Author: Carlos del Cacho
Date: 13:50:55 09/27/00
Go up one level in this thread
On September 27, 2000 at 13:02:49, Vincent Diepeveen wrote: >On September 27, 2000 at 12:10:00, Carlos del Cacho wrote: > >>On September 27, 2000 at 09:26:45, Robert Hyatt wrote: >> >>>On September 27, 2000 at 07:47:18, Bas Hamstra wrote: >>> >>>>Supposing no "lazy-errors" at all were made, does anyone know if there are >>>>serious side-effects to lazy eval? >>> >>>None at all. Except that guaranteeing this is a bit hard. :) >>> >>> >>> >>>> >>>>I one experiment that I yet have to repeat, it seems that NPS increases, but >>>>Depth (as averaged over 300 wac positions) does not. >>>> >>>>I would like to know if others have seen alike or other problems with LE. >>>> >>>> >>>>Regards, >>>>Bas Hamstra. >>> >>>It has been used forever. As a classic trade-off between speed and accuracy. >> >>Just a question related with this. Since I implemented lazy eval in my program I >>don't store the value returned by search in the hash table when there's a fail >>high or a fail low. I just store beta or alpha instead. Is this correct ? >> >>Thanks in advance, >>Carlos > >No it isn't correct as the hashtablestores the moves before that will >pick up the score and store it. > >Apart from that you will give cutoffs with lazy evaluation in positions >where not lazy evaluating will not give a cutoff. > >Now last but not least the more efficient your search will get the less >profit you will get from transpositions as you don't store them! Now that may be a great reason for not getting past 7-8 plys on quiet positions. :) I thought that I couldn't do this beacuse I would experience the side effects that are the subject of this thread. Okay, thinking over it a bit now I see that If I get there again, lazy evaluation will work again so I will be getting the same results (slower...). So I am just storing root alpha and beta all over the tree? I'll correct this. Another thing. I'm using it after evaluating pawn structure (because of passed pawns) and some primitive trapped bishop / knight code. An idea... Would it be theoretically sound to forget also about evaluating pawn structure if too many pawns/pieces on the board? What I'm suggesting is to have two points where lazy evaluation would get activated, one of those disabled in endgame/late middlegame. Mmmm... I don't think this would help much since I'm already getting a lot of hits from the pawn hash table but I wouldn't have to bother in converting my char vectors into real bitboards. Suggestions ? Ideas ? Greets, Carlos
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.