Author: James Robertson
Date: 18:50:28 09/24/98
Go up one level in this thread
On September 24, 1998 at 18:39:02, Brian McKinley wrote: >On September 24, 1998 at 18:15:34, Bruce Moreland wrote: > >> >>On September 24, 1998 at 16:11:54, John Coffey wrote: >> >>>It seems to me that there would come a point where attempts to order >>>the moves at every branch of the tree would slow down the tree search. >>>I could see doing this in the lower levels of the tree, but wouldn't >>>there be a point of diminished returns? I.e. if you were searching >>>N ply deep, maybe you would want to stop ordering at maybe N-3? >> >>Try it and see. But I predict that you will want to order all the way out, >>since you will be messing around out near the frontier a huge percentage of the >>time. >> >>Another idea is to stop ordering if the first few moves you try don't produce a >>cutoff. If you don't cut off, you will probably have to search all of the moves >>anyway, either that or you've proven that you have no idea how to order the >>moves in this position, so why bother? Try this and see, too, if you would >>like. >> >>>This brings up another issue: Crafty (and for all I know, other programs) >>>will generate the entire move list at each branch before searching deeper. >>>It might do some ordering too. But it might not be necessary to look at >>>all these moves, so the time spent generating some of the moves could be wasted. >> >>You are suggesting an incremental move generator. Try it and see. >> >>>My plan, right or wrong, for my Mars engine was to have different search >>>strategies at different levels of the tree. Probably most programs do this >>>already. >> >>Mine doesn't, but this might be of value, and I've thought about doing it. I >>will try it and see ;-) >> >>bruce > >I tried an incremental move generator when I first started. The nps were >significantly faster, but I ended up searching more nodes. I eventually >scrapped it in my losers chess program because move ordering was more important >with so many branches extending 15 to 30 ply, I may try it again now that I am >working on normal chess and the extensions are not as deep. > >Brian At the very least, generate captures and try out the gainful captures before generating anything else. Trying hash and killer moves before generating other stuff is very important, too. James
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.