Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Test Position - Exchange sac - some programs shoot it out...

Author: Joachim Rang

Date: 00:35:44 07/20/04

Go up one level in this thread


On July 20, 2004 at 02:44:33, Roy Brunjes wrote:

>Earlier I posted about this position with the question "Is the exchange
>sacrifice 13. ... Rxf4!? sound?"
>
>[D] r1b3k1/pp2n1b1/1qn1prpp/2pp4/5B1P/2PP1NP1/PPQ2PB1/RN3RK1 b - - 0 13
>
>Some readers thought it should be, others were not so convinced they would play
>it themselves in a game (not necessarily a vote for or against the soundness,
>just that they might not play that move themselves in a game).
>
>I'm a weak player myself, so I thought I would have the programs play it out to
>see how things went when pitting various engines against Chessmaster.  I started
>the games shown below after Black's 14th move.  The assumption here is that
>White cannot afford to decline the sacrifice as he would be a piece down.  Below
>are the first 14 moves as originally played between Shredder 8 on a 2.8 GHz
>Pentium IV and Chessmaster 9000 (Personality=SKR_JR_MIX which is my own creation
>that is derived from the SKR personality developed by Kurt and others) which ran
>when this game was originally played on my Centrino 1.5 GHz laptop.
>
>Today, to make things more interesting and to try to lend more weight to the
>argument that 13. ... Rxf4 is indeed a very strong move, I ran Chessmaster on my
>old Athlon 700 MHz machine while the other ChessBase engines (tried out one
>after the other against Chessmaster on the 700 MHz Athlon) ran on my quite fast
>Centrino 1.5 GHz laptop.  Earlier tests I ran showed that my laptop is between 3
>and 4 times faster than the 700 MHz Athlon; thus the ChessBase engines had a
>significant hardware advantage.  If Chessmaster could achieve winning positions
>aginst these engines while Chessmaster had a big hardware disadvantage, I think
>it would lend more credibility to the argument that 13. ... Rxf4 is indeed a
>very strong move (perhaps best?!)
>
>
>The first 14 moves from the original game played between Shredder 8 and CM9000:
>
>1.e4 e6 2.Qe2 c5 3.g3 Nc6 4.c3 g6 5.Bg2 Bg7 6.Nf3 Nge7 7.0-0 d5 8.d3 0-0 9.e5 h6
>10.h4 Qb6 11.Bf4 f6 12.exf6 Rxf6 13.Qc2 Rxf4 14.gxf4 Bd7
>
>Now I started Shredder 8 up on my machine (a Centrino 1.5 GHz and slightly
>faster than what Shredder 8 ran on during this ICC game the other day).
>
>Here is how that game went:
>
>15.Na3 Rf8 16.Qd2 Qc7 17.h5 Rxf4 18.Rfe1 g5 19.Nb5 Qb8 20.Qe2 Nf5 21.Rac1 g4
>22.Nd2 g3 23.fxg3 Nxg3 24.Qe3 Nf5 25.Qe2 a6 26.Na3 Ncd4 27.Qd1 c4 28.Naxc4 Qa7
>29.Ne3 Nxe3 30.Rxe3 Nf5 31.Qe2 Qxe3+ 32.Qxe3 Nxe3 33.Bf3 Nf5 34.d4 Nh4 35.Be2 e5
>36.dxe5 Bxe5 37.Rf1 Rxf1+ 38.Kxf1 d4 39.cxd4 Bxd4 40.b3 Kf8 41.Nf3 Nxf3 42.Bxf3
>b6 43.Ke2 Ke7 44.Kd3 Bf6 45.Be2 Bb5+ 46.Ke3 Bxe2 47.Kxe2 Ke6 48.Kf2 Kf5 49.Kf3
>Kg5 50.Kf2 Kxh5 0-1
>
>Hmm... Shredder 8 is no slouch as CCC readers and others well know.
>
>Next I tried Hiarcs 8 to see if its positional knowledge might help it avoid
>problems in this position (again, starting after Black's 14th move).  Here are
>the moves of that game:
>
>15.Na3 Rf8 16.Qd2 Qc7 17.Nb5 Qb8 18.Qe3 Rxf4 19.Bh3 Nf5 20.Bxf5 Rxf5 21.Rae1 a6
>22.Na3 d4 23.cxd4 cxd4 24.Qe4 Rf4 25.Qe2 Qf8 26.Nh2 e5 27.Rd1 Rxh4 28.f3 Be6
>29.Nc4 Qf4 30.Nd6 Bxa2 31.Nxb7 Qg5+ 32.Qg2 Qe7 33.Ra1 Bf7 34.Na5 Nb4 35.Qd2 Rh5
>36.Rf2 Nd5 37.Nc6 Qh4 38.Rxa6 Qg3+ 39.Kh1 Nf4 40.b4 Bf8 41.Ra5 Rxh2+ 42.Rxh2
>Qxf3+ 43.Kg1 Qxc6 44.Rxe5 Qf3 45.Re4 Bd6 46.Rxd4 Bd5 47.Rxd5 Nxd5 48.Rf2 Qg4+
>49.Kf1 Qh3+ 50.Kg1 Bg3 51.Rf3 Bh2+ 52.Kf2 Bg1+ 53.Kxg1 Qxf3 54.Qc2 Qg4+ 55.Kh2
>Nxb4 56.Qb3+ Kf8 57.Qc3 Qe2+ 58.Kg1 Qd1+ 59.Kg2 Qc2+ 60.Qxc2 Nxc2 61.Kg3 Nb4 0-1
>
>And again, Chessmaster on the much slower hardware is victorious.
>
>And now for Fritz 8.0.0.23 (same deal -- first 14 moves as before):
>
>15.Bh3 Rf8 16.Re1 Rxf4 17.Nbd2 Nf5 18.Kg2 Qd8 19.Bxf5 Rxf5 20.Rad1 Qb8 21.Qb3 b6
>22.Re3 Qf4 23.Rg1 Ne5 24.Qc2 Nxf3 25.Rxf3 Qxh4 26.Rg3 Ba4 27.b3 Bb5 28.Nf3 Qf4
>29.d4 g5 30.Re1 h5 31.Ne5 Bxe5 32.Rxe5 h4 33.Rge3 Rf6 34.Kh1 Be8 35.Qe2 Bg6
>36.Kg1 Bh5 37.Qf1 Bf3 38.Qh3 Be4 39.Qf1 Qg4+ 40.Kh2 Kf7 41.R5xe4 dxe4  and White
>is definitely in deep trouble here.  I ran out of energy at this point ...
>
>So, with three very strong ChessBase engines all having much faster hardware
>than Chessmaster, I have to believe this strongly suggests that in fact 13. ...
>Rxf4 is a very strong and sound move.
>
>Of course shoot-outs like this are not definitive, but short of hiring a GM to
>analyze this I do not think it is easy to know for sure.  I do not think any GMs
>read CCC, but if you are out there, go ahead and take a shot at it please!!
>
>I also believe this might make an interesting test position for engine
>developers...
>
>Regards,
>
>Roy


thank you for your definitely interesting test. It seems it is indeed an
interesting testposition.

regards Joachim



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.