Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: ChessGenius/Visor Prism vs Fidelity EAG/Mach III , 4.5 - 5.5

Author: Richard A. Fowell

Date: 10:18:41 01/01/01

Go up one level in this thread


The SSDF rating I find for the Mach III suggests the Mach III is rated 1993:

Rank Program            CPU   CPU Speed      Rating Var #Games win% Oppo rating
 124 Fidelity Mach III  68000 16 MHz         1993   14  2410   52%  1980

what is your basis for "[BlackElo "2265"]" for "[Black "Fidelity EAG"]" below?

Is this the USCF CRA rating for the Mach III?

Or is your machine something different from that tested by the SSDF?
Note that the SSDF rating has a very low uncertainty (14 Elo) because
of the large number of games played (2410 of them).

Other data has suggested that ChessGenius 1.2 on your machine ought
to have a rating of perhaps 2000 - 2060 SSDF.

Assuming that your Fidelity EAG/Mach III machine is the
one above, then your match results would indicate a mean performance
rating of 1993 - (16/10)*25 = 1953, with a 95% confidence bound of
14 pts (for Fidelity uncertainty) + 60 points (SSDF uncertainty at 100 games)*
sqrt(10) [variance is higher for fewer games, goes at square root of number
of games), for a total uncertainty number of 204 points, indicating that
Chess Genius rating would be in the range 1749-2157 with 95% confidence.

Note that this uncertainty (204 points) in a rating based on a match of only
10 games is very high, which is why the SSDF ratings won't even print any
ratings for a program that they don't have at least 100 games. However, better
to have some data than no data. Were you planning to continue the match? If
you played another 30 games, then the uncertainty drops to 109 points?


On December 31, 2000 at 12:28:20, Tom Glenn wrote:

>ChessGenius 1.2 / handspring visor prism ,  33 MHz
>Fidelity EAG= old Mach III engine
>All games are  Game/30 mins
>
>
>  CG  .5  1  .5  0  1  0  0  .5  0  1   =  4.5
>
>  EAG .5  0  .5  1  0  1  1  .5  1  0   =  5.5
>
>  CG threw away a clear win in game 7,  or it
>  would have been 5-5.  A couple of the more
>  interesting games below.  Only 10 games , but
>  they feel 'comparable' in strength.
>
>[Event "Interspecies  Championship"]
>[Site "Tom's basement"]
>[Date "12/31/2000"]
>[Round "Game 9"]
>[White "ChessGenius 1.2"]
>[Black "Fidelity EAG"]
>[Result "0-1"]
>[WhiteElo "2250"]
>[BlackElo "2265"]
>
>1.d4 d5 2.Nf3 Nf6 3.Bg5 Ne4 4.Bf4 c5 5.e3 Qb6 6.Qc1 Nc6 7.c3 Bf5 8.Be2 f6 9.Nbd2
>g5 10.Bg3
>Nxg3 11.hxg3 e5 12.dxe5 Nxe5 13.Nxe5 fxe5 14.Nf3 Bg7 15.Nxg5 h6 16.Nf3 O-O
>17.Qd2 Rad8
>18.O-O-O d4 19.cxd4 cxd4 20.Bc4+ Kh8 21.exd4 exd4 22.Bd3 Bg4 23.Be4 Qe6 24.Qd3
>Qxa2 25.Bxb7
>Rb8 26.Qb1 Qxb1+ 27.Kxb1 Rxb7 28.Rd2 Bf5+ 29.Ka2 Rf6 30.b4 Rxb4 31.Rxh6+ Bxh6
>32.Rd3 Ra6+
>33.Ra3 Be6+ 34.Ka1 0-1
>
>[Event "InterSpecies  Championship"]
>[Site "Tom's basement"]
>[Date "12/31/2000"]
>[Round "Game 10"]
>[White "Fidelity EAG"]
>[Black "ChessGenius 1.2"]
>[Result "0-1"]
>[WhiteElo "2265"]
>[BlackElo "2250"]
>
>1.c4 Nf6 2.Nc3 e5 3.g3 Bb4 4.Nf3 Nc6 5.Bg2 O-O 6.O-O e4 7.Ng5 Bxc3 8.bxc3 Re8
>9.Qc2 Qe7 10.d3
>exd3 11.exd3 h6 12.Nf3 d6 13.Re1 Qxe1+ 14.Nxe1 Rxe1+ 15.Bf1 Bh3 16.Bb2 Rae8
>17.Qd2 Ne5
>18.Rxe1 Nf3+ 19.Kh1 Rxe1 20.Qxe1 Nxe1 21.Bxh3 Nxd3 22.Ba1 Nxf2+ 23.Kg2 Nxh3
>24.Kxh3 Ne4
>25.Bb2 Nd2 26.c5 dxc5 27.Kg2 f5 28.Bc1 Ne4 29.Bb2 g5 30.Kf3 Nd2+ 31.Ke2 Nc4
>32.Bc1 Kf7 33.a3
>Ke6 34.h4 gxh4 35.gxh4 h5 36.Kf3 Ne5+ 37.Kg3 Kd5 38.Bg5 Ng6 39.Bf6 b5 40.Bg5 Kc4
>41.Bd2
>a5 42.Be3 b4 43.cxb4 cxb4 44.axb4 axb4 45.Bg5 f4+ 46.Kf2 b3 47.Bf6 Kd3 48.Kf3 c5
>49.Kf2
>c4 50.Ke1 c3 51.Kf2 b2 52.Bxc3 Kxc3 53.Kf3 b1Q 54.Ke2 Qg1 55.Kf3 Qf1+ 56.Ke4
>Qd3# 0-1



This page took 0.01 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.