Author: Andrew Williams
Date: 01:07:38 08/28/00
Go up one level in this thread
On August 28, 2000 at 00:22:15, Will Singleton wrote: >On August 27, 2000 at 17:55:23, Andrew Williams wrote: > >>On August 27, 2000 at 17:41:07, Dan Newman wrote: >> >>>On August 27, 2000 at 08:54:51, Brian Richardson wrote: >>> >>><snip> >>>> >>>>One awkward thing for me right now relates to predicted moves. Generally, if >>>>Tinker ponders for longer than the normal next move search time and the >>>>predicted move is made, then Tinker moves immediately. Unfortunately, the score >>>>and PV from the pondering are lost and bogus values are reported from the >>>>abnormally short search. This is annoying when reporting, and a real problem >>>>when going back to do position learning. I am thinking of just saving the >>>>pondering search results separately for these cases. Any other suggestions? >>>> >>> >>>What I do for pondering is just do a normal search with the time limit set >>>to "infinite". Then when the opponent's move comes in I either break out of >>>the ponder search and start a new search (in case his move is different from >>>the predicted move), or I just set the time limit and continue searching >>>without breaking out of the ponder search. If the time has already expired, >>>the search will immediately terminate as it ordinarily does when it runs out >>>of time, with the full PV and so forth retained. Of course the code that >>>does all this is one of the ugliest parts of my program, very difficult to >>>debug, and I don't entirely understand it :). >>> >> >>Everything you say here applies equally to my program too. > >Ditto for me. That code took the longest time to get right, and I never want to >look at it again. > > >> >>>>The situation is similar but worse when there is only _one_ legal move, which >>>>Tinker makes immediately. I was reporting the raw root eval(), but this caused >>>>Tinker to resign a KNNvK game (which Tinker had seen was drawn according to >>>>EGTB's). I suppose EGTBs should be checked too. >>>> >>> >>>Mine does this too. I end up without any ponder move for the next go, so my >>>program just sits there waiting... I plan to add somthing that will find a >>>ponder move when there is none, but just haven't gotten around to it. >>> >> >>If there's only one legal move, PostModernist just does a depth 4 search, >>then returns. This way I get something to ponder. >> > >Yeah, I should do that too, would take a minute or so to program. But I can't >see it helping that much. Yes. I've not evaluated it in any way, but I can't imagine that the move I get is particularly impressive. I just feel better if my program is actually doing something rather than just sitting there. Andrew
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.