,

Code: Select all

AMD X6 1090T 3200 Mhz
Bram Testsuite 2.0 - 50 games, TC 4m+2s, 4CPU, ponder off

Stockfish 110813 64 SSE4.2   - Komodo 5.1r2 64-bit            30   - 20      +17/=26/-7    60.00%

earlier result with SF 0308:

Code: Select all

Stockfish 030813 64 SSE4.2   - Komodo 5.1r2 64-bit            29.0 - 21.0    +15/=28/-7    58.00%

against Houdini lesser result

Code: Select all

AMD X6 1090T 3200 Mhz
Bram Testsuite 2.0 - 50 games, TC 4m+2s, 4CPU, ponder off
                                
1   Houdini 3 x64               +19/=20/-11 58.00%   29.0/50
2   Stockfish 110813 64 SSE4.2  +11/=20/-19 42.00%   21.0/50

Actually, latest Stockfish versions are regularly BEATING Houdini 3
All that is required is a change in one of the parameters/settings of Stockfish !! For obvious reasons, I won't mention what that parameter is, but you can take my word for it (or not ! ) that the days of Houdini 3 supremacy are over !
Hint :- I got the idea from the post of a member on this very Forum. As is usual here, his idea/hypothesis was ridiculed by the wise members here.
But I checked out his hypothesis for myself and did indeed find that Stockfish with the changed parameter was consistently beating Houdini 3 !
I don't really care whether folk here believe me or not, but I for one am dead certain that its time R. Houdart released Houdini 4 !

[beram]

New Stockfish results in my ongoing test series

Code: Select all

AMD X6 1090T 3200 Mhz Bram Testsuite 2.0 - 50 games, TC 4m+2s, 4CPU, ponder off

Stockfish 300813 64 SSE4.2   - Komodo 5.1r2 64-bit            +17/=23/-10 57.00%   28.5/50

So no improvement for 'Stockie' over the former matches with SF0308 and SF1108, but less draws so more fun to watch

earlier results with SF 0308 and SF1108

Code: Select all

Stockfish 030813 64 SSE4.2   - Komodo 5.1r2 64-bit            29.0 - 21.0    +15/=28/-7    58.00%

Stockfish 110813 64 SSE4.2   - Komodo 5.1r2 64-bit            30   - 20      +17/=26/-7    60.00%

[beram]

First time a version of Stockfish wins from Houdini 3 in my test matches
It happened last night with Stockfish 300813

Code: Select all

Stockfish 300813 64 SSE4.2   - Houdini 3 x64         +11/=30/-9 
                                
1   Stockfish 300813 64 SSE4.2  +11/=30/-9 52.00%   26.0/50
2   Houdini 3 x64               +9/=30/-11 48.00%   24.0/50

Here is one game example.
If there is interest I will post a complete games link

[pgn]
[Event "Blitz 4m+2s 4CPU"]
[Site "AMD X6 1090T 3200 Mhz, (Fritz"]
[Date "2013.09.01"]
[Round "45"]
[White "Stockfish 300813 64 SSE4.2"]
[Black "Houdini 3 x64"]
[Result "1-0"]
[ECO "C92"]
[Annotator "0.24;0.11"]
[PlyCount "137"]
[TimeControl "240+2"]

{AMD Phenom(tm) II X6 1090T Processor 3207 MHz W=24.8 plies; 5.232kN/s B=20.
6 plies; 7.745kN/s} 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1
b5 7. Bb3 d6 8. c3 O-O 9. h3 Bb7 10. d4 Re8 11. Nbd2 Bf8 12. a4 h6 13. Bc2 exd4
14. cxd4 Nb4 15. Bb1 c5 {Both last book move} 16. d5 {[%eval 24,22] [%emt 0:00:
06]} g6 {[%eval 11,19] [%emt 0:00:11]} 17. Nf1 {[%eval 6,25] [%emt 0:00:12]}
Bg7 {[%eval 4,21] [%emt 0:00:22]} 18. axb5 {[%eval 6,23] [%emt 0:00:06]} axb5 {
[%eval 4,21] [%emt 0:00:06]} 19. Rxa8 {[%eval 22,25] [%emt 0:00:06]} Qxa8 {
[%eval 5,22] [%emt 0:00:06]} 20. Bf4 {[%eval 22,26] [%emt 0:00:05]} Qd8 {
[%eval -2,21] [%emt 0:00:13] (Qa6)} 21. Qc1 {[%eval 32,23] [%emt 0:00:07] (Bd2)
} Kh7 {[%eval 7,22] [%emt 0:00:38] (h5)} 22. Bd2 {[%eval 32,20] [%emt 0:00:05]}
Na6 {[%eval 1,20] [%emt 0:00:07] (Nd7)} 23. b3 {[%eval 46,21] [%emt 0:00:04]}
Qe7 {[%eval 3,19] [%emt 0:00:05] (Nc7)} 24. Ng3 {[%eval 40,22] [%emt 0:00:11]
(Bc3)} b4 {[%eval 10,17] [%emt 0:00:06] (Nc7)} 25. h4 {[%eval 50,21] [%emt 0:
00:10]} Qd7 {[%eval 22,20] [%emt 0:00:19] (Qd8)} 26. h5 {[%eval 48,21] [%emt 0:
00:07]} Ng4 {[%eval 22,18] [%emt 0:00:00]} 27. Nh4 {[%eval 42,22] [%emt 0:00:
05] (Qc2)} Bd4 {[%eval 13,17] [%emt 0:00:06]} 28. Rf1 {[%eval 60,23] [%emt 0:
00:06]} Ne5 {[%eval 52,17] [%emt 0:00:09] (Rg8)} 29. Ne2 {[%eval 86,23] [%emt
0:00:11] (hxg6+)} Qg4 {[%eval 65,20] [%emt 0:00:05]} 30. hxg6+ {[%eval 155,24]
[%emt 0:00:18]} fxg6 {[%eval 62,20] [%emt 0:00:05]} 31. Nxd4 {[%eval 163,25]
[%emt 0:00:02]} cxd4 {[%eval 67,20] [%emt 0:00:07]} 32. f4 {[%eval 171,26]
[%emt 0:00:06]} Nd7 {[%eval 68,21] [%emt 0:00:11]} 33. Nf3 {[%eval 161,23]
[%emt 0:00:05]} Nac5 {[%eval 97,21] [%emt 0:00:22] (Ndc5)} 34. Nxd4 {[%eval
175,22] [%emt 0:00:07] (Re1)} Nf6 {[%eval 204,16] [%emt 0:00:06] (Ba6)} 35. f5
{[%eval 296,19] [%emt 0:00:02]} g5 {[%eval 257,17] [%emt 0:00:07]} 36. Ne6 {
[%eval 315,24] [%emt 0:00:08]} Rc8 {[%eval 303,20] [%emt 0:00:21] (Kg8)} 37.
Nxc5 {[%eval 363,25] [%emt 0:00:15]} Ba6 {[%eval 304,19] [%emt 0:00:04]} 38.
Rf3 {[%eval 361,26] [%emt 0:00:03]} Rxc5 {[%eval 317,20] [%emt 0:00:03]} 39.
Qe1 {[%eval 400,26] [%emt 0:00:07]} Qh4 {[%eval 339,21] [%emt 0:00:04]} 40.
Qxh4 {[%eval 367,26] [%emt 0:00:06]} gxh4 {[%eval 307,20] [%emt 0:00:00]} 41.
Re3 {[%eval 381,26] [%emt 0:00:03]} Rb5 {[%eval 271,23] [%emt 0:00:13]} 42. Re1
{[%eval 402,26] [%emt 0:00:04]} Ng4 {[%eval 280,23] [%emt 0:00:03]} 43. Rc1 {
[%eval 347,27] [%emt 0:00:04]} Rb7 {[%eval 268,23] [%emt 0:00:09]} 44. Be1 {
[%eval 383,29] [%emt 0:00:25]} Be2 {[%eval 282,22] [%emt 0:00:03] (Bb5)} 45.
Bxh4 {[%eval 381,29] [%emt 0:00:03]} Ra7 {[%eval 268,23] [%emt 0:00:04]} 46.
Bg3 {[%eval 381,29] [%emt 0:00:05]} Ra1 {[%eval 287,23] [%emt 0:00:03]} 47. Bf4
{[%eval 412,30] [%emt 0:00:22]} Kg7 {[%eval 281,23] [%emt 0:00:03]} 48. Bd2 {
[%eval 426,30] [%emt 0:00:10] (Bc2)} Bb5 {[%eval 276,22] [%emt 0:00:04] (h5)}
49. Re1 {[%eval 414,26] [%emt 0:00:03] (Bxb4)} Kf7 {[%eval 289,22] [%emt 0:00:
03] (Ra6)} 50. Bxb4 {[%eval 410,27] [%emt 0:00:04]} h5 {[%eval 285,24] [%emt 0:
00:03] (Ra6)} 51. g3 {[%eval 426,25] [%emt 0:00:03] (Bc2)} Ne5 {[%eval 299,22]
[%emt 0:00:04] (Ra6)} 52. Bc3 {[%eval 464,28] [%emt 0:00:06]} Ra7 {[%eval 301,
22] [%emt 0:00:03] (Ra3)} 53. Bxe5 {[%eval 478,27] [%emt 0:00:04]} dxe5 {
[%eval 296,20] [%emt 0:00:00]} 54. Rc1 {[%eval 494,27] [%emt 0:00:04]} Ke7 {
[%eval 307,21] [%emt 0:00:04]} 55. Rc5 {[%eval 503,27] [%emt 0:00:03] (Kf2)}
Bd7 {[%eval 321,22] [%emt 0:00:04] (Rb7)} 56. Bd3 {[%eval 527,25] [%emt 0:00:
02] (Kg2)} Kd6 {[%eval 338,20] [%emt 0:00:06] (Rb7)} 57. Rc3 {[%eval 543,26]
[%emt 0:00:05] (Rc2)} Rb7 {[%eval 358,22] [%emt 0:00:03]} 58. Kf2 {[%eval 539,
25] [%emt 0:00:04]} Rb4 {[%eval 374,21] [%emt 0:00:03] (Rb8)} 59. Bc4 {[%eval
585,24] [%emt 0:00:02] (Ke3)} Rb8 {[%eval 401,22] [%emt 0:00:08]} 60. Rc2 {
[%eval 585,26] [%emt 0:00:04] (Be2)} Ra8 {[%eval 396,21] [%emt 0:00:02] (Rg8)}
61. b4 {[%eval 624,22] [%emt 0:00:03] (Be2)} Rc8 {[%eval 433,18] [%emt 0:00:03]
} 62. Bd3 {[%eval 664,24] [%emt 0:00:05]} Rb8 {[%eval 461,20] [%emt 0:00:02]
(Rf8)} 63. b5 {[%eval 715,25] [%emt 0:00:04]} Rb7 {[%eval 489,21] [%emt 0:00:
03] (Rb6)} 64. Be2 {[%eval 771,25] [%emt 0:00:04]} Be8 {[%eval 517,21] [%emt 0:
00:06] (Rb6)} 65. Ke3 {[%eval 795,25] [%emt 0:00:03] (Rb2)} Rb8 {[%eval 540,20]
[%emt 0:00:03] (Ke7)} 66. f6 {[%eval 965,24] [%emt 0:00:05] (Rb2)} Ra8 {[%eval
664,21] [%emt 0:00:05]} 67. b6 {[%eval 1050,25] [%emt 0:00:02]} Rb8 {[%eval
691,21] [%emt 0:00:02] (Ra3+)} 68. Rb2 {[%eval 1351,22] [%emt 0:00:01] (Rc7)}
Kd7 {[%eval 1230,22] [%emt 0:00:09] (h4)} 69. b7 {[%eval 2418,24] [%emt 0:00:
02] (Ba6)} 1-0
[/pgn]

[ernest]

So no improvement for 'Stockie' over the former matches with SF0308 and SF1108

Hi Bram,

What I will say is no fun, but it is difficult to discern "improvement" with only 50-game matches.
(the 68% error-bar is around 5%, the 95% error-bar is around 10%...)

[beram]

So no improvement for 'Stockie' over the former matches with SF0308 and SF1108

Hi Bram,

What I will say is no fun, but it is difficult to discern "improvement" with only 50-game matches.
(the 68% error-bar is around 5%, the 95% error-bar is around 10%...)

Well Ernest, lets keep it in between quotation marks than - no "improvement"

[beram]

First time a version of Stockfish wins from Houdini 3 in my test matches
It happened last night with Stockfish 300813

Code: Select all

Stockfish 300813 64 SSE4.2   - Houdini 3 x64         +11/=30/-9 
                                
1   Stockfish 300813 64 SSE4.2  +11/=30/-9 52.00%   26.0/50
2   Houdini 3 x64               +9/=30/-11 48.00%   24.0/50

And - to no surprise - a second time a Stockfish version wins a match against Houdini 3 in my testing

Code: Select all

Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00% 
AMD X6 1090T 3200 Mhz, Bram Testsuite 2a - 50 games, TC 4m+2s, 4CPU, ponder off
                                
1   Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00%   27.5/50
2   Houdini 3 x64               +7/=31/-12 45.00%   22.5/50

[beram]

First time a version of Stockfish wins from Houdini 3 in my test matches
It happened last night with Stockfish 300813

Code: Select all

Stockfish 300813 64 SSE4.2   - Houdini 3 x64         +11/=30/-9 
                                
1   Stockfish 300813 64 SSE4.2  +11/=30/-9 52.00%   26.0/50
2   Houdini 3 x64               +9/=30/-11 48.00%   24.0/50

And - to no surprise - a second time a Stockfish version wins a match against Houdini 3 in my testing

Code: Select all

Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00% 
AMD X6 1090T 3200 Mhz, Bram Testsuite 2a - 50 games, TC 4m+2s, 4CPU, ponder off
                                
1   Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00%   27.5/50
2   Houdini 3 x64               +7/=31/-12 45.00%   22.5/50

and result against Komodo 5.1r2 - 56%

Code: Select all

 Stockfish 100913 64 SSE4.2  +15/=26/-9 56.00%    
                         
1   Stockfish 100913 64 SSE4.2  +15/=26/-9 56.00%   28.0/50
2   Komodo 5.1r2 64-bit         +9/=26/-15 44.00%   22.0/50

[beram]

First time a version of Stockfish wins from Houdini 3 in my test matches
It happened last night with Stockfish 300813

Code: Select all

Stockfish 300813 64 SSE4.2   - Houdini 3 x64         +11/=30/-9 
                                
1   Stockfish 300813 64 SSE4.2  +11/=30/-9 52.00%   26.0/50
2   Houdini 3 x64               +9/=30/-11 48.00%   24.0/50

And - to no surprise - a second time a Stockfish version wins a match against Houdini 3 in my testing

Code: Select all

Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00% 
AMD X6 1090T 3200 Mhz, Bram Testsuite 2a - 50 games, TC 4m+2s, 4CPU, ponder off
                                
1   Stockfish 100913 64 SSE4.2  +12/=31/-7 55.00%   27.5/50
2   Houdini 3 x64               +7/=31/-12 45.00%   22.5/50

and result against Komodo 5.1r2 - 56%

Code: Select all

 Stockfish 100913 64 SSE4.2  +15/=26/-9 56.00%    
                         
1   Stockfish 100913 64 SSE4.2  +15/=26/-9 56.00%   28.0/50
2   Komodo 5.1r2 64-bit         +9/=26/-15 44.00%   22.0/50

I have no explanation but Houdini 3 now strikes back against latest Stockfish 1209

Code: Select all

AMD X6 1090T 3200 Mhz
Bram Testsuite 2a - 50 games, TC 3m+2s, 4CPU, ponder off
                                
1   Houdini 3 x64               +17/=23/-10 57.00%   28.5/50
2   Stockfish 120913 64 SSE4.2  +10/=23/-17 43.00%   21.5/50

[ernest]

I have no explanation...

...except error-bar!

(do 50 more, you will see...)