Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Remarkable miniature in LTC game Stockfish vs Komodo
- - By Peter Grayson (****) Date 2016-09-30 21:09 Upvotes 2
LTC miniatures with two top engines competing are a rarity these days so when I saw this miniature produced in a Stockfish vs Komodo game while testing some lines for worthiness for inclusion in a new opening test suite it required deeper inspection to check what happened. The more I look at it, the more I enthuse about Stockfish' moves 17-19.

The opening line terminated with Black's 10..Nd7 in a popular Sicilian Najdorf 6.Be3 line. With opposite side castling the thematic of the line is quite clear with attacking opportunities on opposite sides. King side for White and Queen side for Black. With Stockfish playing White and Komodo with Black, the game progressed along well established moves until Stockfish played 17.Kb1! That on its own was fairly remarkable when engines prefer the material balance and would capture by 17.axb. Why was 17.Kb1 remarkable? Because given the subsequent moves it displayed very strong signs of artificial intelligence rather than pure number crunching.

Stockfish' sequence of 17.Kb1, 18 Nc1 with 19.Ka1 seemed to be an "idea" of using Black's "a" pawn to form a pawn shelter and in the process stymied Komodo's Queen side attack leaving Komodo at a loss of how to proceed. Stockfish sacrificed two pawns to enable this! One look at the scoring of the moves during this sequence confirms Stockfish "knew" what it was doing. Komodo had it equal at move 20..Qc6, by move 25 it could have resigned but allowing the game to continue, Stockfish mated Komodo on move 41.

Time control: Game in 120m + 30s
Ponder=off, 4 cores and 2 Gb hash per engine. Both default settings except Komodo where Contempt = 0.
Machine: CPU Q9550 at 3.53GHz with 8 Gb

[Event "Stockfish 240916 vs Komodo 10.1, Slow 1"]
[Site "DEVO4"]
[Date "2016.09.27"]
[Round "1"]
[White "Stockfish 240916 64"]
[Black "Komodo 10.1 64-bit"]
[Result "1-0"]
[ECO "B90"]
[Annotator "0.21;0.29"]
[PlyCount "81"]
[TimeControl "7200+30"]

{Intel(R) Core(TM)2 Quad CPU    Q9550  @ 2.83GHz 3536 MHz  W=54.2 plies; 5,
100kN/s; 260,318 TBAs  B=37.3 plies; 5,282kN/s; 516,300 TBAs} 1. e4 c5 2. Nf3
d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8. f3 Be7 9. Qd2 O-O
10. O-O-O Nbd7 {Both last book move} 11. g4 {[%eval 21,34] [%emt 0:05:29]} b5 {
[%eval 29,32] [%emt 0:04:27]} 12. g5 {[%eval 31,33] [%emt 0:02:36]} b4 {[%eval
24,31] [%emt 0:01:55] (Nh5)} 13. Ne2 {[%eval 45,31] [%emt 0:01:30]} Ne8 {
[%eval 16,29] [%emt 0:01:45]} 14. f4 {[%eval 23,31] [%emt 0:01:16] (h4)} a5 {
[%eval 29,29] [%emt 0:05:45]} 15. f5 {[%eval 13,35] [%emt 0:10:54]} Bxb3 {
[%eval 27,31] [%emt 0:05:03]} 16. cxb3 {[%eval 35,29] [%emt 0:00:34]} a4 {
[%eval 24,31] [%emt 0:02:57]} 17. Kb1 {[%eval 36,31] [%emt 0:03:18] (bxa4)}
axb3 {[%eval -15,30] [%emt 0:03:19]} 18. Nc1 {[%eval 35,30] [%emt 0:00:29]
(axb3)} bxa2+ {[%eval -41,27] [%emt 0:02:33]} 19. Ka1 {[%eval 53,31] [%emt 0:
01:42]} Qc7 {[%eval -22,27] [%emt 0:04:09]} 20. Nb3 {[%eval 61,29] [%emt 0:00:
44]} Qc6 {[%eval 0,30] [%emt 0:05:59] (Nc5)} 21. Rc1 {[%eval 113,29] [%emt 0:
01:38] (Rg1)} Nc5 {[%eval 37,30] [%emt 0:05:09]} 22. Nxc5 {[%eval 168,31]
[%emt 0:02:28]} dxc5 {[%eval 39,30] [%emt 0:02:19]} 23. Bc4 {[%eval 183,30]
[%emt 0:03:11]} Ra5 {[%eval 50,32] [%emt 0:05:56] (Rd8)} 24. f6 {[%eval 350,32]
[%emt 0:02:11] (Qd5)} Bd6 {[%eval 161,28] [%emt 0:02:56] (Nxf6)} 25. h4 {
[%eval 559,28] [%emt 0:01:57] (Bd5)} Ra8 {[%eval 304,27] [%emt 0:08:22] (h5)}
26. h5 {[%eval 764,29] [%emt 0:02:18] (Bd5)} Ra7 {[%eval 538,27] [%emt 0:08:05]
(Rd8)} 27. Qg2 {[%eval 1112,30] [%emt 0:02:17] (fxg7)} Kh8 {[%eval 742,28]
[%emt 0:03:48]} 28. fxg7+ {[%eval 1200,30] [%emt 0:00:36]} Nxg7 {[%eval 1224,
31] [%emt 0:06:59]} 29. g6 {[%eval 1547,33] [%emt 0:06:10]} Rc7 {[%eval 1338,
30] [%emt 0:05:13] (Rd7)} 30. Rcg1 {[%eval 12827,36] [%emt 0:02:06] (h6)} Qd7 {
[%eval 25000,31] [%emt 0:08:08] (Qb7)} 31. gxh7 {[%eval 32746,55] [%emt 0:01:
44]} f5 {[%eval 32748,28] [%emt 0:00:22]} 32. h6 {[%eval 32748,58] [%emt 0:00:
34]} Re8 {[%eval 32750,27] [%emt 0:00:15] (Rd8)} 33. hxg7+ {[%eval 32750,71]
[%emt 0:02:40]} Qxg7 {[%eval 32752,5] [%emt 0:00:00]} 34. Qd2 {[%eval 32752,76]
[%emt 0:00:36]} f4 {[%eval 32754,31] [%emt 0:00:16]} 35. Qxd6 {[%eval 32754,
109] [%emt 0:02:34]} Ra7 {[%eval 32756,36] [%emt 0:00:15] (Rce7)} 36. Rxg7 {
[%eval 32756,127] [%emt 0:01:28]} Rxg7 {[%eval 32758,99] [%emt 0:00:13]} 37.
Qf6 {[%eval 32758,127] [%emt 0:00:00]} fxe3 {[%eval 32760,99] [%emt 0:00:00]}
38. Rg1 {[%eval 32760,127] [%emt 0:00:00]} Ree7 {[%eval 32762,99] [%emt 0:00:
00]} 39. Qf8+ {[%eval 32762,127] [%emt 0:00:00]} Kxh7 {[%eval 32764,99] [%emt
0:00:00]} 40. Rh1+ {[%eval 32764,127] [%emt 0:00:00]} Kg6 {[%eval 32766,5]
[%emt 0:00:00]} 41. Qf5# {[%eval 32766,127] [%emt 0:00:00]} 1-0
Parent - By Peter Grayson (****) Date 2016-09-30 21:11 Upvotes 2
Of course it crossed my mind that despite database games to the contrary, maybe the line is an easy win for White. So what was even more remarkable with the return game was that Stockfish also saw off Komodo with reversed colours and the notable similarity to the first game was that Stockfish used Komodo's advanced "h" pawn to form a King shelter just as it had done with Komodo's "a" pawn in the first game. So Stockfish was able to win with both sides.

In this return game, Stockfish deviated with 12..Nh5

[Event "Stockfish 240916 vs Komodo 10.1, Slow 1"]
[Site "DEVO4"]
[Date "2016.09.27"]
[Round "2"]
[White "Komodo 10.1 64-bit"]
[Black "Stockfish 240916 64"]
[Result "0-1"]
[ECO "B90"]
[Annotator "0.30;0.35"]
[PlyCount "158"]
[TimeControl "7200+30"]

{Intel(R) Core(TM)2 Quad CPU    Q9550  @ 2.83GHz 3536 MHz  W=33.9 plies; 5,
112kN/s; 28,692,093 TBAs  B=44.5 plies; 5,655kN/s; 43,296,308 TBAs} 1. e4 c5 2.
Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e5 7. Nb3 Be6 8. f3 Be7 9. Qd2
O-O 10. O-O-O Nbd7 {Both last book move} 11. g4 {[%eval 30,32] [%emt 0:03:13]}
b5 {[%eval 35,33] [%emt 0:05:37]} 12. g5 {[%eval 29,33] [%emt 0:04:32]} Nh5 {
[%eval 16,31] [%emt 0:02:11] (b4)} 13. Nd5 {[%eval 32,32] [%emt 0:07:24]} Bxd5
{[%eval 14,33] [%emt 0:03:01]} 14. exd5 {[%eval 16,31] [%emt 0:01:47]} f6 {
[%eval 19,30] [%emt 0:00:50]} 15. h4 {[%eval 23,31] [%emt 0:03:46] (g6)} Ng3 {
[%eval -7,32] [%emt 0:01:55] (fxg5)} 16. Rg1 {[%eval 28,30] [%emt 0:02:08]}
Nxf1 {[%eval 0,31] [%emt 0:00:30]} 17. Rdxf1 {[%eval 32,30] [%emt 0:01:42]} a5
{[%eval 0,36] [%emt 0:02:37]} 18. Qd3 {[%eval 24,32] [%emt 0:04:34]} b4 {
[%eval 0,36] [%emt 0:01:26] (a4)} 19. g6 {[%eval 75,27] [%emt 0:01:40] (Kb1)}
f5 {[%eval 0,35] [%emt 0:02:31]} 20. gxh7+ {[%eval 60,29] [%emt 0:02:43]} Kh8 {
[%eval 0,36] [%emt 0:01:00] (Kxh7)} 21. Rg2 {[%eval 39,28] [%emt 0:07:46] (h5)}
a4 {[%eval -41,31] [%emt 0:01:45]} 22. Nd2 {[%eval 31,27] [%emt 0:01:44]} Bxh4
{[%eval -38,33] [%emt 0:00:57]} 23. Rfg1 {[%eval 21,29] [%emt 0:04:07]} Rf7 {
[%eval -47,35] [%emt 0:02:08]} 24. Nc4 {[%eval 14,29] [%emt 0:03:56]} Qc7 {
[%eval -62,31] [%emt 0:00:34] (Qe7)} 25. Kb1 {[%eval 0,30] [%emt 0:06:11]} Be7
{[%eval -41,36] [%emt 0:21:46] (f4)} 26. Rd1 {[%eval 26,28] [%emt 0:04:43]} Nc5
{[%eval -45,33] [%emt 0:03:04] (Rc8)} 27. Qe2 {[%eval 15,28] [%emt 0:05:12]
(Qf1)} e4 {[%eval -82,30] [%emt 0:03:25] (b3)} 28. f4 {[%eval 0,30] [%emt 0:04:
02]} b3 {[%eval -95,34] [%emt 0:03:31] (Bf6)} 29. Bd4 {[%eval -34,28] [%emt 0:
04:24] (a3)} Bf6 {[%eval -104,34] [%emt 0:05:41]} 30. Bxf6 {[%eval -39,29]
[%emt 0:04:50]} Rxf6 {[%eval -112,33] [%emt 0:00:31]} 31. Rdg1 {[%eval -63,29]
[%emt 0:05:43]} Rf7 {[%eval -86,39] [%emt 0:05:55]} 32. Qe3 {[%eval -66,29]
[%emt 0:02:55]} Rd8 {[%eval -86,36] [%emt 0:00:37]} 33. Rd1 {[%eval -52,30]
[%emt 0:05:28] (Qd4)} Qa7 {[%eval -122,30] [%emt 0:01:41] (Qe7)} 34. Qc3 {
[%eval -88,30] [%emt 0:06:58]} bxc2+ {[%eval -135,30] [%emt 0:00:34]} 35. Rxc2
{[%eval -73,32] [%emt 0:02:11]} Qe7 {[%eval -150,34] [%emt 0:02:08]} 36. Rf1 {
[%eval -86,32] [%emt 0:03:53] (Qa5)} Rff8 {[%eval -137,35] [%emt 0:05:04] (Nd3)
} 37. Rg1 {[%eval -90,29] [%emt 0:03:18] (Na5)} Nd3 {[%eval -183,33] [%emt 0:
01:58]} 38. Rcg2 {[%eval -115,29] [%emt 0:03:06]} Rf7 {[%eval -205,34] [%emt 0:
03:16]} 39. Na5 {[%eval -99,30] [%emt 0:02:18]} Rd7 {[%eval -212,29] [%emt 0:
00:26] (Re8)} 40. Rf1 {[%eval -111,28] [%emt 0:02:14] (Qd2)} Qf6 {[%eval -225,
32] [%emt 0:02:15]} 41. Qa3 {[%eval -133,29] [%emt 0:02:32] (Nc6)} Rf8 {[%eval
-247,33] [%emt 0:01:43] (Qd4)} 42. Qxa4 {[%eval -184,30] [%emt 0:03:41] (Nc6)}
Ra7 {[%eval -297,35] [%emt 0:01:38]} 43. Qb5 {[%eval -198,30] [%emt 0:00:50]}
Rfa8 {[%eval -314,32] [%emt 0:00:52]} 44. Nc4 {[%eval -198,31] [%emt 0:00:44]}
Qd4 {[%eval -325,36] [%emt 0:02:31]} 45. Rd1 {[%eval -213,32] [%emt 0:02:59]}
Rc7 {[%eval -380,35] [%emt 0:01:35]} 46. Nxd6 {[%eval -244,33] [%emt 0:02:07]}
Qa7 {[%eval -418,32] [%emt 0:00:43]} 47. a3 {[%eval -248,33] [%emt 0:00:51]}
Rb8 {[%eval -482,34] [%emt 0:02:10]} 48. Rh1 {[%eval -267,35] [%emt 0:02:16]}
Rxb5 {[%eval -491,33] [%emt 0:00:39]} 49. Nxb5 {[%eval -267,34] [%emt 0:00:46]}
Qb6 {[%eval -505,35] [%emt 0:01:28] (Qc5)} 50. Nxc7 {[%eval -270,33] [%emt 0:
00:25]} Qxc7 {[%eval -533,32] [%emt 0:01:09]} 51. d6 {[%eval -290,35] [%emt 0:
02:47]} Qb7 {[%eval -537,29] [%emt 0:00:31]} 52. Rc2 {[%eval -299,32] [%emt 0:
01:19]} e3 {[%eval -556,32] [%emt 0:02:20]} 53. Rh3 {[%eval -299,33] [%emt 0:
01:08]} Nxf4 {[%eval -574,30] [%emt 0:02:01]} 54. Rxe3 {[%eval -305,32] [%emt
0:00:17]} Qd5 {[%eval -574,25] [%emt 0:00:26] (Qh1+)} 55. Rh2 {[%eval -305,27]
[%emt 0:00:47] (Rc1)} Ne6 {[%eval -602,27] [%emt 0:01:16]} 56. Rc3 {[%eval
-331,30] [%emt 0:02:44] (b4)} Qxd6 {[%eval -637,28] [%emt 0:01:38]} 57. Rh1 {
[%eval -380,27] [%emt 0:02:16] (Rc8+)} f4 {[%eval -682,26] [%emt 0:01:02]} 58.
Rc8+ {[%eval -403,29] [%emt 0:00:43]} Nf8 {[%eval -727,31] [%emt 0:01:21]} 59.
Ka1 {[%eval -432,28] [%emt 0:01:12]} g5 {[%eval -824,30] [%emt 0:01:39] (Qf6)}
60. Rhc1 {[%eval -377,27] [%emt 0:00:13]} Kxh7 {[%eval -806,27] [%emt 0:00:28]}
61. R8c6 {[%eval -432,28] [%emt 0:00:35]} Qd3 {[%eval -933,29] [%emt 0:01:38]}
62. R1c5 {[%eval -451,29] [%emt 0:00:49] (R1c3)} f3 {[%eval -1321,27] [%emt 0:
01:08] (Qe4)} 63. Rf6 {[%eval -2984,27] [%emt 0:00:56] (Rc1)} Nd7 {[%eval
-12842,44] [%emt 0:01:31]} 64. Rc7 {[%eval -25000,35] [%emt 0:00:06]} Kg7 {
[%eval -12847,44] [%emt 0:00:46] (Kg8)} 65. Rxf3 {[%eval -25000,36] [%emt 0:00:
27]} Qd1+ {[%eval -12848,53] [%emt 0:01:26]} 66. Ka2 {[%eval -25000,35] [%emt
0:00:13]} Qxf3 {[%eval -12849,54] [%emt 0:00:28]} 67. a4 {[%eval -25000,35]
[%emt 0:00:21] (Rxd7+)} Qd5+ {[%eval -32737,44] [%emt 0:00:55]} 68. Ka1 {
[%eval -25000,36] [%emt 0:01:20] (Kb1)} g4 {[%eval -32739,56] [%emt 0:01:23]
(Qd1+)} 69. Rc3 {[%eval -25000,26] [%emt 0:00:06] (Kb1)} Ne5 {[%eval -32743,61]
[%emt 0:01:27] (Qe5)} 70. Kb1 {[%eval -32637,29] [%emt 0:00:57] (Rg3)} Nd3 {
[%eval -32747,68] [%emt 0:01:17]} 71. Rc7+ {[%eval -32749,23] [%emt 0:00:01]
(Ra3)} Kf6 {[%eval -32749,80] [%emt 0:01:22] (Kg6)} 72. Kc2 {[%eval -32751,23]
[%emt 0:00:02]} g3 {[%eval -32751,84] [%emt 0:00:26]} 73. Rc3 {[%eval -32753,
29] [%emt 0:00:02]} Nb4+ {[%eval -32753,127] [%emt 0:01:10]} 74. Kc1 {[%eval
-32755,38] [%emt 0:00:03]} g2 {[%eval -32755,127] [%emt 0:00:01]} 75. Rg3 {
[%eval -32757,95] [%emt 0:00:03]} Qe5 {[%eval -32757,127] [%emt 0:00:17]} 76.
b3 {[%eval -32759,99] [%emt 0:00:00] (Rf3+)} Qxg3 {[%eval -32759,127] [%emt 0:
00:00]} 77. Kb2 {[%eval -32761,99] [%emt 0:00:00]} Qe5+ {[%eval -32761,127]
[%emt 0:00:00]} 78. Ka3 {[%eval -32763,99] [%emt 0:00:00]} Nc2+ {[%eval -32763,
127] [%emt 0:00:00]} 79. Ka2 {[%eval -32765,5] [%emt 0:00:00]} Qa1# {[%eval
-32765,127] [%emt 0:00:00]} 0-1
Parent - - By MarshallArts (***) Date 2016-09-30 22:59

> an "idea" of using Black's "a" pawn to form a pawn shelter and in the process stymied Komodo's Queen side attack leaving Komodo at a loss of how to proceed


The Stockfish team added code that helps SF identify enemy a/h-pawns as part of its own king shelter.
This was done a couple of years ago, following a suggestion by Ljudmil Tzvetkov over at talkchess.

:fat:
Parent - By Peter Grayson (****) Date 2016-10-01 11:04

> The Stockfish team added code that helps SF identify enemy a/h-pawns as part of its own king shelter.
> This was done a couple of years ago, following a suggestion by Ljudmil Tzvetkov over at talkchess.


I try to look through the games produced from my matches to identify anything odd, unusual or particularly interesting but must admit that if this has been coded into Stockfish for some time then it has gone unnoticed by me. It certainly worked well in the two games I posted and it is encouraging to find the engine has this degree of sophistication coded into it.

PeterG
Parent - - By MarshallArts (***) Date 2016-10-01 05:30
Furthermore, here's a recently concluded LTC game that has the latest Komodo losing to Stockfish in some 29 moves,

http://www.talkchess.com/forum/viewtopic.php?topic_view=threads&p=687906&t=61516

:eek:
Parent - By Peter Grayson (****) Date 2016-10-01 10:56

> Furthermore, here's a recently concluded LTC game that has the latest Komodo losing to Stockfish in some 29 moves,


> http://www.talkchess.com/forum/viewtopic.php?topic_view=threads&p=687906&t=61516


Fascinating to see how from the first moves that looked guarded to say the least, by move 10..Re8 Stockfish already looked better when the immediate visual impression was that Stockfish controlled the important King side space and the position looked dynamic whereas although Komodo had Queen side space it looked cramped.

PeterG
Parent - - By Venator (Silver) Date 2016-10-01 06:58
Remarkable. One could say that a black pawn on a2 is protecting the white king even better than a white pawn on a2. A white pawn on a2 can be attacked by black, while he cannot take his own pawn on a2.
Parent - - By Peter Grayson (****) Date 2016-10-01 11:15
The strange thing is that on my machines, whether or not Stockfish will play that move 17-19 sequence for some reason depends on the hash level it is given.

When I analysed the game on the same machine but with 4 Gb hash, the same engine would not consider anything other than maintaining the material balance by 17.bxa. Similarly so on my dual Xeon machine when the POPCNT engine would play the game moves with 8 Gb and 16 Gb hash but again not with 4 Gb hash.

Must also mention that with a different hash setting Komodo also prefers 12..Nh5 rather than 12..b4, so there would seem to be a number of influences determining what moves an engine will play not excluding the multi-core variability.

Peter
Parent - - By gsgs (***) Date 2016-10-03 15:54
isn't it just random ? I get different moves with the same settings
and cleared hash
Parent - - By Peter Grayson (****) Date 2016-10-04 11:34

> isn't it just random ? I get different moves with the same settings
> and cleared hash


When analysing a game it is important to try and reproduce the hash content for the position in the game to be analysed. With cleared hash I use the engine in infinite analysis from the out of line position and click on opponent's move at the appropriate ply depth completion. Hash will not be exactly the same but fairly close. Analysing the same position with clean hash can often give different results because the analysis conditions are different to that in the game especially if, as now seems the case, the engine can follow a particular plan for several moves until completion.

I would recommend analysing from at least 5 moves back from the position to be analysed. Number of threads and hash can also influence output because it can influence the time spent analysing a particular branch of the position.

I conducted much testing of hash influence in single core mode and it was surprising how changing hash influenced the engine's output when running through tests sets. e.g. the same engine with 512 Mg and 1Gb hash the output could look as though it were from different engines despite starting each test run with clean hash. The GUI can affect output too with something as simple as whether or not it supports sampled search in the output display.

PeterG
Parent - By gsgs (***) Date 2016-10-08 16:24
I usually do this with droidfish on a smartphone.
I enter the position in analysis-mode, then I touch the bottom right button in Android ("back")
and it exits droidfish with stockfish with cleared hash (I suppose) ,
and goes to the main smartphone menu.
Then I click the droidfish icon again, it goes to the position and starts immedialy analyzing
apparently with clear hash.
The critical, known, good move in test-positions can be found earlier or later - unpredictable.
Later runs needn't be better than earlier runs.
Parent - - By beram (*) Date 2016-10-01 10:58
Here is another example of a game where SF evaluates much better a white h-pawn passer blocked by black king
(see my original post at http://www.talkchess.com/forum/viewtopic.php?topic_view=threads&p=665324&t=59563)

[Event "SF 020316 - Komodo 9.42(ct=0), Rapid 15"]
[Site "AMD X6 1090T (Fritzmark 23,1)"]
[Date "2016.03.21"]
[Round "4"]
[White "Komodo 9.42 64-bit"]
[Black "Stockfish 020316 64"]
[Result "0-1"]
[ECO "B90"]
[Annotator "0.55;0.37"]
[PlyCount "140"]
[TimeControl "900+10"]

{AMD Phenom(tm) II X6 1090T Processor 3208 MHz  W=26.0 plies; 5.239kN/s  B=28.

7 plies; 5.711kN/s} 1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3

e5 7. Nb3 Be7 8. Qd2 O-O 9. f3 Be6 10. O-O-O Qc7 11. g4 Rc8 {Beide laatste zet

uit boek} 12. g5 {

[%eval 55,26]
[%emt 0:00:35]

} Nh5 {

[%eval 37,29]
[%emt 0:00:

40]

} 13. Qf2 {

[%eval 44,27]
[%emt 0:01:40]

(Kb1)} Qd8 {

[%eval 15,25]
[%emt 0:

00:17]

} 14. Qd2 {

[%eval 53,27]
[%emt 0:00:37]

(h4)} Qc7 {

[%eval 0,32]
[%emt 0:

00:16]

(Pd7)} 15. Kb1 {

[%eval 58,26]
[%emt 0:00:15]

} Nd7 {

[%eval 39,29]
[%emt

0:00:45]

} 16. Rg1 {

[%eval 50,28]
[%emt 0:00:50]

} g6 {

[%eval 33,28]
[%emt 0:00:

40]

} 17. h4 {

[%eval 46,27]
[%emt 0:00:50]

} b5 {

[%eval 31,29]
[%emt 0:00:22]

}

18. Nd5 {

[%eval 52,28]
[%emt 0:00:23]

} Bxd5 {

[%eval 29,28]
[%emt 0:00:17]

} 19.

exd5 {

[%eval 39,27]
[%emt 0:00:14]

} Nb6 {

[%eval 29,29]
[%emt 0:00:41]

} 20. Na5

{

[%eval 31,28]
[%emt 0:00:25]

} Bf8 {

[%eval 31,26]
[%emt 0:00:24]

} 21. Bh3 {

[%eval 42,28]
[%emt 0:00:46]

(c3)} Re8 {

[%eval 16,28]
[%emt 0:00:39]

(Tcb8)}

22. Bxb6 {

[%eval 41,27]
[%emt 0:00:51]

} Qxb6 {

[%eval 31,27]
[%emt 0:00:14]

} 23.

Nc6 {

[%eval 26,24]
[%emt 0:00:16]

} Qc7 {

[%eval 16,30]
[%emt 0:00:28]

(Db7)} 24.

Rh1 {

[%eval 45,25]
[%emt 0:00:23]

(Lg4)} Nf4 {

[%eval 19,33]
[%emt 0:00:39]

(Le7)} 25. Rde1 {

[%eval 41,26]
[%emt 0:00:48]

(Lg4)} h5 {

[%eval 0,26]
[%emt 0:

00:25]

} 26. gxh6 {

[%eval 37,26]
[%emt 0:01:01]

(Lf1)} Kh7 {

[%eval -4,29]
[%emt

0:00:18]

} 27. h5 {

[%eval 27,26]
[%emt 0:00:38]

(Teg1)} Nxh5 {

[%eval -27,29]
[%emt 0:00:21]

} 28. Bf1 {

[%eval 22,27]
[%emt 0:00:56]

} Nf4 {

[%eval -37,27]
[%emt 0:00:14]

} 29. c4 {

[%eval 25,29]
[%emt 0:00:37]

(Te3)} bxc4 {

[%eval -48,

29]
[%emt 0:01:13]

(Db6)} 30. Bxc4 {

[%eval 33,26]
[%emt 0:00:17]

} Qc8 {

[%eval

-39,32]
[%emt 0:01:02]

(a5)} 31. Bb3 {

[%eval 36,23]
[%emt 0:00:22]

} Be7 {

[%eval -43,31]
[%emt 0:00:44]

(Dd7)} 32. Ka1 {

[%eval 47,24]
[%emt 0:01:03]

(Tc1) almost one pawn disagreement in evaluation. Komodo thinks white is +0.47

pawn better while SF thinks black is better by -0.44  The greatest difference

probably must be found in the evaluation of the white blocked but free pawn on

h6. Black can create a pawnfront in centre while h6 goes nowhere. I have seen

this same pawn blocked pawn on h6 theme more in games where Komodo evaluates

this to positively. Point is that black can play alongside this pawn and white

cannot create sufficient attacking threats to compensate for that and cannot

progress the pawn any further neither} Qc7 {

[%eval -44,30]
[%emt 0:00:49]

(Lf8)

} 33. Bc2 {

[%eval 31,23]
[%emt 0:01:10]

(Tc1)} Bg5 {

[%eval -48,27]
[%emt 0:00:

34]

(Db6)} 34. Reg1 {

[%eval 19,24]
[%emt 0:00:39]

(Te4)} Bf6 {

[%eval -42,28]
[%emt 0:00:38]

} 35. Rg4 {

[%eval 22,25]
[%emt 0:01:06]

(Tc1)} a5 {

[%eval -81,27]
[%emt 0:00:20]

} 36. Kb1 {

[%eval 0,28]
[%emt 0:00:30]

(La4)} a4 {

[%eval -97,28]
[%emt 0:00:19]

} 37. Bd3 {

[%eval -7,30]
[%emt 0:01:17]

} Bh8 {

[%eval -91,26]
[%emt 0:00:05]

(a3)} 38. a3 {

[%eval -43,25]
[%emt 0:00:47]

(Te1)} f5 {

[%eval

-142,27]
[%emt 0:00:34]

here comes the flood} 39. Rgh4 {

[%eval -60,26]
[%emt 0:

00:26]

(Tgg1)} Bf6 {

[%eval -167,28]
[%emt 0:00:33]

(Db6)} 40. R4h2 {

[%eval -64,

25]
[%emt 0:00:20]

} Nxd3 {

[%eval -176,26]
[%emt 0:00:07]

} 41. Qxd3 {

[%eval -59,

24]
[%emt 0:00:05]

} e4 {

[%eval -189,27]
[%emt 0:00:21]

} 42. fxe4 {

[%eval -91,

23]
[%emt 0:00:23]

(De3)} Rxe4 {

[%eval -194,27]
[%emt 0:00:20]

and all becomes

clear here black bishop dominates and black has strong free pawn pair and h6

goes nowhere} 43. Qb5 {

[%eval -94,23]
[%emt 0:00:22]

(Tc1)} f4 {

[%eval -218,27]
[%emt 0:01:24]

(Tae8)} 44. Rc2 {

[%eval -107,23]
[%emt 0:00:26]

} Qf7 {

[%eval

-243,27]
[%emt 0:00:38]

} 45. Ka2 {

[%eval -116,24]
[%emt 0:00:24]

(Tf1)} Rae8 {

[%eval -244,26]
[%emt 0:00:15]

} 46. Rf1 {

[%eval -127,25]
[%emt 0:00:19]

} g5 {

[%eval -247,24]
[%emt 0:00:12]

(Lh4)} 47. Nb4 {

[%eval -104,20]
[%emt 0:00:14]

}

R8e7 {

[%eval -248,26]
[%emt 0:00:32]

(Kxh6)} 48. Qxa4 {

[%eval -92,22]
[%emt 0:

00:27]

(Tcf2)} g4 {

[%eval -309,24]
[%emt 0:00:12]

} 49. Qc6 {

[%eval -102,21]
[%emt 0:00:06]

} f3 {

[%eval -325,24]
[%emt 0:00:06]

(Le5)} 50. Qxd6 {

[%eval

-155,22]
[%emt 0:00:15]

} Bh4 {

[%eval -370,26]
[%emt 0:00:15]

(Te2)} 51. Rh2 {

[%eval -154,23]
[%emt 0:00:12]

(Tc8)} g3 {

[%eval -525,28]
[%emt 0:00:15]

} 52.

Rxh4 {

[%eval -193,23]
[%emt 0:00:06]

} g2 {

[%eval -539,30]
[%emt 0:00:08]

} 53.

Rxe4 {

[%eval -204,26]
[%emt 0:00:05]

} Rxe4 {

[%eval -565,30]
[%emt 0:00:16]

} 54.

Rg1 {

[%eval -227,27]
[%emt 0:00:09]

} f2 {

[%eval -569,30]
[%emt 0:00:04]

} 55.

Rxg2 {

[%eval -242,27]
[%emt 0:00:04]

} f1=Q {

[%eval -576,33]
[%emt 0:00:17]

} 56.

Qh2 {

[%eval -257,28]
[%emt 0:00:20]

(Tc2)} Q7f4 {

[%eval -589,31]
[%emt 0:00:12]

} 57. Rg7+ {

[%eval -266,28]
[%emt 0:00:06]

} Kh8 {

[%eval -589,1]
[%emt 0:00:00]

}

58. Qc2 {

[%eval -266,30]
[%emt 0:00:13]

} Qc4+ {

[%eval -600,32]
[%emt 0:00:17]

}

59. Qxc4 {

[%eval -276,27]
[%emt 0:00:03]

} Rxc4 {

[%eval -612,28]
[%emt 0:00:06]

}

60. Rg6 {

[%eval -279,31]
[%emt 0:00:10]

(Tg1)} Rd4 {

[%eval -653,33]
[%emt 0:01:

05]

(Te4)} 61. Rg2 {

[%eval -279,33]
[%emt 0:00:14]

(Tc6)} Qf1 {

[%eval -692,31]
[%emt 0:00:23]

(Te4)} 62. Rc2 {

[%eval -298,27]
[%emt 0:00:13]

} Kh7 {

[%eval

-692,32]
[%emt 0:00:05]

} 63. Kb3 {

[%eval -311,26]
[%emt 0:00:15]

(Tc6)} Qd1 {

[%eval -778,30]
[%emt 0:00:40]

(Kxh6)} 64. Ka2 {

[%eval -313,27]
[%emt 0:00:06]

}

Kxh6 {

[%eval -792,30]
[%emt 0:00:16]

} 65. Rc6+ {

[%eval -337,27]
[%emt 0:00:08]

}

Kg5 {

[%eval -817,28]
[%emt 0:00:05]

} 66. Rc5 {

[%eval -338,29]
[%emt 0:00:05]

(d6)} Rd2 {

[%eval -858,32]
[%emt 0:00:15]

} 67. d6+ {

[%eval -338,27]
[%emt 0:00:

04]

} Kf6 {

[%eval -897,27]
[%emt 0:00:05]

(Kh6)} 68. Rc8 {

[%eval -693,27]
[%emt

0:00:33]

(d7)} Qe2 {

[%eval -4013,28]
[%emt 0:00:12]

} 69. d7 {

[%eval -755,27]
[%emt 0:00:07]

} Qe6+ {

[%eval -4312,32]
[%emt 0:00:04]

} 70. Kb1 {

[%eval -1094,

28]
[%emt 0:00:04]

} Qxd7 {

[%eval -4425,36]
[%emt 0:00:13]

} 0-1
Parent - - By Peter Grayson (****) Date 2016-10-01 12:50
Thanks for games link.

In the first game, Black's ..Nh5 seems key to prevent White the opportunity of f4. That was the deviation point in the second game I posted. In your game the King sheltered behind the "h" pawn and once again it was very effective in stifling Komodo's attack. One basic chess rule that stood out was Pawn Islands where generally less is better and so it proved here.

An observation in the second game of your Talkchess post was the often seen when it loses 0.00 scoring of the position by Stockfish from 20..Rce8 through to 27..Be3 when you also commented on the 0.00 score. The jump to adverse 0.86 for move 28..Qg4 suggests a blind spot when it is too late by the time it identifies the position is not 0.00. I cannot recall if anything has been posted on the reasons for this but have recorded it can occur in the Komodo engine's scoring during a game too.

PeterG
Parent - By MarshallArts (***) Date 2016-10-02 02:10
I'm quite surprised that Komodo appears to lack any "edge-pawn-blocked-by-king as part of king shelter" code that's known to work so well in SF, especially with Larry on K's team.

The 0.00 score can be unreliable and doesn't always mean sure equality if the position is complex. In such cases 0.00 should be more in line to the unclear/infinity Informant symbol.

:neutral:
Parent - By beram (*) Date 2017-05-25 08:30
Edge pawnshelter theme with double enemy pawns continues
A very nice recent example arose at move 27
Komodo also cannot find best move 25 ..Kh8! it hangs on the only other possible move 25 ..Kh7 even after 44ply...
25...Kg8xh7 26.Td1-d3 d6-d5 27.e4xd5 f7-f5 28.Pc1xb3 Db7-b5 29.Le3-d2 Ld8-e7 30.f3-f4 Le7xa3 31.Ld2-c1 Pc4xb2 32.Lc1xb2 La3xb2 33.Kb1xb2 Pe8-d6 34.f4xe5 Pd6-c4+ 35.Kb2-b1 Pc4-a3+ 36.Kb1-b2 Pa3-c4+
  =  (0.00)   Diepte: 44   00:50:10  50101MN

[Event "pedantFish240417-K104 AMD Ryzen1800X 14"]
[Site "AMD Ryzen 1800X"]
[Date "2017.05.05"]
[Round "50"]
[White "Komodo 10.4 64-bit"]
[Black "pedantFishW_2017-04-26_popcnt"]
[Result "0-1"]
[ECO "B90"]
[Annotator "0.36;0.18"]
[PlyCount "158"]
[TimeControl "180+2"]

{AMD Ryzen 7 1800X Eight-Core Processor          3593 MHz  W=23.7 plies; 14.
872kN/s  B=24.7 plies; 20.081kN/s} 1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5.
Nc3 a6 6. Be3 e5 7. Nb3 Be6 8. f3 Be7 9. Qd2 O-O 10. O-O-O Nbd7 11. g4 b5 {
Beide laatste zet uit boek} 12. g5 {[%eval 36,26] [%emt 0:00:16]} b4 {[%eval
18,25] [%emt 0:00:10]} 13. Ne2 {[%eval 41,23] [%emt 0:00:03]} Ne8 {[%eval 23,
22] [%emt 0:00:01]} 14. Kb1 {[%eval 21,24] [%emt 0:00:13] (f3-f4)} a5 {[%eval
22,23] [%emt 0:00:07]} 15. h4 {[%eval 40,24] [%emt 0:00:16]} Qb8 {[%eval 18,24]
[%emt 0:00:14] (a5-a4)} 16. Ng3 {[%eval 46,24] [%emt 0:00:06]} a4 {[%eval 5,24]
[%emt 0:00:05] (Le7-d8)} 17. Nc1 {[%eval 49,24] [%emt 0:00:06]} Bd8 {[%eval 30,
24] [%emt 0:00:10]} 18. h5 {[%eval 49,23] [%emt 0:00:04] (Lf1-h3)} b3 {[%eval
41,23] [%emt 0:00:09]} 19. cxb3 {[%eval 45,22] [%emt 0:00:05]} axb3 {[%eval 16,
23] [%emt 0:00:08]} 20. a3 {[%eval 55,21] [%emt 0:00:03]} Qb7 {[%eval 0,24]
[%emt 0:00:08] (Pd7-b6)} 21. Bh3 {[%eval 41,23] [%emt 0:00:08]} Nb6 {[%eval 4,
22] [%emt 0:00:02]} 22. g6 {[%eval 41,21] [%emt 0:00:09] (Le3xb6)} Nc4 {[%eval
-17,25] [%emt 0:00:11] (Pb6-a4)} 23. Qe2 {[%eval 47,22] [%emt 0:00:05]} Bxh3 {
[%eval 0,25] [%emt 0:00:03]} 24. Rxh3 {[%eval 44,22] [%emt 0:00:04]} Rc8 {
[%eval 0,29] [%emt 0:00:06]} 25. gxh7+ {[%eval 52,23] [%emt 0:00:11]} Kh8 {
[%eval 0,28] [%emt 0:00:01] (Kg8xh7)} 26. Rhh1 {[%eval 59,23] [%emt 0:00:19]}
d5 {[%eval -35,28] [%emt 0:00:31] (Db7-b5)} 27. h6 {[%eval 0,25] [%emt 0:00:11]
(Td1-d3)} g6 {[%eval -104,22] [%emt 0:00:04]} 28. exd5 {[%eval 0,25] [%emt 0:
00:06]} Be7 {[%eval -76,20] [%emt 0:00:02]} 29. Rd3 {[%eval -14,22] [%emt 0:00:
03]} f5 {[%eval -102,21] [%emt 0:00:03]} 30. Rxb3 {[%eval -4,24] [%emt 0:00:19]
} Qxd5 {[%eval -90,20] [%emt 0:00:02]} 31. Rd1 {[%eval -14,25] [%emt 0:00:09]}
Qe6 {[%eval -72,24] [%emt 0:00:13]} 32. Bd4 {[%eval -10,26] [%emt 0:00:10]} f4
{[%eval -86,22] [%emt 0:00:05]} 33. Nf1 {[%eval -6,24] [%emt 0:00:06] (Pg3-h1)}
Kxh7 {[%eval -66,24] [%emt 0:00:10] (Pe8-f6)} 34. Bc3 {[%eval -5,25] [%emt 0:
00:07]} Nf6 {[%eval -56,25] [%emt 0:00:07]} 35. Nd2 {[%eval 0,25] [%emt 0:00:
03]} Ne3 {[%eval -24,26] [%emt 0:00:09]} 36. Rg1 {[%eval -22,24] [%emt 0:00:05]
} Nfd5 {[%eval -35,24] [%emt 0:00:01]} 37. Na2 {[%eval -7,26] [%emt 0:00:10]
(Pd2-e4)} Rb8 {[%eval -74,25] [%emt 0:00:06]} 38. Nb4 {[%eval -1,26] [%emt 0:
00:05]} Nf6 {[%eval -53,22] [%emt 0:00:01] (Pd5xb4)} 39. Qd3 {[%eval -36,24]
[%emt 0:00:04]} Rg8 {[%eval -59,26] [%emt 0:00:04] (Pf6-d7)} 40. Bxe5 {[%eval
-58,24] [%emt 0:00:10] (Kb1-a2)} Rbd8 {[%eval -141,25] [%emt 0:00:05]} 41. Bd4
{[%eval -96,24] [%emt 0:00:06]} Nf5 {[%eval -122,28] [%emt 0:00:10]} 42. Nc2 {
[%eval -93,25] [%emt 0:00:02]} Nxd4 {[%eval -134,25] [%emt 0:00:01]} 43. Nxd4 {
[%eval -86,25] [%emt 0:00:02]} Qd5 {[%eval -131,29] [%emt 0:00:10]} 44. Qc4 {
[%eval -90,24] [%emt 0:00:03]} Qxd4 {[%eval -103,26] [%emt 0:00:01]} 45. Qf7+ {
[%eval -92,23] [%emt 0:00:02]} Kxh6 {[%eval -112,26] [%emt 0:00:03]} 46. Rh1+ {
[%eval -87,23] [%emt 0:00:02]} Nh5 {[%eval -111,24] [%emt 0:00:01]} 47. Ne4 {
[%eval -84,25] [%emt 0:00:05]} Rge8 {[%eval -117,27] [%emt 0:00:03]} 48. Rc3 {
[%eval -89,22] [%emt 0:00:01] (Df7xf4+)} Qd5 {[%eval -139,23] [%emt 0:00:02]}
49. Qxf4+ {[%eval -96,22] [%emt 0:00:02]} Kg7 {[%eval -133,23] [%emt 0:00:01]}
50. Qc1 {[%eval -95,23] [%emt 0:00:04] (Df4-e3)} Rd7 {[%eval -120,25] [%emt 0:
00:06]} 51. Qc2 {[%eval -95,22] [%emt 0:00:03] (Th1-g1)} Red8 {[%eval -139,24]
[%emt 0:00:02] (Te8-f8)} 52. Rc6 {[%eval -94,21] [%emt 0:00:02]} Qf5 {[%eval
-145,23] [%emt 0:00:06]} 53. Rc1 {[%eval -105,23] [%emt 0:00:08]} Nf4 {[%eval
-131,24] [%emt 0:00:06] (Td7-d3)} 54. Qc3+ {[%eval -92,21] [%emt 0:00:03]} Kf7
{[%eval -142,27] [%emt 0:00:14] (Td7-d4)} 55. Rb6 {[%eval -99,20] [%emt 0:00:
01]} Ne6 {[%eval -164,28] [%emt 0:00:06]} 56. Re1 {[%eval -115,21] [%emt 0:00:
01] (Tc1-h1)} Rd1+ {[%eval -169,25] [%emt 0:00:04]} 57. Rxd1 {[%eval -126,22]
[%emt 0:00:01]} Rxd1+ {[%eval -178,20] [%emt 0:00:01]} 58. Ka2 {[%eval -118,25]
[%emt 0:00:03]} Qd5+ {[%eval -167,26] [%emt 0:00:04]} 59. Qb3 {[%eval -137,24]
[%emt 0:00:01]} g5 {[%eval -178,23] [%emt 0:00:01] (Dd5xb3+)} 60. Nf2 {[%eval
-133,22] [%emt 0:00:02] (Db3xd5)} Nc5 {[%eval -200,25] [%emt 0:00:02] (Dd5xb3+)
} 61. Qxd5+ {[%eval -110,19] [%emt 0:00:01]} Rxd5 {[%eval -202,22] [%emt 0:00:
01]} 62. Ng4 {[%eval -147,21] [%emt 0:00:02] (Tb6-c6)} Bd6 {[%eval -218,24]
[%emt 0:00:03]} 63. Ne3 {[%eval -137,23] [%emt 0:00:02]} Rd3 {[%eval -204,25]
[%emt 0:00:06]} 64. Nc4 {[%eval -148,23] [%emt 0:00:01]} Bh2 {[%eval -252,24]
[%emt 0:00:00]} 65. a4 {[%eval -140,23] [%emt 0:00:01] (Tb6-b5)} Rxf3 {[%eval
-281,24] [%emt 0:00:04] (Td3-d4)} 66. a5 {[%eval -133,22] [%emt 0:00:01]} Nd3 {
[%eval -293,22] [%emt 0:00:01] (Tf3-f6)} 67. Kb3 {[%eval -187,20] [%emt 0:00:
04] (a5-a6)} Bc7 {[%eval -347,22] [%emt 0:00:03]} 68. Rb5 {[%eval -221,21]
[%emt 0:00:01]} g4 {[%eval -411,23] [%emt 0:00:03]} 69. Nd2 {[%eval -246,23]
[%emt 0:00:01]} Re3 {[%eval -362,21] [%emt 0:00:01]} 70. Nc4 {[%eval -246,23]
[%emt 0:00:02]} Rh3 {[%eval -431,22] [%emt 0:00:02] (Te3-e8)} 71. Kc2 {[%eval
-246,23] [%emt 0:00:03] (a5-a6)} g3 {[%eval -512,23] [%emt 0:00:02]} 72. Ne3 {
[%eval -265,23] [%emt 0:00:02]} Ke6 {[%eval -543,23] [%emt 0:00:01] (Pd3-e1+)}
73. Kxd3 {[%eval -299,23] [%emt 0:00:05]} g2 {[%eval -586,27] [%emt 0:00:01]}
74. Rg5 {[%eval -321,26] [%emt 0:00:01]} Bg3 {[%eval -610,26] [%emt 0:00:01]}
75. Rg6+ {[%eval -332,28] [%emt 0:00:02] (Pe3xg2)} Kf7 {[%eval -660,30] [%emt
0:00:03]} 76. Nxg2 {[%eval -359,29] [%emt 0:00:02] (Tg6-g7+)} Kxg6 {[%eval
-671,34] [%emt 0:00:02]} 77. b4 {[%eval -662,29] [%emt 0:00:04] (Kd3-e4)} Bd6+
{[%eval -822,27] [%emt 0:00:02]} 78. Kc4 {[%eval -787,32] [%emt 0:00:02]} Rg3 {
[%eval -4562,25] [%emt 0:00:01]} 79. Ne1 {[%eval -835,33] [%emt 0:00:02]} Rg4+
{[%eval -4562,41] [%emt 0:00:02]} 0-1
- By Lazy Frank (****) Date 2016-11-08 18:08
Today i got one more nice example of king shelter theme. Additionally manual long castling king walk e1->b1.

Here is game:


[Site "Raivis"]
[White "Stockfish"]
[Black "Komodo"]
[Result "1-0"]
[ECO "B31"]
[Opening "Sicilian"]
[Variation "Rossolimo, 3...g6 4.Bxc6 dxc6 5.d3 Bg7 6.h3"]
[Termination "normal"]
[PlyCount "161"]
[WhiteType "engine"]
[BlackType "engine"]

1. e4 c5 2. Nf3 Nc6 3. Bb5 g6 4. Bxc6 dxc6 5. d3 Bg7 6. h3 Nf6 7. Nc3 O-O
8. Be3 b6 9. Qd2 e5 10. Bh6 Qd6 11. g4 a5 12. Ne2 Ne8 13. Qe3 a4 14. Nd2
Qe7 15. a3 b5 16. Bxg7 Nxg7 17. f4 exf4 18. Qxf4 Be6 19. Rf1 f6 20. Qg3 b4
21. Rf2 Qb7 22. Qd6 Qb6 23. d4 bxa3 24. Rxa3 Qxb2 25. Qxc5 Qb5 26. Qc3 Bd7
27. Nf4 Rae8 28. Kd1 Ne6 29. Nxe6 Bxe6 30. Rf1 Ra8 31. Kc1 Kg7 32. g5 Qxg5
33. Qxc6 Bb3 34. Qb7+ Bf7 35. d5 Qe5 36. Qb2 Qe7 37. Raf3 a3 38. Qd4 Ra6
39. Kb1 Rb8+ 40. Rb3 Rxb3+ 41. Nxb3 Qd7 42. Rf3 a2+ 43. Ka1 Qc7 44. c4 Qh2
45. Nd2 Be8 46. Qc3 Bd7 47. c5 Qe2 48. Rf1 Bxh3 49. Rc1 Qh2 50. c6 Ra8 51.
c7 h5 52. e5 Qxe5 53. Qxe5 fxe5 54. d6 Ra5 55. Ne4 Rd5 56. Rc5 Rd1+ 57.
Kxa2 Bc8 58. Rxe5 Kf8 59. Kb2 h4 60. Ra5 Kf7 61. Ra8 Bd7 62. Rd8 Bf5 63. d7
Rxd7 64. c8=Q Rd2+ 65. Nxd2 Bxc8 66. Rxc8 g5 67. Kc3 Ke6 68. Kd4 g4 69. Rh8
Kf5 70. Rxh4 Kg5 71. Rh2 Kf6 72. Kd5 Kg5 73. Ke5 g3 74. Rg2 Kg4 75. Ne4 Kf3
76. Rxg3+ Ke2 77. Kd4 Kd1 78. Kd3 Kc1 79. Rg2 Kb1 80. Nc3+ Ka1 81. Ra2# 1-0
Up Topic The Rybka Lounge / Computer Chess / Remarkable miniature in LTC game Stockfish vs Komodo

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill