Author: Larry Griffiths
Date: 17:25:14 02/13/01
Go up one level in this thread
On February 13, 2001 at 08:55:23, Carmelo Calzerano wrote: >On February 13, 2001 at 06:09:02, Larry Griffiths wrote: > > >>It looks more like this, Bob. >>BestScoreSoFar=alpha >> >>for(do capture moves) > >[snip] > >> return BestScoreSoFar; > > >(It seems to be ok) > >>>>I see all three types added to my hash table, but it seems that the >>>>upper_bound entries never cause any cutoffs to occur when searching >>>>the hash table. >>>> >>>>If I do not add lower_bound hash table entries, then I see some >>>>upper_bound cutoffs occuring. It acts like the upper_bound entries >>>>get overlayed so often by lower_bound entries so that they never >>>>get seen when searching the hash table. >>>>Is this normal, or is something wrong with my code? > >I guess it's normal, although you should definitely see some upper_bound >cutoffs from time to time. Matter is that the number of fail-high nodes >(which stores lower bounds in the HT) is usually about one order of magnitude >_bigger_ than the number of fail-low nodes visited in search; so upper_bound >HT entries are often replaced by lower_bound ones. > >So it should be normal, IMHO, if lower_bound cutoffs occur more often (say >10 times, for example) than upper_bound ones; but the number of upper_bound >cutoffs should not be zero anyway... >:-) > >Bye, >Carmelo Thanks Carmelo, My program performs about the same with, or without this hash table. It hardly makes any impact on endgame positions. Seems like the only thing that really works well in my program are my killers. Larry.
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.