unsound sacrifices of chess programs

[Uri Blass]

I looked at a game from the infinite loop and I found that movei sacrificed a pawn for developement of the pieces and in the line
1.Nc3 d5 2.Nf3 movei played 2...e5

I do not think that this sacrifice is good(movei lost the game) and
I wonder if there are other programs that play this sacrifice.

Here is analysis by movei with 128 mbytes hash
Movei needs depth 16 to avoid this mistake:

Rybkav2.3.2a.w32 - Movei00_8_438
[d]rnbqkbnr/ppp1pppp/8/3p4/8/2N2N2/PPPPPPPP/R1BQKB1R b KQkq - 0 1

Analysis by Movei00_8_438:

2...Nf6
= (-0.11) Depth: 1 00:00:00
2...Nf6
= (0.19) Depth: 2 00:00:00
2...Nf6 3.d4
² (0.46) Depth: 2 00:00:00
2...d4
² (0.45) Depth: 2 00:00:00
2...d4 3.Nb5
² (0.43) Depth: 2 00:00:00
2...d4 3.Nb5
² (0.43) Depth: 2 00:00:00
2...d4
= (0.13) Depth: 3 00:00:00
2...d4 3.Nb5 Nc6
= (0.05) Depth: 3 00:00:00
2...Nf6
= (0.04) Depth: 3 00:00:00
2...Nf6 3.d4 Nc6
= (0.00) Depth: 3 00:00:00
2...Nc6
= (-0.01) Depth: 3 00:00:00
2...Nc6 3.e4 Nf6
= (-0.03) Depth: 3 00:00:00
2...Nc6 3.e4 Nf6
= (-0.03) Depth: 3 00:00:00
2...Nc6
² (0.27) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nf6
² (0.47) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Bf5
² (0.47) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nc6
= (0.18) Depth: 5 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4 Nf6
= (0.00) Depth: 5 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4 Nf6
= (0.00) Depth: 5 00:00:00
2...Nc6
² (0.30) Depth: 6 00:00:00 6kN
2...Nc6 3.d4 Bf5 4.Bf4 Nf6 5.e3
² (0.30) Depth: 6 00:00:00 9kN
2...d4
² (0.29) Depth: 6 00:00:00 11kN
2...d4 3.Nb5 Nc6 4.e3 dxe3 5.fxe3
= (0.24) Depth: 6 00:00:00 14kN
2...d4 3.Nb5 Nc6 4.e3 dxe3 5.fxe3
= (0.24) Depth: 6 00:00:00 19kN
2...d4 3.Nb5 Nc6 4.e3 e5 5.Bc4 Bf5 6.exd4 exd4
= (0.05) Depth: 7 00:00:00 28kN
2...d4 3.Nb5 Nc6 4.e3 e5 5.Bc4 Bf5 6.exd4 exd4
= (0.05) Depth: 7 00:00:00 35kN
2...d4 3.Nb5 e5 4.Nxe5 Nc6 5.Nxc6 bxc6 6.Na3 Bxa3 7.bxa3
= (0.05) Depth: 8 00:00:00 54kN
2...d4 3.Nb5 e5 4.Nxe5 Nc6 5.Nxc6 bxc6 6.Na3 Bxa3 7.bxa3
= (0.05) Depth: 8 00:00:00 91kN
2...d4 3.Ne4 Nc6 4.d3 Nf6
= (-0.11) Depth: 9 00:00:01 189kN
2...d4 3.Ne4 Nc6 4.d3 Nf6
= (-0.11) Depth: 9 00:00:01 232kN
2...d4 3.Ne4 Nc6 4.d3 e5 5.Bg5 Qd5 6.c3 Bf5 7.Qc2 dxc3 8.bxc3
= (-0.12) Depth: 10 00:00:01 360kN
2...d4 3.Ne4 Nc6 4.d3 e5 5.Bg5 Qd5 6.c3 Bf5 7.Qc2 dxc3 8.bxc3
= (-0.12) Depth: 10 00:00:02 457kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4
= (-0.14) Depth: 11 00:00:03 778kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4
= (-0.14) Depth: 11 00:00:04 939kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4 8.c3
= (-0.05) Depth: 12 00:00:10 2182kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4 8.c3
= (-0.05) Depth: 12 00:00:14 3013kN
2...d4 3.Ne4 Nc6 4.e3 dxe3 5.dxe3 Qxd1+ 6.Kxd1 Bg4 7.Bd2 Nf6 8.Nc5 0-0-0 9.Bd3 e5
= (-0.10) Depth: 13 00:00:27 5666kN
2...e5
= (-0.11) Depth: 13 00:00:43 8955kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bxf4 8.exf4 Nc6
= (-0.15) Depth: 13 00:00:49 10455kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bxf4 8.exf4 Nc6
= (-0.15) Depth: 13 00:00:52 11120kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bb4 8.Bd2 Bg4 9.Bb5+ Nc6
= (-0.13) Depth: 14 00:01:10 14989kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bb4 8.Bd2 Bg4 9.Bb5+ Nc6
= (-0.13) Depth: 14 00:01:31 19342kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Bg4 7.Nxd4 0-0-0 8.Nxc6 Qxc6 9.d4 Nf6
= (-0.12) Depth: 15 00:02:49 35274kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Bg4 7.Nxd4 0-0-0 8.Nxc6 Qxc6 9.d4 Nf6
= (-0.12) Depth: 15 00:04:07 51096kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Nf6 7.cxd4 Nxd4 8.Nc3 Nxf3+ 9.exf3 Qe6+ 10.Be2 Bd6 11.Qb3 Qe5
= (0.14) Depth: 16 00:07:19 90632kN
2...d4
= (0.13) Depth: 16 00:07:56 98478kN
2...d4 3.Ne4 e5 4.e3 Nf6 5.Nxf6+ Qxf6 6.exd4 exd4 7.Bb5+ c6 8.Qe2+ Kd8 9.Bc4 b5 10.Bb3 Bd6 11.0-0 Re8 12.Qd3
= (-0.13) Depth: 16 00:13:42 168708kN

(Uri, MyTown 07.09.2007)

[Eraserheads]

On one hand, I am pleased movei played this move. I have a liking for programs that can play speculatively, and this shows that movei is likely to play similar moves in the middlegame too.

[bob]

I looked at a game from the infinite loop and I found that movei sacrificed a pawn for developement of the pieces and in the line
1.Nc3 d5 2.Nf3 movei played 2...e5

I do not think that this sacrifice is good(movei lost the game) and
I wonder if there are other programs that play this sacrifice.

Here is analysis by movei with 128 mbytes hash
Movei needs depth 16 to avoid this mistake:

Rybkav2.3.2a.w32 - Movei00_8_438
[d]rnbqkbnr/ppp1pppp/8/3p4/8/2N2N2/PPPPPPPP/R1BQKB1R b KQkq - 0 1

Analysis by Movei00_8_438:

2...Nf6
= (-0.11) Depth: 1 00:00:00
2...Nf6
= (0.19) Depth: 2 00:00:00
2...Nf6 3.d4
² (0.46) Depth: 2 00:00:00
2...d4
² (0.45) Depth: 2 00:00:00
2...d4 3.Nb5
² (0.43) Depth: 2 00:00:00
2...d4 3.Nb5
² (0.43) Depth: 2 00:00:00
2...d4
= (0.13) Depth: 3 00:00:00
2...d4 3.Nb5 Nc6
= (0.05) Depth: 3 00:00:00
2...Nf6
= (0.04) Depth: 3 00:00:00
2...Nf6 3.d4 Nc6
= (0.00) Depth: 3 00:00:00
2...Nc6
= (-0.01) Depth: 3 00:00:00
2...Nc6 3.e4 Nf6
= (-0.03) Depth: 3 00:00:00
2...Nc6 3.e4 Nf6
= (-0.03) Depth: 3 00:00:00
2...Nc6
² (0.27) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nf6
² (0.47) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Bf5
² (0.47) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4
² (0.48) Depth: 4 00:00:00
2...Nc6
= (0.18) Depth: 5 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4 Nf6
= (0.00) Depth: 5 00:00:00
2...Nc6 3.d4 Bf5 4.Bf4 Nf6
= (0.00) Depth: 5 00:00:00
2...Nc6
² (0.30) Depth: 6 00:00:00 6kN
2...Nc6 3.d4 Bf5 4.Bf4 Nf6 5.e3
² (0.30) Depth: 6 00:00:00 9kN
2...d4
² (0.29) Depth: 6 00:00:00 11kN
2...d4 3.Nb5 Nc6 4.e3 dxe3 5.fxe3
= (0.24) Depth: 6 00:00:00 14kN
2...d4 3.Nb5 Nc6 4.e3 dxe3 5.fxe3
= (0.24) Depth: 6 00:00:00 19kN
2...d4 3.Nb5 Nc6 4.e3 e5 5.Bc4 Bf5 6.exd4 exd4
= (0.05) Depth: 7 00:00:00 28kN
2...d4 3.Nb5 Nc6 4.e3 e5 5.Bc4 Bf5 6.exd4 exd4
= (0.05) Depth: 7 00:00:00 35kN
2...d4 3.Nb5 e5 4.Nxe5 Nc6 5.Nxc6 bxc6 6.Na3 Bxa3 7.bxa3
= (0.05) Depth: 8 00:00:00 54kN
2...d4 3.Nb5 e5 4.Nxe5 Nc6 5.Nxc6 bxc6 6.Na3 Bxa3 7.bxa3
= (0.05) Depth: 8 00:00:00 91kN
2...d4 3.Ne4 Nc6 4.d3 Nf6
= (-0.11) Depth: 9 00:00:01 189kN
2...d4 3.Ne4 Nc6 4.d3 Nf6
= (-0.11) Depth: 9 00:00:01 232kN
2...d4 3.Ne4 Nc6 4.d3 e5 5.Bg5 Qd5 6.c3 Bf5 7.Qc2 dxc3 8.bxc3
= (-0.12) Depth: 10 00:00:01 360kN
2...d4 3.Ne4 Nc6 4.d3 e5 5.Bg5 Qd5 6.c3 Bf5 7.Qc2 dxc3 8.bxc3
= (-0.12) Depth: 10 00:00:02 457kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4
= (-0.14) Depth: 11 00:00:03 778kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4
= (-0.14) Depth: 11 00:00:04 939kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4 8.c3
= (-0.05) Depth: 12 00:00:10 2182kN
2...d4 3.Ne4 Nc6 4.e3 Qd5 5.d3 Nf6 6.exd4 Nxd4 7.Nxd4 Qxd4 8.c3
= (-0.05) Depth: 12 00:00:14 3013kN
2...d4 3.Ne4 Nc6 4.e3 dxe3 5.dxe3 Qxd1+ 6.Kxd1 Bg4 7.Bd2 Nf6 8.Nc5 0-0-0 9.Bd3 e5
= (-0.10) Depth: 13 00:00:27 5666kN
2...e5
= (-0.11) Depth: 13 00:00:43 8955kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bxf4 8.exf4 Nc6
= (-0.15) Depth: 13 00:00:49 10455kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bxf4 8.exf4 Nc6
= (-0.15) Depth: 13 00:00:52 11120kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bb4 8.Bd2 Bg4 9.Bb5+ Nc6
= (-0.13) Depth: 14 00:01:10 14989kN
2...e5 3.e3 e4 4.Ne5 Bd6 5.d4 exd3 6.Nxd3 Nf6 7.Nf4 Bb4 8.Bd2 Bg4 9.Bb5+ Nc6
= (-0.13) Depth: 14 00:01:31 19342kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Bg4 7.Nxd4 0-0-0 8.Nxc6 Qxc6 9.d4 Nf6
= (-0.12) Depth: 15 00:02:49 35274kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Bg4 7.Nxd4 0-0-0 8.Nxc6 Qxc6 9.d4 Nf6
= (-0.12) Depth: 15 00:04:07 51096kN
2...e5 3.Nxe5 d4 4.Nb1 Qd5 5.Nf3 Nc6 6.c3 Nf6 7.cxd4 Nxd4 8.Nc3 Nxf3+ 9.exf3 Qe6+ 10.Be2 Bd6 11.Qb3 Qe5
= (0.14) Depth: 16 00:07:19 90632kN
2...d4
= (0.13) Depth: 16 00:07:56 98478kN
2...d4 3.Ne4 e5 4.e3 Nf6 5.Nxf6+ Qxf6 6.exd4 exd4 7.Bb5+ c6 8.Qe2+ Kd8 9.Bc4 b5 10.Bb3 Bd6 11.0-0 Re8 12.Qd3
= (-0.13) Depth: 16 00:13:42 168708kN

(Uri, MyTown 07.09.2007)

It looks like your score for a "tempo" is too high. Development is good, but not _that_ good...

[Eelco de Groot]

I think it is not necessarily such a very bad move, Rybka 1.0 Beta only gives White +0.04 after 2... e5 at 20 plies deep. I think it may be playable but for my taste Black should probably try to hold on to its d-pawn, after playing 3... d4.

rnbqkbnr/ppp2ppp/8/3pp3/8/2N2N2/PPPPPPPP/R1BQKB1R w KQkq -

Engine: Rybka 1.0 Beta 32-bit (256 MB)
gemaakt door Vasik Rajlich

3.01 0:00 +0.49 3.Pxe5 (640) 40

4.01 0:00 +0.36 3.Pxe5 (1.273) 81

5.01 0:00 +0.43 3.Pxe5 Pf6 4.Pb5 (7.545) 97

6.01 0:00 +0.31 3.Pxe5 d4 4.Pb5 (17.964) 97

7.01 0:00 +0.31 3.Pxe5 d4 4.Pb5 c6 5.e3 (29.356) 96

8.01 0:00 +0.22 3.Pxe5 d4 4.Pb5 a6 5.Pa3 (50.511) 97

9.01 0:01 +0.13 3.Pxe5 d4 4.Pb5 Pc6 5.Pxc6 bxc6
6.Pa3 (141.789) 91

9.01 0:02 +0.16 3.e3 e4 4.Pd4 Pf6 5.Lb5+ c6 6.Le2 (189.177) 90

10.01 0:02 +0.09 3.e3 e4 4.Pd4 Pf6 5.Lb5+ c6 6.Le2 c5 (265.586) 92

11.01 0:05 +0.09 3.e3 e4 4.Pd4 Pf6 5.f3 c5 6.Lb5+ Pbd7
7.Pf5 (509.342) 93

12.01 0:09 +0.03 3.e3 e4 4.Pd4 Pf6 5.Lb5+ c6 6.Le2 c5
7.Pb3 d4 (848.619) 93

13.01 0:22 -0.01 3.e3 e4 4.Pd4 Pf6 5.h3 c5 6.Lb5+ Pbd7
7.Pf5 a6 (1.968.968) 93

13.02 0:27 +0.03 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Pf6
6.Pxd6+ Dxd6 7.Pa3 Pc6 (2.363.215) 90

14.01 0:52 +0.05 3.Pxe5 d4 4.Pb1 Dd5 5.Pf3 Pc6 6.e3 Pf6
7.exd4 Pxd4 (4.442.582) 87

15.01 1:28 +0.05 3.Pxe5 d4 4.Pb1 Dd5 5.Pf3 Pc6 6.c3 d3
7.exd3 Pf6 (7.467.294) 86

16.01 3:22 +0.08 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Pf6 6.c3 d3
7.exd3 O-O (16.074.568) 81

17.01 6:13 +0.05 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Pf6 6.c3 d3
7.exd3 O-O (29.051.713) 79

18.01 14:42 +0.08 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Le7 6.c3 c5
7.d3 Pf6 (66.599.389) 77

19.01 29:49 +0.06 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Le7 6.c3 c5
7.d3 Pf6 (131.674.561) 75

20.01 60:12 +0.04 3.Pxe5 d4 4.Pb1 Ld6 5.Pc4 Le7 6.c3 c5
7.d3 Pf6 (260.468.566) 73

This reminded me of a thread in the CSS-forum about the line 1. Nf3 Nf6 2. c4 e5 that is in some Opening books, Fritz Power Book 8 is mentioned. http://f23.parsimony.net/forum50826/messages/169384.htm 1. c4 e5 2. Nf3 Nf6 is more commonly played and White does not always take the pawn, as Dieter Eberle posted in the thread later.

Eelco