LCTII.CMB.11 -> Nxf7!!

[Vinvin]

Still interesting, in 1994, Vaisser killed ChessGenius2 with Nxf7!!

[d]rnbq2k1/p1r2p1p/1p1p1Pp1/1BpPn1N1/P7/2P5/6PP/R1B1QRK1 w - - 0 1

LCTII epd : http://www.blitzchess.fr/fr/downloads/lct2.epd
LCTII diag : http://demetrix.pagesperso-orange.fr/Chess/LCTIIcmb.htm

[ernest]

Indeed, Houdini is far from finding immediately...
(dual-core, 3GHz)

but maybe other moves (a5, Bf4) also lead to win.

rnbq2k1/p1r2p1p/1p1p1Pp1/1BpPn1N1/P7/2P5/6PP/R1B1QRK1 w - - 0 1
Analysis by Houdini 2.0c w32 Nlm:
...
1.a5
+- (1.42 !) Depth: 18/63 00:00:37 92604kN
1.a5
+- (1.65 !) Depth: 18/63 00:00:59 149mN
1.a5 h5 2.Bf4 Nbd7 3.Ne4 a6 4.axb6 Nxb6 5.Nxd6 axb5 6.Rxa8 Nxa8 7.Bxe5 Bb7 8.Qe3 Bxd5 9.Nf5 Ra7 10.Qh6 Qf8 11.Ne7+ Rxe7 12.fxe7 Qxh6 13.e8Q+ Qf8 14.Qxb5 Kh7 15.Rd1 Be6 16.c4 Qe7
+- (1.78) Depth: 18/63 00:01:21 205mN
1.a5
+- (1.90 !) Depth: 19/63 00:01:52 285mN
1.a5 c4 2.Ne4 Rc5 3.Nxc5 bxc5 4.Qh4 a6 5.Qh6 Qf8 6.Qxf8+ Kxf8 7.Ba4 Bd7 8.Bf4 Bxa4 9.Bxe5 dxe5 10.Rxa4 Nd7 11.Rxc4 Rb8 12.Rh4 Kg8 13.c4 h5 14.g4 hxg4 15.Rxg4 Rb3 16.Kg2 Ra3 17.Rg3 Ra2+ 18.Rf2 Rxa5
+- (1.87) Depth: 19/63 00:02:33 388mN
1.a5
+- (1.99 !) Depth: 20/63 00:02:58 454mN
1.a5
+- (2.22 !) Depth: 20/64 00:04:32 695mN
1.a5 Rb7 2.Qh4 h5 3.a6 Rc7 4.Be2 Bg4 5.Bxg4 Nxg4 6.Ne6 Qc8 7.Qg5 Nd7 8.Nxc7 Qxc7 9.h3 Ngxf6 10.Rxf6 Nxf6 11.Qxf6 Qd8 12.Qxd8+ Rxd8 13.Bg5 Re8 14.Bf4 Kg7 15.c4 Rd8 16.Re1 Kf6 17.Kf2 h4 18.Bc1 Rd7 19.Kf3 Kf5 20.Bb2 f6
+- (2.18) Depth: 20/64 00:05:43 880mN
1.Nxh7
+- (2.75 !) Depth: 20/64 00:06:09 952mN
1.Nxh7
+- (4.18 !) Depth: 20/70 00:06:37 1026mN
1.Nxh7
+- (8.00 !) Depth: 20/70 00:07:25 1158mN
1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 4.Bg7 Bh5 5.Qg5 Ng4 6.h3 Nxf6 7.Bxf6 Qf8 8.g4 a6 9.gxh5 axb5 10.hxg6 fxg6 11.Qxg6+ Rg7 12.Bxg7 Qxg7 13.Qxg7+ Kxg7 14.axb5 Rxa1 15.Rxa1 Kf6 16.Ra7 Ke5 17.Rb7 Kxd5 18.Rxb8 Kc4 19.Rxb6 d5 20.Rd6 Kxb5 21.Rxd5 Kc4 22.Rg5 Kxc3 23.Rxc5+ Kd4 24.Rf5

2nd try:
Analysis by Houdini 2.0c w32 Nlm:
...
1.Bf4
+- (2.87 !) Depth: 19/61 00:01:14 194mN
1.Bf4 Nbd7 2.Rd1 Qxf6 3.Bxe5 Qxe5 4.Nxf7 Qxe1 5.Rdxe1 Nf8 6.Nxd6 Bf5 7.Nxf5 gxf5 8.d6 Rg7 9.Rxf5 Rd8 10.Re7 Ng6 11.Re8+ Rxe8 12.Bxe8 Rb7 13.d7 Rb8 14.Kf2 Ne7 15.Re5 Nc6 16.Re2 Kf8 17.Bh5 Ne7 18.Kg3 Rd8 19.Rf2+ Kg8
+- (3.01) Depth: 19/68 00:01:42 269mN
1.Bf4 Nbd7 2.Rd1 Nxf6 3.Bxe5 dxe5 4.d6 Rb7 5.d7 Bxd7 6.Qh4 Qe7 7.Bxd7 Rxd7 8.Rxd7 Qxd7 9.Rxf6 Qd1+ 10.Rf1 Qh5 11.Qe4 Rf8 12.Qxe5 h6 13.Nxf7 Qxe5 14.Nxe5 Re8 15.Nxg6 Kg7 16.Nf4 Re4 17.g3 Rxa4 18.Re1 Ra2 19.Re7+ Kf6 20.Rh7 Ra1+ 21.Kg2 Ra2+ 22.Kf3 Rxh2 23.Rxa7
+- (2.95) Depth: 20/68 00:01:59 317mN
1.Bf4
+- (3.10 !) Depth: 21/68 00:02:59 479mN
1.Bf4 Nbd7 2.Rd1 Bb7 3.Ne6 Qc8 4.Nxc7 Qxc7 5.Bxd7 Nxd7 6.Qe7 Qd8 7.Bxd6 Qxe7 8.fxe7 f6 9.g4 Kf7 10.g5 f5 11.Kg2 Rh8 12.c4 h6 13.gxh6 Rxh6 14.Rfe1 Rh8 15.Re6 Re8 16.Kg3 Nf6 17.Rf1
+- (3.04) Depth: 21/68 00:03:19 534mN
1.Bf4 Nbd7 2.Rd1 Qxf6 3.Bxe5 Qxe5 4.Nxf7 Qxe1 5.Rdxe1 Ne5 6.Nxd6 a6 7.Rxe5 axb5 8.Nxb5 Rd7 9.Re8+ Kg7 10.c4 Bb7 11.Rxa8 Bxa8 12.Re1 Bxd5 13.cxd5 Rxd5 14.Re6 Rd1+ 15.Kf2 Ra1 16.Rxb6 Rxa4 17.Rb7+ Kh6 18.Nd6 c4 19.Rc7 Ra2+ 20.Kg3 Ra3+ 21.Kf4 c3 22.Ne4 Ra2 23.Rxc3 Rxg2 24.h3 g5+ 25.Kf5
+- (3.15) Depth: 22/68 00:04:47 774mN
1.Bf4
+- (3.30 !) Depth: 23/70 00:08:33 1373mN
1.Bf4 Nbd7 2.Rd1 Bb7 3.Ne6 fxe6 4.dxe6 Nxf6 5.Bxe5 Nd5 6.Bc4 dxe5 7.Qxe5 Re7 8.Rf7 Rxf7 9.exf7+ Kxf7 10.Rxd5 Bxd5 11.Bxd5+ Kf8 12.Bxa8 Qe8 13.Qf6+ Qf7 14.Qh8+ Ke7 15.Bf3 Kd6 16.Qa8 Qc7 17.g3 Qd7 18.Qb8+ Qc7 19.Qg8 Qd7 20.Qd5+ Kc7 21.Qb7+ Kd6 22.Qb8+ Qc7
+- (3.58) Depth: 23/72 00:12:54 2074mN
1.Bf4
+- (3.76 !) Depth: 24/72 00:22:35 3532mN
1.Bf4 Nbd7 2.Rd1 Bb7 3.Ne6 fxe6 4.dxe6 Nxf6 5.Bxe5 Nd5 6.Bc4 dxe5 7.Qxe5 Re7 8.Rf7 Rxf7 9.exf7+ Kxf7 10.Rxd5 Bxd5 11.Bxd5+ Kf8 12.Bxa8 Qe8 13.Qf6+ Qf7 14.Qh8+ Ke7 15.Bf3 Kd6 16.Qa8 Qc7 17.g3 Ke6 18.Bd5+ Kd6 19.Bc4 Qd7 20.Bb5 Qe7 21.Qc6+ Ke5 22.Bc4 Kf5 23.Bd3+ Kg4 24.Kg2 Kg5 25.Qf3
+- (3.54) Depth: 24/76 00:23:40 3702mN
1.Nxh7
+- (4.12 !) Depth: 24/76 00:25:45 4035mN
1.Nxh7
+- (5.00 !) Depth: 24/76 00:28:40 4504mN

[Houdini]

Indeed, Houdini is far from finding immediately...
(dual-core, 3GHz)

but maybe other moves (a5, Bf4) also lead to win.

Yes, the very good moves Bf4 or Qh4 will hide the even more crushing Nxh7 for quite a while.
In the first run you were "lucky" that Bf4!/Qh4! were not picked up, that's why Nxh7!! was found a lot faster than in the second run.
Below a run with the current Houdini sticking to Qh4 up to depth 25:

Code: Select all

rnbq2k1/p1r2p1p/1p1p1Pp1/1BpPn1N1/P7/2P5/6PP/R1B1QRK1 w - -

Engine: Houdini DEV 3T (2048 MB)

18/52  0:07   +0.97    1.Qe4 h5 2.Ne6 fxe6 3.dxe6 Bxe6 
                       4.Qxa8 Nbc6 5.Qxd8+ Nxd8 6.Bf4 Kf7 
                       7.Bxe5 dxe5 8.a5 bxa5 9.Rxa5 e4 
                       10.Re1 Bf5 11.Ra6 c4 12.Kf2 Rc5 
                       13.Rxa7+ Kxf6 14.Ra6+ (38.349.917) 4963 

18/52  0:09   +1.07++  1.Qh4 (45.800.962) 4983 

18/52  0:09   +1.26++  1.Qh4 (47.358.176) 4978 

15/52  0:09   +1.47    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qxd5 10.hxg6 fxg6 11.Qh6 Qf7 
                       12.Ba3 Nbc6 13.Bxd6 Nd3 14.Rg1 (48.469.223) 4968 

16/52  0:09   +1.47    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qxd5 10.hxg6 fxg6 11.Qh6 Qf7 
                       12.Ba3 Nbc6 13.Bxd6 Nd3 14.Rg1 (48.842.930) 4963 

17/52  0:10   +1.59++  1.Qh4 (50.557.215) 4963 

17/52  0:10   +1.59    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qxd5 10.hxg6 fxg6 11.Qh6 Qf7 
                       12.Ba3 Nbc6 13.Bxd6 Nd3 14.Rg1 (50.822.249) 4957 

18/52  0:10   +1.70++  1.Qh4 (52.882.107) 4951 

18/53  0:11   +1.94++  1.Qh4 (56.024.936) 4947 

18/55  0:12   +1.96    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 Qc8 8.gxh5 Nbd7 
                       9.Qg3 Qa6 10.hxg6 Nxg6 11.Qg5 Re8 
                       12.Bf4 Qd3 13.Bxd6 Qe3+ 14.Qxe3 (61.026.345) 4957 

19/58  0:15   +2.08++  1.Qh4 (80.316.664) 5038 

19/58  0:16   +2.24    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 Qc8 8.gxh5 Nbd7 
                       9.Qg3 Nf8 10.Bf4 Qf5 11.Bxe5 Qxe5 
                       12.Qxe5 dxe5 13.a5 Rd8 14.hxg6 (84.500.405) 5028 

20/58  0:18   +2.36++  1.Qh4 (93.591.203) 5020 

20/58  0:20   +2.41    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 Qc8 8.gxh5 Nbd7 
                       9.Qg3 Nf8 10.Bf4 Qf5 11.Bxe5 Qxe5 
                       12.Qxe5 dxe5 13.a5 Rd8 14.hxg6 (103.520.952) 5011 

21/58  0:26   +2.53++  1.Qh4 (130.715.259) 5010 

21/62  0:32   +2.78++  1.Qh4 (162.165.342) 5041 

21/64  0:42   +3.36++  1.Qh4 (213.527.517) 5023 

21/76  0:50   +3.38    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 Qc8 8.gxh5 a6 
                       9.hxg6 Nxg6 10.Qg3 Qf8 11.Bf4 Nd7 
                       12.Bxd6 Qh6 13.Rae1 b5 14.axb5 (255.381.579) 5033 

22/76  1:04   +3.55++  1.Qh4 (324.090.631) 5023 

22/77  1:19   +3.89++  1.Qh4 (400.834.831) 5026 

22/77  1:30   +3.91    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qc8 10.hxg6 Nxg6 11.Qh6 Qf8 
                       12.Qxf8+ Kxf8 13.Ba3 Ke8 14.a5 (455.023.860) 5020 

23/77  1:47   +3.91    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qc8 10.hxg6 Nxg6 11.Qh6 Qf8 
                       12.Qxf8+ Kxf8 13.Ba3 Ke8 14.a5 (544.038.532) 5044 

24/77  2:31   +4.10++  1.Qh4 (776.614.879) 5131 

24/77  2:35   +4.10    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc8 
                       9.hxg6 Nxg6 10.Qh6 Qf8 11.Qxf8+ Nxf8 
                       12.Ba3 Na6 13.Bxd6 Rd8 14.Bxf8 (799.145.598) 5136 

25/77  4:09   +4.26    1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4 4.Ne6 Qc8 
                       5.h3 Ne5 6.Nxc7 Qxc7 7.g4 c4 8.gxh5 Qc5+ 
                       9.Kh2 Qc8 10.hxg6 Nxg6 11.Qh6 Qf8 
                       12.Qxf8+ Nxf8 13.Ba3 Na6 14.Bxd6 (1.302.298.681) 5210 

25/77  4:29   +4.30++  1.Nxh7 (1.411.762.661) 5232 

25/77  4:39   +4.70++  1.Nxh7 (1.461.811.609) 5236 

22/77  4:43   +5.72++  1.Nxh7 (1.483.852.588) 5235 

22/77  5:02   +8.31++  1.Nxh7 (1.583.214.285) 5235 

22/77  5:59   +12.15   1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 
                       4.Bg7 Bh5 5.Qg5 Ng4 6.Be2 Nxf6 
                       7.Bxf6 Qf8 8.Bxh5 Nd7 9.Bxg6 Nxf6 
                       10.Rxf6 Qg7 11.Rxd6 Rb7 12.Rf1 Rab8 
                       13.Rd8+ Rxd8 14.Qxd8+ (1.905.959.424) 5296 

Robert

[Vinvin]

Correction : Nxh7!! (not f7)

[kgburcham]

Nxh7
[D] rnbq2k1/p1r2p1p/1p1p1Pp1/1BpPn1N1/P7/2P5/6PP/R1B1QRK1 w - -

Houdini 2.0c Pro x64
12 threads
HT off

18/54 0:01 +1.41 1.Bf4 Nbd7 2.Ne4 Qf8 3.c4 Bb7 (39.876.828) 23498
18/54 0:02 +1.53 1.Nxh7 (48.248.331) 23992
18/54 0:02 +2.06 1.Nxh7 (50.658.230) 24145
18/54 0:02 +3.48 1.Nxh7 (63.538.387) 24965
18/65 0:05 +7.38 1.Nxh7 (147.504.325) 25923
18/65 0:06 +7.78 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (166.491.183) 26030
19/65 0:07 +8.08 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (192.686.247) 26471
20/65 0:09 +8.41 1.Nxh7 (249.457.902) 27221
20/65 0:10 +8.43 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (284.891.590) 27451
21/65 0:11 +8.43 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (330.534.773) 27799
22/65 0:16 +8.62 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (473.055.373) 28595
23/68 0:27 +8.97 1.Nxh7 (806.176.660) 29588
23/72 0:37 +9.27 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (1.131.431.475) 29906
24/72 1:07 +9.65 1.Nxh7 (2.070.490.793) 30668
24/72 1:55 +10.41 1.Nxh7 (3.625.556.765) 31289
24/72 3:42 +12.54 1.Nxh7 (7.183.672.824) 32235
24/72 11:28 +18.78 1.Nxh7 (23.818.673.472) 34574
24/81 23:47 +20.67 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (50.225.757.482) 35179
24/81 32:01 +20.67 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 4.Bg7 Bh5 5.Qg5 Ng4 (67.620.982.403) 35192

[ernest]

18/54 0:02 +1.53 1.Nxh7 (48.248.331) 23992

Hi KG,

I think you got lucky on this!

If you run it 3 times, will you always get Nxh7 at depth 18?

[kgburcham]

18/54 0:02 +1.53 1.Nxh7 (48.248.331) 23992

Hi KG,
I think you got lucky on this!
If you run it 3 times, will you always get Nxh7 at depth 18?

yes you have a good eye Ernest.
I paid extra for the Lucky Option as seen below.

12 processor(s) found, POPCNT available
NUMA configuration with 2 node(s), offset 0
4096 MB Large Page Hash
Nalimov 6 men EGTB available - 512 MB cache
Lucky Mode Enabled
Engine: Houdini 2.0c Pro x64 (4096 MB)
by Robert Houdart

[kgburcham]

18/54 0:02 +1.53 1.Nxh7 (48.248.331) 23992

Hi KG,
I think you got lucky on this!
If you run it 3 times, will you always get Nxh7 at depth 18?

Of course you are correct.
Tried three more times, on the third try after the original post, I get this.
So it might hit it 1 out of 4 or 5 times, but at depth 20 still fast.

Engine: Houdini 2.0c Pro x64 (4096 MB)
by Robert Houdart

18/60 0:04 +2.04 1.Qh4 h5 2.Be2 Bg4 3.Bxg4 Nxg4
18/60 0:04 +3.30++ 1.Nxh7 (123.592.937) 25483
18/65 0:05 +7.22++ 1.Nxh7 (146.968.123) 25915
18/65 0:07 +8.78 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (209.837.302) 26724
19/65 0:08 +8.78 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (242.995.992) 27065
20/65 0:12 +9.14++ 1.Nxh7 (347.789.751) 28033
20/65 0:16 +9.54 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (474.502.785) 28606
21/65 0:22 +9.82 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (661.998.179) 29195
22/65 0:34 +10.08 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (1.042.028.451) 29868
22/65 0:53 +10.08 1.Nxh7 Kxh7 2.Qh4+ Kg8 3.Bh6 Bg4 (1.639.980.229) 30495

[MikeGL]

I tried 5 times with Hannibal1.2 in my 32bit machine, clearing the hash in each try, and Hannibal1.2 consistently finds Nxh7 every after 1 second.
Makes me wonder if this position is hardcoded, or maybe the pawn shelter of king has different piece-value.




[D] rnbq2k1/p1r2p1p/1p1p1Pp1/1BpPn1N1/P7/2P5/6PP/R1B1QRK1 w - - 0 1

Hannibal1.2x32:
8 00:00 7.315 457.187 +1.19 Qh4 h5 Ne4 Bf5 Be2 Qf8 Be3 Ng4 Bxg4 Bxe4
9 00:00 20.005 317.539 +1.17 Qh4 h5 Ne4 Bf5 Rxf5 gxf5 Qg5+ Ng6 Qh6 Qf8
10 00:00 68.435 364.015 +1.24 a5 h6 axb6 Rb7 Ne4 g5 Rxa7 Rbxa7 bxa7 Rxa7
11+ 00:00 118.852 379.718 +1.40 Qh4 h5 Ne4 Bf5 Bh6 a6 Be2 Kh8
11 00:00 300.311 426.578 +1.11 Qh4 h5 Ne4 Bb7 Bh6 Bxd5 Rad1 Bxe4 Qxe4 Nbc6 Kh1 Kh8
12+ 00:00 367.718 435.684 +1.27 Qh4 h5 Ne4 Bb7 Bh6 Bxd5 Rad1 Bxe4 Qxe4 Nbc6 Bg7 a6 Be2 g5
12 00:01 471.530 450.362 +1.19 Qh4 h5 Ne4 Bb7 Rd1 a6 Be2 Qf8 Qf4 Nbd7 Be3 Kh8 Rab1 Re8
13+ 00:01 575.890 472.428 +1.35 Nxh7 Kxh7 Qh4+ Kg8 Bh6 Nbd7 Bxd7 c4 Bxc8 Raxc8
13+ 00:01 631.966 481.314 +1.83 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Ba6 Bxg6 Bxf1 Bd3+ Kh8 Bxf1 c4 Qh4 Nd7 Bxc4 Rxc4 Qxc4 Qxf6
13+ 00:01 733.680 499.441 +2.79 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Bh3
13+ 00:01 992.739 516.513 +4.71 Nxh7 Kxh7 Qh4+ Kg8 Bh6 Bg4 Bg7 Bh5 Qf4 Nf3+ gxf3 Na6 Bxa6 Qe8
13 00:02 1.431.453 548.449 +4.45 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Qe8 Bxg6 fxg6 f7+ Qxf7 Rxf7 Kxf7 Re1
14+ 00:02 1.567.461 554.068 +4.61 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Bf5 Bxf5 a6 Bxg6 Kh8
14 00:03 1.991.666 569.047 +5.66 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Qe8 Bxg6 fxg6 f7+ Qxf7 Rxf7 Kxf7 Re1 Na6 Qd8
15 00:03 2.118.844 577.027 +4.45 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Qe8 Bxg6 fxg6 f7+ Qxf7 Rxf7 Kxf7 Re1
16+ 00:04 2.371.387 592.846 +4.61 Nxh7 Kxh7 Qh4+ Kg8 Bh6 g5 Qxg5+ Ng6 Bd3 Qe8 Bxg6 fxg6 f7+ Qxf7 Rxf7 Kxf7 Qf4+ Bf5 Qxd6 Na6 g4 Rd7 Qf4 Rxd5 gxf5 c4
16 00:06 4.115.531 601.334 +7.70 Nxh7 Nbd7 Qh4 Ng4 h3 Nf8 Nxf8 Qxf8 hxg4 Qd8 Bd3 c4 Bc2 Qe8 Bh6 Bf5 gxf5 Qe2 Bg7
17 00:08 5.284.458 606.085 +7.70 Nxh7 Nbd7 Qh4 Ng4 h3 Nf8 Nxf8 Qxf8 hxg4 Qd8 Bd3 c4 Bc2 Qe8 Bh6 Bf5 gxf5 Qe2 Bg7
18 00:15 9.266.988 617.799 +8.69 Nxh7 Nbd7 Qh4 Ng4 h3 a6 hxg4 axb5 Qh6 Nxf6 Nxf6+ Qxf6 Rxf6 bxa4 Bf4 Bxg4 Bxd6 Rd7




also, I tried with R4 and Nxh7 is found every time it reaches Depth 13 at random times,
00:03:07
00:03:17
00:06:49
00:01:27
00:02:29

old T3200 Dual Core 32bit

[zullil]

I tried 5 times with Hannibal1.2 in my 32bit machine, clearing the hash in each try, and Hannibal1.2 consistently finds Nxh7 every after 1 second.

Since Hannibal doesn't use a multithreaded search, it seems quite likely that Hannibal's search is deterministic. This would explain the consistency of your five results, though not how the engine is finding Nxh7 so quickly.

[EDIT] Compare the output you posted to the output here:

http://talkchess.com/forum/viewtopic.ph ... 28&t=44020

Same score and same number of nodes, since the search is deterministic.