home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.games.chess
- Path: sparky!uunet!stanford.edu!nntp.Stanford.EDU!928s4
- From: 928s4@leland.Stanford.EDU (Benjamin Sloss)
- Subject: Re: MChess Pro vs. ChessMachine
- Message-ID: <1992Dec28.170621.21802@leland.Stanford.EDU>
- Sender: ?@leland.Stanford.EDU
- Organization: DSG, Stanford University, CA 94305, USA
- Date: Mon, 28 Dec 92 17:06:21 GMT
- Lines: 29
-
- uzun@crash.cts.com (Roger Uzun) writes:
- >ralf@ark.abg.sub.org (R.Stephan) writes:
- >> One CM win in one game is by no means a "defeat". Here you see 30 games.
- >> And more coming...
- >
- >OK, perhaps the tourney can be declared invalid, but in fact MChess Pro
- >and the chessmachine only met in 1 game, and in that game the result
- >was decisive, The Chessmachine won the game outright. The fact
- >is, if they had met 30 times, perhaps MChess pro would have won,
- >but IN MADRID THEY ONLY MET ONCE, and MChess Pro LOST THE GAME.
-
- Statistically speaking, it is irrelevant whether MChess or ChessMachine
- won the Madrid game. Even with a rating difference of 100 points, the
- weaker player is expected to win a substantial percentage of the games, so
- you cannot begin to estimate the strength of the two programs from a single
- data point. In reality, MChess Pro (50 mhz) and ChessMachine (30mhz) are
- probably within 30 points of each other.
-
- MChess's excellent result in the 30-game match is very impressive, but
- computer-computer tests tend to exaggerate the rating difference between the
- machines by roughly a factor of 2. Applying this standard contraction
- gives MChess-Pro a 2475-2500 rating on a 486/50, which is probably
- an an accurate picture of its strength.
-
- --
- Ben Sloss Email: ben@osc.versant.com Work: (415) 329-7500 x172
- Quote: "It's all fun and games until someone gets paid."
- Transportation: '83 CB650SC (straight roads) '93 CBR 900RR (twisty roads)
- '91 XR 250L (dirty roads) '87 Mustang 5.0 (wet roads)
-