Rxd7! Looks like a nice positional sac.

[Laskos]

But what would interest me, is from BEFORE Rxd7.

Stuart, here is the game 40/40 before Rxd7, Houdini 3 vs Houdini 3, white won in a somehow methodical fashion. So, the sack seems a bad idea.

[Event "40m/40+40m/40+"]
[Site "PC"]
[Date "2012.11.13"]
[Round "1"]
[White "Houdini 3 Pro x64"]
[Black "Houdini 3 Pro x64"]
[Result "1-0"]
[Annotator "0.12;0.11"]
[SetUp "1"]
[FEN "r2q1r1k/3nbppp/p7/3RpP2/Pp2P3/1PN1Q3/1P5P/1K3BR1 w - - 0 19"]
[PlyCount "159"]
[TimeControl "40/2400:40/2400:2400"]

19. Qg3 {0.12/20 60} Rg8 {0.11/23 87} 20. Nb5 {0.20/22 87} Qc8 {0.11/23 59} 21.
Nd6 {0.20/20 0} Bxd6 {0.11/22 0} 22. Rxd6 {0.20/19 7} Qc7 {0.10/24 52} 23. Rd4
{0.19/22 51 (Rxa6)} Nb6 {0.22/21 85 (Qb7)} 24. Be2 {0.33/21 70 (Rd2)} Qe7 {0.
29/21 60} 25. Rd2 {0.48/22 85} Nd7 {0.42/23 66} 26. Rc1 {0.57/23 144} Nf6 {0.
43/23 111} 27. Qe3 {0.57/21 0} a5 {0.41/22 59 (h6)} 28. Rdc2 {0.59/24 83} Qb7 {
0.59/21 114 (Rad8)} 29. Bf3 {0.80/22 87 (Bb5)} Ra7 {0.65/21 43 (Rac8)} 30. Rc6
{0.88/23 85} Re8 {0.84/20 71 (Rd8)} 31. h4 {0.98/22 52} Ng8 {0.91/21 52 (Qd7)}
32. h5 {1.14/22 44} Qe7 {1.22/21 222} 33. Qb6 {1.11/20 48} h6 {1.54/20 108} 34.
Rc8 {1.91/22 66} Ra8 {1.92/21 50} 35. Rxe8 {1.91/20 0} Qxe8 {1.80/22 39} 36.
Rc5 {1.79/23 62} Qd7 {1.76/22 59} 37. Rd5 {1.79/21 0} Qa7 {1.73/24 37} 38. Qxa7
{1.80/25 73} Rxa7 {1.55/23 0} 39. Rxe5 {1.55/24 0} g6 {1.52/25 62} 40. hxg6 {
1.61/26 102} fxg6 {1.51/23 0} 41. Kc2 {1.69/26 190} gxf5 {1.48/25 43 (Kg7)} 42.
Rxf5 {1.66/24 50} Kg7 {1.55/25 45} 43. Kd2 {1.75/26 59 (Kd3)} Ne7 {1.58/24 41}
44. Rb5 {1.78/25 142 (Rc5)} Kf6 {1.64/24 228} 45. Be2 {1.91/26 52 (Ke3)} Ng6 {
1.62/26 33} 46. Rf5+ {1.89/26 59} Kg7 {1.73/27 55} 47. Ke3 {1.95/27 65} Ne7 {
1.82/25 34} 48. Rh5 {1.94/24 59 (Rc5)} Ng6 {1.81/26 36 (Nc6)} 49. Bc4 {1.85/26
112} Ne7 {1.78/24 34} 50. Kf4 {1.89/24 51 (Rc5)} Ng6+ {1.84/23 35} 51. Kg4 {1.
99/24 44} Ne7 {1.91/25 73} 52. Rb5 {2.01/25 54} Nc6 {2.05/23 85 (Nc8)} 53. e5 {
2.04/25 33 (Rc5)} Re7 {2.30/21 45 (Ne7)} 54. Kf4 {2.63/24 78} Ra7 {2.50/23 57}
55. Rc5 {2.69/23 30 (e6)} Ne7 {2.71/21 31} 56. e6 {2.69/21 0} Ra8 {2.80/21 34}
57. Ke5 {2.80/25 87} Ng6+ {2.94/22 40} 58. Kd6 {3.13/21 26} Kf6 {3.05/20 14}
59. Rh5 {3.13/19 0 (Rc7)} Rd8+ {3.19/21 102} 60. Kc7 {3.36/22 150} Rh8 {3.32/
22 239} 61. Kb6 {3.36/20 0} Nf4 {3.46/21 288} 62. Rh2 {3.22/23 376} h5 {3.36/
20 172 (Ng6)} 63. Kxa5 {3.55/19 83} Nxe6 {3.41/20 62 (Rb8)} 64. Kxb4 {3.39/20
42} Ng5 {3.19/19 46 (h4)} 65. a5 {3.81/20 41 (Rh4)} Ke5 {3.93/19 50 (h4)} 66.
a6 {3.93/20 40} h4 {4.34/20 76} 67. Kc5 {3.93/19 0 (Kb5)} Ne4+ {3.95/19 38} 68.
Kc6 {4.00/22 47 (Kb6)} Rh6+ {3.90/20 37} 69. Kc7 {4.00/20 0} Rh7+ {3.90/19 0}
70. Kb6 {4.03/20 9} Nd6 {4.23/21 78} 71. Bd3 {4.08/21 77 (Bf1)} Nc8+ {3.96/19
38} 72. Kc5 {4.08/19 0} Rd7 {4.04/20 36} 73. Rh3 {4.23/22 117} Rc7+ {4.09/21 39
} 74. Kb5 {4.73/22 130 (Kb4)} Kd6 {4.74/20 67 (Kd4)} 75. b4 {5.24/22 138 (Rxh4)
} Rf7 {4.87/19 62} 76. Rxh4 {5.24/20 0} Kc7 {6.53/21 108} 77. Ka5 {8.22/20 36
(Rh8)} Rf2 {6.19/20 29 (Re7)} 78. Rh7+ {9.23/19 13} Kd6 {6.59/23 135} 79. Rh8 {
9.45/18 5 (b3)} Kc7 {8.83/22 108 (Ne7)} 80. Be4 {10.24/17 2 (b5)} Rf7 {13.59/
20 64 (Ne7)} 81. b5 {12.13/16 3} Kd7 {13.27/17 18 (Re7)} 82. b6 {14.43/15 2}
Nxb6 {16.92/18 39} 83. Kxb6 {14.43/14 0} Rf6+ {#14/25 23} 84. Ka5 {14.23/13 0
(Kb5)} Rf1 {#11/27 61} 85. a7 {14.23/12 0 (Bc2)} Ra1+ {#15/27 27} 86. Kb6 {14.
23/11 0} Ke7 {#14/26 33 (Kd6)} 87. a8=Q {14.23/10 0} Rxa8 {#11/27 27} 88. Rxa8
{14.23/9 0 (Bxa8)} Kf6 {#9/27 32} 89. b4 {13.65/9 0 (Ra5)} Ke5 {#9/28 26} 90.
Bc6 {13.65/9 0 (b5)} Kd4 {#8/26 12 (Ke6)} 91. b5 {16.11/9 0} Kd3 {#7/24 1 (Kc4)
} 92. Kc5 {#12/9 0} Ke3 {#6/22 0} 93. b6 {#8/9 0} Kf4 {#5/20 0 (Kf2)} 94. b7 {
#7/9 0} Kf5 {#4/18 0} 95. b8=Q {#6/9 0} Kg6 {#3/16 0 (Kg4)} 96. Qf4 {#3/15 0}
Kg7 {#2/14 0} 97. Qg5+ {#2/13 0 (Ra7+)} Kf7 {#1/12 0 (Kh7)} 98. Bd5# {#1/11 0}
1-0

[S.Taylor]

Thanks!
Very interesting.

[pilgrimdan]

Both engines running with 4-Gb hash, all evals given from white's POV.

[D] r2q1r1k/3nbppp/p7/3RpP2/Pp2P3/1PN1Q3/1P5P/1K3BR1 w - - 0 19

Ivanhoe 9.46h Analysis

Code: Select all

29	08:26	5 361 661 963	10 613 000	+0,21	Nc3-b5 Be7-c5 Rd5xc5 Nd7xc5 Qe3xc5 Qd8-d1+ Kb1-a2 a6xb5 Bf1xb5 Qd1-d4 Qc5xd4 e5xd4 Ka2-b1 g7-g6 f5-f6 Ra8-d8 a4-a5 Rd8-d6 Rg1-f1 Rd6-e6 Bb5-d3 h7-h5 a5-a6 Rf8-a8 Kb1-c2 Kh8-h7 Rf1-f2 Kh7-h6 Bd3-c4 Re6xa6 Bc4xa6 Ra8xa6 Kc2-d3 Ra6-d6 h2-h4 g6-g5 e4-e5

Critter 1.6a

Code: Select all

 27/68	08:21	5 019 239 913	10 003 886	+0,19	Nc3-b5 Be7-c5 Rd5xc5 Nd7xc5 Qe3xc5 a6xb5 Bf1xb5 f7-f6 Qc5xb4 Qd8-b6 Qb4-e1 Rf8-c8 h2-h4 Ra8-a5 Qe1-g3 Ra5-a7 Bb5-c4 Qb6-d4 Qg3-g2 Ra7-d7 a4-a5 Qd4-a7 Bc4-e6

Both engines agree that Nb5 is what white should play.

I then ran an analysis in Multi PV (6) mode:

Ivanhoe 9.46h Analysis:

Code: Select all

 26	08:51	5 556 646 450	10 467 000	-0,41	Nc3-e2 Qd8-c7 Qe3-g3 Rf8-g8 Bf1-g2 Ra8-c8 Rg1-c1 Qc7-a7 Rc1-d1 Nd7-f6 Rd5-d2 Qa7-b8 Ne2-c1 Be7-c5 Rd2-c2 Bc5-d4 Rc2-c4 Rg8-d8 Nc1-d3 a6-a5 Rd1-c1 Rc8xc4 b3xc4 Qb8-c7 c4-c5 Kh8-g8 c5-c6 b4-b3 h2-h4
  26	08:51	5 556 646 450	10 467 000	-0,24	Nc3-a2 Qd8-c7 Qe3-g3 Rf8-g8 Bf1-c4 Nd7-f6 Rg1-c1 Nf6xd5 Bc4xd5 Qc7-d6 Bd5xa8 Rg8xa8 Rc1-c4 Ra8-b8 Na2-c1 Rb8-d8 Qg3-f2 Kh8-g8 Qf2-a7 Rd8-d7 Rc4-c8+ Be7-d8 Qa7-c5 g7-g6 Qc5xd6 Rd7xd6 Kb1-c2 g6xf5 e4xf5
  26	08:51	5 556 646 450	10 467 000	-0,20	Qe3-d3 Ra8-a7 Nc3-d1 Qd8-c7 Qd3-g3 Rf8-g8 Rd5-d2 Nd7-f6 Bf1-g2 Nf6-h5 Qg3-f3 Nh5-f4 Nd1-e3 Rg8-d8 Ne3-d5 Nf4xd5 Rd2xd5 Rd8xd5 e4xd5 Be7-c5 Rg1-c1 Qc7-d6 f5-f6 g7-g6 Bg2-f1 Kh8-g8 h2-h4 Ra7-c7
  26	08:51	5 556 646 450	10 467 000	+0,09	Qe3-g3 Rf8-g8 f5-f6 Be7xf6 Nc3-a2 Qd8-e7 Na2xb4 Qe7xb4 Rd5xd7 Qb4xe4+ Bf1-d3 Qe4-b4 Bd3-c4 Rg8-f8 Rg1-d1 a6-a5 Bc4xf7 Qb4-e4+ Kb1-a2 Qe4-f5 Bf7-c4 Ra8-d8 Rd7-d5 Rd8xd5 Rd1xd5 Qf5-c2 Bc4-d3 Qc2-d2 Bd3-e4 Qd2-b4 Qg3-d3 g7-g6 Rd5-d7
  26	08:51	5 556 646 450	10 467 000	+0,20	f5-f6 Be7xf6 Bf1-h3 Qd8-b6 Nc3-d1 Qb6xe3 Nd1xe3 Nd7-b8 Rd5-c5 Bf6-e7 Rc5-c7 Be7-d6 Rc7-c4 g7-g6 Rg1-d1 Bd6-e7 Ne3-d5 Ra8-a7 Nd5xe7 Ra7xe7 Rc4xb4 Nb8-c6 Rb4-b6 Nc6-d4 Bh3-f1 a6-a5 b3-b4 a5xb4 a4-a5 Kh8-g7 Rb6xb4 f7-f5 e4xf5 g6xf5 a5-a6 f5-f4 Rb4-b6
  26	08:51	5 556 646 450	10 467 000	+0,30	Nc3-b5 Be7-c5 Rd5xc5 Nd7xc5 Qe3xc5 a6xb5 Bf1xb5 f7-f6 Qc5xb4 Qd8-b6 Qb4-e1 Rf8-c8 Rg1-g2 Rc8-d8 Bb5-c4 Ra8-a5 Rg2-c2 Rd8-c8 Rc2-c1 Ra5-a7 Qe1-g1 Qb6-a5

Critter 1.6a:

Code: Select all

26/75	29:08	17 543 806 772	10 036 209	-0,46	Nc3-e2 Qd8-c7 Qe3-g3 Rf8-g8 Bf1-g2 Ra8-c8 Rg1-c1 Qc7-a7 Kb1-a2 h7-h6 h2-h4 Be7-c5 Rd5-d3 a6-a5 Rd3-d5 Rg8-e8 Rc1-c2 Kh8-g8 Rc2-d2 Rc8-c7 Rd2-d1 Bc5-f2 Qg3-h2 Re8-c8 Qh2-h3 Bf2-c5 Qh3-g3 Qa7-a6 Ne2-c3 Bc5-d4 Nc3-b5
 26/75	29:08	17 543 806 772	10 036 209	-0,21	Qe3-d3 Ra8-a7 Nc3-d1 Qd8-c8 Qd3-g3 Rf8-g8 Rd5-d2 Be7-c5 Nd1-e3 Qc8-c6 Bf1-g2 Qc6-h6 Rg1-e1 Bc5-d4 Rd2-d1 a6-a5 Ne3-c2 Bd4-c5 Rd1-d5 Ra7-c7 Bg2-f1 Rg8-c8 Re1-d1 Bc5-f2 Qg3xf2 Rc7xc2 Rd1-d2 Rc2xd2 Qf2xd2 Qh6xd2 Rd5xd2
 26/75	29:08	17 543 806 772	10 036 209	-0,12	Nc3-a2 Qd8-c7 Qe3-g3 Rf8-g8 Bf1-c4 Nd7-f6 Rg1-c1 Nf6xd5 Bc4xd5 Qc7-d6 Bd5xa8 Rg8xa8 Rc1-c4 Ra8-b8 Na2-c1 Rb8-d8 Qg3-f2 Kh8-g8 Qf2-a7 Be7-g5 Qa7-b7 g7-g6 h2-h4 Bg5-f4 Rc4-c6 Qd6-d4 Rc6-c4 Qd4-f2 Qb7-e7 Rd8-f8 f5xg6 h7xg6 Qe7xb4 Qf2xh4 Rc4-c7 Bf4xc1 Rc7xc1 Rf8-d8 Rc1-c7 a6-a5 Qb4-b7 Qh4-f2 b3-b4
 26/75	29:08	17 543 806 772	10 036 209	+0,14	Qe3-g3 Rf8-g8 Nc3-b5 Qd8-c8 Nb5-d6 Be7xd6 Rd5xd6 Qc8-c7 Rd6-d4 a6-a5 Bf1-b5 Nd7-f6 Rg1-c1 Qc7-b8 Rd4-c4 Ra8-a7 Qg3-g2 Ra7-e7 Bb5-c6 Qb8-b6 Bc6-d5 Re7-d7 Rc4-c5 Rg8-d8 Rc1-g1 Rd8-g8 Rc5-b5 Qb6-d8 Rg1-d1 h7-h6 Rd1-c1 Nf6xd5 e4xd5
 26/75	29:08	17 543 806 772	10 036 209	+0,17	f5-f6 Be7xf6 Bf1-h3 Qd8-b6 Nc3-d1 Qb6xe3 Nd1xe3 Nd7-b8 Rd5-c5 Bf6-e7 Rc5-c7 Be7-d6 Rc7-c4 g7-g6 Rg1-d1 Bd6-e7 Ne3-d5 Ra8-a7 Nd5xe7 Ra7xe7 Rc4xb4 Nb8-c6 Rb4-b6 Nc6-d4 Bh3-f1 a6-a5 b3-b4 a5xb4 a4-a5 Re7-a7 a5-a6 b4-b3 Bf1-c4 Rf8-c8 Bc4-d5 Kh8-g7 Rd1-c1 Rc8-d8 Rc1-c5 f7-f5 Bd5-b7 f5-f4 Rc5xe5
 26/75	29:08	17 543 806 772	10 036 209	+0,29	Nc3-b5 Be7-c5 Rd5xc5 Nd7xc5 Qe3xc5 Qd8-d1+ Kb1-a2 a6xb5 Bf1xb5 Qd1-d4 Qc5xd4 e5xd4 Ka2-b1 g7-g6 Kb1-c2 Rf8-c8+ Kc2-d3 g6xf5 e4xf5 Rc8-c5 f5-f6 Rc5-f5 Kd3xd4 Ra8-d8+ Kd4-c4 Rf5-f4+ Kc4-c5 Rd8-c8+ Kc5-b6 Rf4xf6+ Kb6-a5 Rf6-f4 Bb5-c4 Rc8-b8

Here again, Nb5 appears as the best move, f6 and Qg3 being less good but still playable. The other moves (Qd3, Na2 and Ne2) just seem bad.

[D] r2q1r1k/3Rbppp/p7/4pP2/Pp2P3/1PN1Q3/1P5P/1K3BR1 b - - 0 19

This is the position after the Rxd7 sac is played. Here is the evaluation by both engines:

Ivanhoe 46h x64 Analysis:

Code: Select all

 30	08:20	5 557 809 423	11 131 000	-0,93	Qd8xd7 Nc3-d5 Qd7-d6 Bf1-c4 a6-a5 Qe3-g3 Be7-f6 Nd5-e3 Qd6-b6 Bc4-d5 Ra8-a7 Ne3-c4 Qb6-d4 Rg1-g2 Rf8-e8 Rg2-d2 Qd4-c5 Rd2-d1 Ra7-d7 h2-h4 Re8-d8 Qg3-g2 h7-h6 Kb1-a2 Rd7-e7 Qg2-g3 Qc5-c7 Rd1-c1

Critter_1.6a_64bit:

Code: Select all

 29/73	10:17	6 574 707 052	10 641 921	-0,83	Qd8xd7 Nc3-d5 Qd7-d6 Bf1-c4 a6-a5 Rg1-g3 Be7-h4 Rg3-h3 Qd6-d8 Qe3-e2 Bh4-g5 Qe2-h5 h7-h6 Rh3-d3 Ra8-c8 Kb1-a2 Rc8-b8 Qh5-g4 Qd8-d6 h2-h4 Bg5-d8 Rd3-d1 Bd8-f6 Nd5-f4 Qd6-b6 Nf4-d5 Qb6-c6 Bc4-b5 Qc6-c2 Nd5xf6 g7xf6 Bb5-c4

The evaluation is pretty similar and doesn't change much over time so I think that White could draw the game with accurate play, but certainly not win it.

So while the sac is not disastrous, it doesn't seem like a sound move to play

Chessmaster 10th Ed. personality Tal...

Celeron 1.5 GHz (1 processor)

2GB memory 64 bit OS

Windows 7


Time Depth Score Positions Moves
5:37 4/13 0.88 138082677 1.Qg3 Bf6 2.Nb5 Qe7 3.Nc7 Nb6 4.Nxa8
Nxd5 5.exd5 Rxa8 6.Bc4 Kg8 7.Qe3
Qd7 8.Qf2

8:53 4/13 0.96 219083127 1.Nb5 Qc8 2.Qg3 Bf6 3.Nd6 Qc7 4.a5
Qa7 5.Be2 Rad8 6.Rc1 Kg8 7.Qg1
Qxg1 8.Rxg1

13:57 4/13 1.10 352158783 1.f6 Bxf6 2.Bh3 Qb6 3.Nd1 Qxe3
4.Nxe3 Nb6 5.Rd6 Be7 6.Rc6 Rfb8
7.Rc7 Re8

18:12 5/14 1.28 459224613 1.f6 Bxf6 2.Bh3 Qb6 3.Nd1 Qxe3
4.Nxe3 Nb6 5.Rd6 Rfb8 6.Ng4 Be7
7.Nxe5 Bxd6 8.Nxf7+ Kg8 9.Nxd6

Chessmaster likes f6...

[Thomas Lagershausen]

Sorry guys, i have bad news for you.

Chessprograms are not sophisticated enough to evaluate a position like this.

Once again. Computer didn´t have any evaluation for longterm details.

White has a eternity knight on d5 and this knight is stronger as the black rook.

Computer didn´t understand this.

Sorry to disturb your funny playing.

[Karlo Bala]

Sorry guys, i have bad news for you.

Chessprograms are not sophisticated enough to evaluate a position like this.

Once again. Computer didn´t have any evaluation for longterm details.

White has a eternity knight on d5 and this knight is stronger as the black rook.

Computer didn´t understand this.

Sorry to disturb your funny playing.

I think you are correct about this particular position. From a human-human point of view, I think that the sacrifice on d7 is an excellent choice. At the first sight, I thought that black is winning, but then I realize that it is very hard for black to make any progress. Moreover, I think that the white can play almost without the risk of making any blunder.

[DocNZ]

Sorry guys, i have bad news for you.

Chessprograms are not sophisticated enough to evaluate a position like this.

Once again. Computer didn´t have any evaluation for longterm details.

White has a eternity knight on d5 and this knight is stronger as the black rook.

Computer didn´t understand this.

Sorry to disturb your funny playing.

Really?

Are chess programs sophisticated enough to analyse this position?

Well I think they are - well at least some are. Knight outposts are an important part of the evaluation of a position.

But I don't think this knight on d5 is all that strong. Maybe getting the knight to d5 without the exchange sacrifice wold have been better.

I think that one of the inherent weaknesses in whites position after the exchange is the backward e pawn and the weak q side. One strategy for black would be to give back the exchange and put pressure on the e pawn.

Rxd7 Qxd7 / Nd5 Qd6 / Bc4 Rfd8 / Qg3 Bf6 all kind of forced / then maybe
h4 a5 / Ka2 Rac8 / Qf2 (!) h6 forced / Qg2 Rc6 / Qg4 Rdc8 an this is the point / if Rd1 Rxc4 / bxc4 Rxc4 and I think black has chances.

I've put in a lot of effort with this initial position, and I have really enjoyed it. My final impression is that Rxd7 in over-the-board play is a good move. However its not a winning move, as I can't see how white can profit from it. I'm sure from a purely psychological POV Rxd7 may have been devastating. Blacks strategy after Rxd7 IMHO is to return the favor and put pressure on the e pawn.

I finally think that the computer analysis of Rxd7 as -0.8 (Blacks favor) is probably correct.

[S.Taylor]

Sorry guys, i have bad news for you.

Chessprograms are not sophisticated enough to evaluate a position like this.

Once again. Computer didn´t have any evaluation for longterm details.

White has a eternity knight on d5 and this knight is stronger as the black rook.

Computer didn´t understand this.

Sorry to disturb your funny playing.

Are YOU the voice of authority?
Are you looking at the position objectively and comprehensively if you are?

I think, that in this position, White would do better keeping the rook as there is much he can do with it.

White can still get an eternity BIshop or eternity Knight on that square without losing the Rook.

So, as far as I'm concerned, computers are fine too, in this position.
you seem to be suggesting that Houdini should jump in its evaluation to +2 Pawns, right after the sac. Well, it didn't go to -2 Pawns (which would have been linear logic), but i think +2 would have been incorrect too.

[Long term? So you say white can do things more freely? and houdini vs Houdini didn't notice this when it was playing itself? So you say that all whites other pieces become stronger without Blacks Knight on the board? and that even Whites King becomes strong as an attacking player? And that the opposite coloured Bishops gives White an extra piece in this position? And that without the sac Black maybe can draw but with the sac he certainly can't? If that's what you mean, I might rethink!]

[Dr.Wael Deeb]

Sorry guys, i have bad news for you.

Chessprograms are not sophisticated enough to evaluate a position like this.

Once again. Computer didn´t have any evaluation for longterm details.

White has a eternity knight on d5 and this knight is stronger as the black rook.

Computer didn´t understand this.

Sorry to disturb your funny playing.

Are YOU the voice of authority?
Are you looking at the position objectively and comprehensively if you are?

I think, that in this position, White would do better keeping the rook as there is much he can do with it.

White can still get an eternity BIshop or eternity Knight on that square without losing the Rook.

So, as far as I'm concerned, computers are fine too, in this position.
you seem to be suggesting that Houdini should jump in its evaluation to +2 Pawns, right after the sac. Well, it didn't go to -2 Pawns (which would have been linear logic), but i think +2 would have been incorrect too.

[Long term? So you say white can do things more freely? and houdini vs Houdini didn't notice this when it was playing itself? So you say that all whites other pieces become stronger without Blacks Knight on the board? and that even Whites King becomes strong as an attacking player? And that the opposite coloured Bishops gives White an extra piece in this position? And that without the sac Black maybe can draw but with the sac he certainly can't? If that's what you mean, I might rethink!]

Well said Shimon....

Can't agree more with your statement......
Dr.D

[S.Taylor]

And indeed, If he is saying he is claiming most of those things, OR simply that he is 2400+ at this present time, then i would believe i have a lesson to learn from him in this position and would try to see what gave me the blindspots and become more aware [and indeed, i DID start seeing how that knight would hamper Blacks movements, but i don't know how much HE intended with his strong statements]. But if not, then it is he, being out of touch with human and computer chess, speaking so strongly.

[Thanar]

[d]r2q1r1k/3nbppp/p7/3RpP2/Pp2P3/1PN1Q3/1P5P/1K3BR1 w - - 0 19
I decided to investigate the Qg3 line from the initial position (instead of Rxd7). Here is one plausible line (though I'm sure there could be improvements for both sides):

19.Qg3 Rg8 20.Nb5 Qc8 21.Nd6 Bxd6 22.Rxd6 Qc7 23.Rd4 Qb7 24.Rd1 Qxe4+ 25.Ka2 Qb7 26.Qh4 Nf6 27.Rxg7 Rxg7 28.Qxf6 e4
[d]r6k/1q3prp/p4Q2/5P2/Pp2p3/1P6/KP5P/3R1B2 w - - 0 29
Here, white has the far-reaching 29.h4! which looks winning to me. One continuation would be 29...Qc7 30.Bxa6 e3 31.h5 Rxa6 32.Rd8+ Qxd8 33.Qxd8+ Rg8 34.Qd4+ Rg7 35.Qxe3
[d]7k/5prp/r7/5P1P/Pp6/1P2Q3/KP6/8 b - - 0 35
Anyone else investigate lines like this?