Author: Shep
Date: 08:48:22 02/15/00
Go up one level in this thread
On February 15, 2000 at 09:36:18, Enrique Irazoqui wrote: >On February 15, 2000 at 09:12:25, Shep wrote: > >Hi Shep, > >Your math is interesting, but don't you think that the easiest and most accurate >procedure in the case of uneven hardware is the simple swapping of machines? >Each program would play half the match on the fast computer and half on the slow >one. I think the more uneven the machines are, the more uninteresting the results become if you're just swapping machines. Although I'm exaggerating here, imagine you have a P5-90 and an Athlon 1000. Then take two programs with a noticeable strength difference, e.g. Junior and CSTal (if Thorsten is offended, you can also take Virtual Chess II instead :). If you play using my calculations, you will get an interesting match which would probably be won by Junior like 60:40 or so. If you play without adjusting the time controls and just swap machines, you will see either of this happening: a) both programs wins their series on the fast machine 48:2 for an overall result of 50:50. This would mean boring one-sided matches and a skewed overall result. b) Junior wins 48:2 on the fast machine and loses 12:38 on the slow machine. This would mean boring one-sided matches when Junior is on the fast machine and a bunch of one-sided matches when it's on the slow machine. Either case is not too appealing, especially if you don't use autoplayer. --- Shep
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.