Author: Uri Blass
Date: 03:21:18 06/15/02
Go up one level in this thread
On June 15, 2002 at 05:18:14, Peter Berger wrote: >The games were played with 30 minutes sudden death (blitz) time control; match >was scheduled for 12 games. > >Genius used 3 MB for hashtables. The Palm Vx has been overclocked with >AfterBurner - Tiger displays Speed index of 1.10 for this device. > >The hardware advantage for Genius in this match should have been somewhere >around factor 4-6. > >An interesting follow-up IMHO to earlier matches against strong amateur programs >(Yace,Comet) with same hardware difference that had much closer results and >seemed to make sense. >Also previous longer matches against some of the weaker WinBoard engines on >Athlon1333 showed consistent results in the 2100-2200 rating region. > >In the match against Genius 4 Chess Tiger was without any chances though . > >Looking at the games themselve gave a similar impression. Even when Tiger >managed to reach good positions sometimes it was always Genius who had the >better answers, especially when it is about tactics, but its play also looked >much sounder. >In fact most of the time Tiger even displayed higher depths but the Genius >search seemed to be much more efficient. > >SSDF lists Genius 4 with > >85 Genius 4.0 DOS 486/50-66 MHz 2265 23 -23 900 49% 2270 > >Taking into account past callibrations of the rating level this is definitely >very strong. But I would have expected Chess Tiger on the Palm to score better >if it wants to reach 2100 SSDF level. >I think Genius is clearly underrated on this hardware btw. > >Is Genius the very strongest program on the 486? Well possible. > >Or is it Tiger itself?! > >62 Chess Tiger 11.8 Pentium 90 MHz 2382 43 -43 261 50% 2383 >69 Genius 5.0 DOS Pentium 90 MHz 2329 18 -18 1558 47% 2348 > >In fact I assume that Genius profitted much less from the 486 to Pentium change, >but that is guesswork. Even if Tiger earned more from the p90 you should remember that it is an old tiger and the latest tiget is supposed to be significantly better. I guess that Tiger is clearly better than Genius even on 486 if you use 120/40 What happens in G/30 is a different question. It is also possible that a match between the 2 programs overestimate the difference and it may be interesting to test both of them against other opponents for example movei on faster hardware(you can give movei the white color in 20 games when you use option c and movei will changes it's choice of the first move in every game). Uri
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.