Author: Bruce Moreland
Date: 01:33:54 11/08/99
Go up one level in this thread
On November 08, 1999 at 04:20:56, Jeremiah Penery wrote: >On November 08, 1999 at 04:05:45, Jouni Uski wrote: > >>To get the definitive(?) answer is selected 20 random middle game position from >>ECM suite and run all 5 minutes. After that I read NPS and finally counted >>average for program. Test was done in 450 Mhz AMD with 52 MB hash. >> >>Fritz5.32 510kNPS >>Goliath2000a 610kNPS >> >>So Goliath has beaten Fritz in pure speed and is NPS champion!! >> >>Jouni > >Most of those positions were probably tactical ones. I think Goliath uses some >very lazy-eval in these types of positions, so NPS goes way up. Try the test >using "positional" positions, and see if this is the same result. It may still >be the same thing - who knows? :) > >Jeremiah Why do people care? I can understand being concerned about speed, but this isn't speed. Here is what I mean. If you want to talk about race cars and how fast they go, you talk about miles or kilometers per hour rather than RPM's. RPM's can be increased by decreasing the gear ratio, while also potentially *decreasing* speed. RPM's is a nice number, but it doesn't mean much unless you know more about the drive train. NPS is just some number that is easy to collect. It is not an indicator of tactical speed, much less chess strength. bruce
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.