Computer Chess Club Archives


Search

Terms

Messages

Subject: Rebel 10 Anti-GM, BS2830

Author: Ralph E. Carter

Date: 12:33:24 10/11/98


To those interested in Rebel 10:

ALL "RATINGS" QUOTED HERE ARE MERELY NUMBERS OUTPUT BY BENCHMARKS.

NOTES:
1: PII 300, 64M, 60M hash under DOS.
2: The Rebel 10 tested is the temporary, downloadable version.
3: Rebel 10 has an option "COMBINATIONS". It has no effect in games, but is
activated automatically to solve problems. I am not interested in this function,
and set it OFF FOR THESE TESTS.
4: Rebel 10 can test itself automatically using BS2830 (and other EPD test
suites). It uses the old rating formula, which gives a higher number. I used the
new formula.
5: The rating it gives itself after sensing the hardware is "2713".
6: The rating it gives itself after running its built in benchmark is "2529".

RESULTS:
Rebel 10 Anti-GM off    2551
Rebel 10 Anti-GM Smart  2524

CONCLUSIONS:
TACTICALLY, during actual play, this difference is probably not noticeable.
POSITIONALLY (or strategically, or stylistically) Anti-GM is still an
interesting mystery.

I don't want to be MERELY torn to smithereens tactically by computer programs.
They can all do this. I bought Rebel 10, hoping for the most chess intelligence.
Maybe it has this, I don't know yet. But here's a bonus: it is also tactically
(according to BS2830) only a meaningless 3 points below "that other program" I
have, which is supposedly the best calculator.

Since they played Anand with Anti-GM on, I think I will leave it on. I think of
this as a way to measure myself against Anand.






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.