Author: blass uri
Date: 06:13:29 08/13/98
Go up one level in this thread
On August 13, 1998 at 06:29:50, Eran wrote: >>I think the User guide warn you too much about this. >>I found that Fritz5 earns significantly from time when the hash tables are full. >> >>Uri > >How did you find that Fritz5 earns significantly from time when the hash tables >are full? What did you see? How did you know? Please give me a good example. I believe the results of the nunn test will be a good example. I found in the nunn test when I did some games 1000 seconds per move on my pentium200MMX that Fritz5 has good results against Junior4.6. after 8 games the result was 6:2 for fritz5 Fritz5 often found good moves when the hash tables were full. I know Mark young did test games between fritz5 and other programs and found that fritz5 is not the best program at blitz but very good at long time control. He did a match between Junior4.6 and Fritz5 with 24Mbytes hash tables when most of the time the hash tables were full and posted the games before some months 3 hours per 40 moves on pentium300Mh the result was 6:5 for fritz5 after 11 games. The only explanation for the result is that fritz5 earns significantly when the hash tables are full. Uri >I will try your example in my Fritz5 and understand it. > >Eran
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.