Houdini 3 wins nice game over Rybka 4.1.

[mwyoung]

Houdini 3 unleashes a clever counter attack on Rybka 4 with a nice exchange sac, rook for bishop. The game looks like a positionally lost endgame after the exchange sac, but it takes many moves to prove the win.

Time control 40/2.5 hours 16/1 hour.

[Dr.Wael Deeb]

Houdini 3 unleashes a clever counter attack on Rybka 4 with a nice exchange sac, rook for bishop. The game looks like a positionally lost endgame after the exchange sac, but it takes many moves to prove the win.

Time control 40/2.5 hours 16/1 hour.

Nice......
Dr.D

[Nay Lin Tun]

It is the obvious mistake of Rybka by exchanging one major piece with two pawns. If Rybka makes 36. R*a6 then 37. B*b4, it is a dead draw with different coloured bishops. It seems that Rybka has no codes or ideas about the theory of different coloured bishops.

[beram]

It is the obvious mistake of Rybka by exchanging one major piece with two pawns. If Rybka makes 36. R*a6 then 37. B*b4, it is a dead draw with different coloured bishops. It seems that Rybka has no codes or ideas about the theory of different coloured bishops.

[d] 8/1bp5/n3p1k1/4Pp2/1ppP4/5P1P/3B2K1/R7 w - - 0 36

Interesting position; only Deep Rybka 4.1 wants to play Kg3 here. All the other engines I ve tested choose almost immediately Rxa6 which, I agree on, is a dead draw, because off the opposite coloured bishops.
Since Rybka 3 also wants to play Rxa6 it has to be something specific for Rybka 4.1(and 4.0)
------------------
But what I think is a really nice move in this game, is the inventive pawn sacrifice 23...h6! followed by Rg8 intending an exchange sac 26 ..Rxg2
I don't see other engines choose this move
[d] b3r2k/2pq3p/nn2p3/1p2PpB1/2pP4/2P2N2/5PBP/1Q2R2K b - - 0 23

[beram]

Houdini 3 unleashes a clever counter attack on Rybka 4 with a nice exchange sac, rook for bishop. The game looks like a positionally lost endgame after the exchange sac, but it takes many moves to prove the win.

Time control 40/2.5 hours 16/1 hour.

The first loss of Deep Critter 1.6 against Houdini 3 in eight games. The decisive mistake occurred in this position. Despite the long time control DC1.6 could just not search deep enough to avoid the loosing move Rxa5. After 70 Rxa5? - d4! white is in great trouble because black invades the white position with his pieces or creates a free pawn on the d-line

Mark thx a lot for these LTC match publication here and on youtube. These games are very high level, the programs reach very high ply level. Btw what are the specs of your machine. It must be a lot better than my AMD 1090T 6core

Critter 1.6a 64-bit - Houdini 3 x64, LTC Chess Match 2012
[d] 8/2r3pk/1q2p1p1/p2pPp2/R4P1P/1rP3PK/1P2R3/Q7 w - - 0 1

Code: Select all

Analysis by Critter 1.6a 64-bit:

70.Rg2 Rb5 
  =  (0.00)   Depth: 18/41   00:00:28  11021kN
70.Rg2 Rb5 
  =  (0.00)   Depth: 19/41   00:00:31  12828kN
70.Rg2 Rb5 
  =  (0.00)   Depth: 20/41   00:00:38  21714kN
70.Rxa5 d4 71.Ra3 Rxa3 72.Qxa3 Qb7 73.Qa1 Qf3 74.Qe1 g5 75.fxg5 dxc3 76.Rc2 Rc4 77.Kh2 Rd4 78.bxc3 Rd1 79.Rf2 Qxg3+ 80.Kxg3 Rxe1 81.c4 Rg1+ 82.Kf3 Rc1 83.h5 Rc3+ 84.Ke2 Rxc4 85.Kd3 Rh4 
  =  (0.00)   Depth: 21/48   00:00:59  85206kN
70.Rxa5 d4 71.Ra3 Rxa3 72.Qxa3 Qb7 73.Qa1 Qf3 74.Qe1 dxc3 75.Rc2 g5 76.fxg5 cxb2 77.Rxc7 Qg4+ 78.Kg2 b1R 79.Rxg7+ Kxg7 80.Qxb1 Qe2+ 81.Kh3 Qg4+ 82.Kh2 Qe2+ 83.Kg1 Qe3+ 84.Kg2 Qe2+ 
  =  (0.00)   Depth: 22/48   00:01:16  120mN
70.Rxa5 d4 71.Ra3 Rxa3 72.Qxa3 Qb7 73.Qa1 Qf3 74.Qe1 dxc3 75.Rc2 g5 76.fxg5 cxb2 77.Rxc7 Qg4+ 78.Kg2 b1R 79.Rxg7+ Kxg7 80.Qxb1 Qe2+ 81.Kh3 Qg4+ 82.Kh2 Qe2+ 83.Kg1 Qe3+ 84.Kg2 Qe2+ 
  =  (-0.08 --)   Depth: 23/59   00:02:17  487mN
70.Rg2 Rc5 71.Qg1 Qc6 72.Qa1 Qb5 73.Rd4 Qb6 74.Qa2 Rcb5 75.Rdd2 a4 76.Qxa4 Rxb2 77.Rxb2 Rxb2 78.Qd4 Qb8 79.Rxb2 Qxb2 80.Qe3 Qc2 81.Qe1 Qd3 82.Kg2 Kg8 83.Qa1 Qe4+ 84.Kh3 Qf3 85.Qb2 Qh1+ 86.Qh2 Qe1 87.Qc2 Qf1+ 
  =  (-0.08)   Depth: 23/59   00:02:36  563mN
70.Rg2 Rc5 71.Qg1 Qc6 72.Qa1 Qb7 73.Rd4 Rcb5 74.Rdd2 Qb6 75.Qa2 a4 76.Qxa4 Rxb2 77.Rxb2 Rxb2 78.Qd4 Qb8 79.Rd2 Rb1 80.Rd1 Rxd1 81.Qxd1 Qb6 82.Qe1 Qb2 83.Qe3 Qc2 84.Qe1 Qd3 85.Kh2 Kh6 86.Kg2 Kh5 87.Kh3 Qf3 88.Qb1 Qg4+ 89.Kg2 Qe2+ 90.Kh3 Qf3 91.Qe1 Kh6 92.Kh2 Qd3 
  =  (-0.08)   Depth: 24/59   00:02:42  593mN
70.Rg2 Rc5 71.Qg1 Qc6 72.Qa1 Qb7 73.Rd4 Rcb5 74.Rdd2 Qb6 75.Qa2 a4 76.Qxa4 Rxb2 77.Rxb2 Rxb2 78.Qd4 Qb8 79.Rxb2 Qxb2 80.Qe3 Qc2 81.Qe1 Qd3 82.Kh2 Kh6 83.Kg2 Kh5 84.Kh3 Qf3 85.Qb1 Qg4+ 86.Kg2 Qe2+ 87.Kh3 Qf3 88.Qe1 Kh6 89.Kh2 Qd3 
  =  (-0.11)   Depth: 25/59   00:03:09  715mN

[mwyoung]

My computer is a 2 year old i7 cpu, with 8 gigs of ram. And it is a LAPTOP, what is called a desktop replacement laptop. The fastest I could get 2 years ago. It is still ok fast by todays standards.

I like to take my desktop with me so I buy the fastest laptops. When at home it is hook up to a Vizio 65" Smart 3d TV as my monitor and I wireless keyboard and mouse. Awesome to play and watch chess on.

[mwyoung]

I missed that move with my limited skills looking at the game. After seeing it RxN it is a clear draw.

Strange Rybka 4.1 missed that move. At this level of play the programs sometimes lose by some strange hole in their programming. It maybe true that this "hole" could help Rybka 4 more then it hurts Rybka 4 in the long run...

Not like vas to miss this in testing. Need to see if this happens again. I don't remember this happing in the TCEC elite chess match.