Komodo 8 vs Stockfish dev’s (sept14)

[shrapnel]

OK, here we go, for some reason, Komodo 8 plays very well with Black...and really seems to like the French Defence !
[Event "120s/Move"]
[Site "www.come2play.com"]
[Date "2014.09.28"]
[Round "?"]
[White "Stockfish 230914 64 SSE 4.2"]
[Black "Komodo 8 64-bit"]
[Result "0-1"]
[ECO "C13"]
[PlyCount "73"]

{16384MB, Perfect2014t.ctg, i7 3930 k} 1. e4 e6 2. d4 d5 {B} 3. Nc3 Nf6 {B} 4.
Bg5 dxe4 {B} 5. Nxe4 Be7 {B} 6. Bxf6 gxf6 {B} 7. Nf3 f5 {B} 8. Nc3 a6 {B} 9.
Be2 Qd6 {-0.03/23} 10. O-O Nd7 {-0.10/24} 11. a4 {-0.03/19} b6 {-0.11/26} 12.
Re1 Bb7 {-0.16/26} 13. d5 O-O-O {-0.29/26} 14. dxe6 {-0.38/23} fxe6 {-0.33/25}
15. Qxd6 {-0.38/18} Bxd6 {-0.30/27} 16. Bc4 Rhg8 {-0.32/28} 17. Kf1 {-0.32/20}
Nc5 {-0.36/27} 18. Bxe6+ {-0.29/21} Nxe6 {-0.45/29} 19. Rxe6 {-0.44/24} Rg6 {
-0.35/28} 20. Rxg6 {-0.33/24} hxg6 {-0.35/28} 21. a5 b5 {-0.37/28} 22. h3 {-0.
43/25} Bxf3 {-0.43/27} 23. gxf3 {-0.43/19} c6 {-0.35/28} 24. Re1 Bb4 {-0.41/28}
25. Re6 {-0.33/20} Kb7 {-0.45/27} 26. Ne2 Rd1+ {-0.56/27} 27. Kg2 {-0.48/15}
Rd2 {-0.29/29} 28. h4 Rxc2 {-0.77/27} 29. b3 c5 {-1.03/34} 30. Nf4 c4 {-1.41/32
} 31. bxc4 {-1.44/32} Rxc4 {-1.56/32} 32. Nxg6 {-1.57/34} Bxa5 {-1.49/31} 33.
h5 {-1.83/32} Bc3 {-1.96/34} 34. f4 b4 {-2.39/29} 35. h6 a5 {-2.75/30} 36. h7 {
-2.75/33} Rc8 {-2.81/31} 37. Ne7 {-2.90/30} 0-1

[shrapnel]

And here's one with the English Opening with Komodo 8 as White....don't really like 1. e4. In my opinion, 1. e4 gives Black too many options.
The English is unique in that White has a large say in what direction the game will take...
[Event "120s/Move"]
[Site "www.come2play.com"]
[Date "2014.09.28"]
[Round "?"]
[White "Komodo 8 64-bit(a)"]
[Black "Stockfish 230914 64 SSE 4.2"]
[Result "1-0"]
[ECO "D76"]
[PlyCount "97"]

{16384MB, Perfect2014t.ctg, i7 3930 k} 1. c4 g6 2. d4 {B} Bg7 3. Nf3 {B} Nf6 4.
g3 {B} d5 5. cxd5 {B} Nxd5 6. Bg2 {B} O-O 7. O-O {B} Nb6 8. Nc3 {B} Nc6 9. e3 {
0.27/22} Re8 10. d5 {0.15/24} Na5 {0.15/22} 11. Nd4 {0.12/27} Bd7 {0.12/19} 12.
Nb3 {0.10/27} Nxb3 {0.06/21} 13. axb3 {0.11/28} Qc8 14. e4 {0.19/25} Bh3 {0.22/
23} 15. Bxh3 {0.28/26} Qxh3 {0.27/14} 16. Be3 {0.30/26} Qd7 17. Qc2 {0.26/27}
e6 {0.26/19} 18. Rfd1 {0.17/27} a6 19. Rac1 {0.40/26} Rac8 {0.36/21} 20. d6 {
0.37/28} Na8 {0.38/19} 21. dxc7 {0.40/28} Qxc7 {0.36/17} 22. Rd3 {0.36/27} Red8
{0.39/23} 23. Na4 {0.43/27} Rxd3 24. Qxd3 {0.44/27} Qb8 25. Rd1 {0.62/26} Bf6
26. Nc5 {1.13/30} Qc7 {1.02/21} 27. Nd7 {1.26/29} Bxb2 {1.28/25} 28. e5 {1.42/
32} Kh8 29. Qe2 {4.37/35} Rd8 {4.76/38} 30. Qxb2 {4.92/37} Rxd7 {5.01/38} 31.
Rd6 {5.07/39} f6 32. Rxe6 {5.56/31} Kg8 33. Rxf6 {10.49/32} Rd1+ 34. Kg2 {12.
37/32} Qd8 35. Bh6 {59.97/36} Nc7 36. e6 {#32/41} Nxe6 37. Rxe6 {#27/36} Kf7
38. Re4 {#18/36} Qf6 39. Rf4 {#15/43} Qxf4 {#14/43} 40. gxf4 {#11/34} Re1 {#10/
50} 41. Qd2 {#10/49} Re7 {#9/53} 42. Qd6 {#9/54} b6 43. Bg5 {#8/59} Ra7 {#7/88}
44. Qf6+ {#7/91} Ke8 45. Qe6+ {#6/120} Kf8 {#5/120} 46. Bh6+ {#5/120} Rg7 {#4/
120} 47. Qd7 {#4/120} a5 {#3/120} 48. Qxg7+ {#3/120} Ke8 {#2/120} 49. Bg5 {#2/
120} 1-0
These are REAL games, not the gobbledygook most Testers post, which truly show how strong Komodo 8 is !

[Guenther]

OK, here we go, for some reason, Komodo 8 plays very well with Black...and really seems to like the French Defence !
[Event "120s/Move"]
[Site "www.come2play.com"]
[Date "2014.09.28"]
[Round "?"]
[White "Stockfish 230914 64 SSE 4.2"]
[Black "Komodo 8 64-bit"]
[Result "0-1"]
[ECO "C13"]
[PlyCount "73"]

{16384MB, Perfect2014t.ctg, i7 3930 k} 1. e4 e6 2. d4 d5 {B} 3. Nc3 Nf6 {B} 4.
Bg5 dxe4 {B} 5. Nxe4 Be7 {B} 6. Bxf6 gxf6 {B} 7. Nf3 f5 {B} 8. Nc3 a6 {B} 9.
Be2 Qd6 {-0.03/23} 10. O-O Nd7 {-0.10/24} 11. a4 {-0.03/19} b6 {-0.11/26} 12.
Re1 Bb7 {-0.16/26} 13. d5 O-O-O {-0.29/26} 14. dxe6 {-0.38/23} fxe6 {-0.33/25}
15. Qxd6 {-0.38/18} Bxd6 {-0.30/27} 16. Bc4 Rhg8 {-0.32/28} 17. Kf1 {-0.32/20}
Nc5 {-0.36/27} 18. Bxe6+ {-0.29/21} Nxe6 {-0.45/29} 19. Rxe6 {-0.44/24} Rg6 {
-0.35/28} 20. Rxg6 {-0.33/24} hxg6 {-0.35/28} 21. a5 b5 {-0.37/28} 22. h3 {-0.
43/25} Bxf3 {-0.43/27} 23. gxf3 {-0.43/19} c6 {-0.35/28} 24. Re1 Bb4 {-0.41/28}
25. Re6 {-0.33/20} Kb7 {-0.45/27} 26. Ne2 Rd1+ {-0.56/27} 27. Kg2 {-0.48/15}
Rd2 {-0.29/29} 28. h4 Rxc2 {-0.77/27} 29. b3 c5 {-1.03/34} 30. Nf4 c4 {-1.41/32
} 31. bxc4 {-1.44/32} Rxc4 {-1.56/32} 32. Nxg6 {-1.57/34} Bxa5 {-1.49/31} 33.
h5 {-1.83/32} Bc3 {-1.96/34} 34. f4 b4 {-2.39/29} 35. h6 a5 {-2.75/30} 36. h7 {
-2.75/33} Rc8 {-2.81/31} 37. Ne7 {-2.90/30} 0-1

Missing hardware(s) core number hash ponder etc.
Strange depths for Stockfish, sometimes even nada is given for Stockfish moves, looks like crap not *real games*.
Looks like a typical Anul post. Stockfish obviously was severly handicapped or played on a toaster.

[shrapnel]

You don't understand.
This Server is not like Playchess.
Humans meet and decide to play each other, after telling the opponent what he/she is using.
The game you see, is as reported by Fritz 13 GUI on my machine; obviously details to be found on my opponent's machine will be missing.
Follow ?

[Guenther]

You don't understand.
This Server is not like Playchess.
Humans meet and decide to play each other, after telling the opponent what he/she is using.
The game you see, is as reported by Fritz 13 GUI on my machine; obviously details to be found on my opponent's machine will be missing.
Follow ?

Sorry you don't understand. It's just crap games for the given reasons.
You advertised the games as 'real games'. You have no clue if you cannot read the output.
Your opponent propably played during a virus scan or whatever forced his computer to puke. This is all fake...

[lkaufman]

It is clear from the blitz (40/4) lists of CCRL and CEGT that SF5 and K8 are very close, with a slight edge to SF, so it's no surprise that newest SF, which is supposedly +23 over SF5 according to SF website, should beat K8 at blitz levels. But at the main levels of these two lists (40/40 and 40/20 respectively) Komodo 8 has a substantial lead over SF5 on both one core and 4 cores; the average lead for Komodo8 of these four ratings is 25 elo. So even if SF has gained 23 elo, K8 would have a 2 elo lead based on those numbers. It would be interesting to see some matches run at intermediate levels like maybe 20' plus 10" increment for example, slow enough to measure strength at a level typical of how they are used, but fast enough to get big samples on good hardware.

[beram]

It is clear from the blitz (40/4) lists of CCRL and CEGT that SF5 and K8 are very close, with a slight edge to SF, so it's no surprise that newest SF, which is supposedly +23 over SF5 according to SF website, should beat K8 at blitz levels. But at the main levels of these two lists (40/40 and 40/20 respectively) Komodo 8 has a substantial lead over SF5 on both one core and 4 cores; the average lead for Komodo8 of these four ratings is 25 elo. So even if SF has gained 23 elo, K8 would have a 2 elo lead based on those numbers. It would be interesting to see some matches run at intermediate levels like maybe 20' plus 10" increment for example, slow enough to measure strength at a level typical of how they are used, but fast enough to get big samples on good hardware.

Isn't this match of 50 games at TC150m(!) 30 sec not telling enough ?
LTC 150m30 s (6x) WSET (Sedat Canbaz and Tom Casanovas) - 55% for SF - 50 games
http://www.talkchess.com/forum/viewtopic.php?t=53582

[lkaufman]

It is clear from the blitz (40/4) lists of CCRL and CEGT that SF5 and K8 are very close, with a slight edge to SF, so it's no surprise that newest SF, which is supposedly +23 over SF5 according to SF website, should beat K8 at blitz levels. But at the main levels of these two lists (40/40 and 40/20 respectively) Komodo 8 has a substantial lead over SF5 on both one core and 4 cores; the average lead for Komodo8 of these four ratings is 25 elo. So even if SF has gained 23 elo, K8 would have a 2 elo lead based on those numbers. It would be interesting to see some matches run at intermediate levels like maybe 20' plus 10" increment for example, slow enough to measure strength at a level typical of how they are used, but fast enough to get big samples on good hardware.

Isn't this match of 50 games at TC150m(!) 30 sec not telling enough ?
LTC 150m30 s (6x) WSET (Sedat Canbaz and Tom Casanovas) - 55% for SF - 50 games
http://www.talkchess.com/forum/viewtopic.php?t=53582

Hardly. I'm sure you know enough about margin of error to know that this is not statistically significant. We need at least several hundred games at some reasonable intermediate time control.

[beram]

It is clear from the blitz (40/4) lists of CCRL and CEGT that SF5 and K8 are very close, with a slight edge to SF, so it's no surprise that newest SF, which is supposedly +23 over SF5 according to SF website, should beat K8 at blitz levels. But at the main levels of these two lists (40/40 and 40/20 respectively) Komodo 8 has a substantial lead over SF5 on both one core and 4 cores; the average lead for Komodo8 of these four ratings is 25 elo. So even if SF has gained 23 elo, K8 would have a 2 elo lead based on those numbers. It would be interesting to see some matches run at intermediate levels like maybe 20' plus 10" increment for example, slow enough to measure strength at a level typical of how they are used, but fast enough to get big samples on good hardware.

Isn't this match of 50 games at TC150m(!) 30 sec not telling enough ?
LTC 150m30 s (6x) WSET (Sedat Canbaz and Tom Casanovas) - 55% for SF - 50 games
http://www.talkchess.com/forum/viewtopic.php?t=53582

Hardly. I'm sure you know enough about margin of error to know that this is not statistically significant. We need at least several hundred games at some reasonable intermediate time control.

55-57 % for latest SF dev vs K8 is about 35 - 50 ELO stronger
and you mentioned CEGT 40/20 and CCRL, but you didnt mention CEGT 40/120 where the score between SF5 and K8 is almost even 51% for K8
It is difficult to find statistically enough games at intermediate level but I think that the 300 games at10m 6s between SF5 and K8 (from FGRL 149-151 score) show about equal strenght
http://www.fastgm.de/10min.html

[lkaufman]

In general it seems that Stockfish scores better against Komodo than would be expected based on their results against other engines. Contempt may contribute to this, but our tests indicate that it's a pretty minor factor. Perhaps it's just something about their styles. My theory is that because Stockfish is so different from all the other top engines in terms of trading accuracy for depth, it tends to lose or draw a similar number of games to all strong opponents due to the occasional pruned good move, while Komodo just tends to see more than all other engines besides SF most of the time. So Komodo might win a three way event including Houdini or a 4 way event including Gull while still losing the direct match to SF. Of course this all depends on the details like time limit, hardware, etc. Whether strength is best measured by a match of the top two or by a RR of the top three or four is a matter of opinion. Both methods have been used to determine the human World Champion.