Author: Johan de Koning
Date: 00:59:23 08/16/03
Go up one level in this thread
On August 15, 2003 at 14:58:05, Dieter Buerssner wrote: >On August 15, 2003 at 02:03:28, Johan de Koning wrote: > >>On August 14, 2003 at 12:12:37, Dieter Buerssner wrote: >> >>>On August 14, 2003 at 02:51:36, Johan de Koning wrote: >>> >>>>On August 13, 2003 at 17:53:30, Dieter Buerssner wrote: >>> >>>>>What are the other good reasons for clearing TTs? >>>> >>>>1. Predictability >>>> IMHO an engine should just search when a search is needed. >>>> After all it is a tool, not a living creature. >>> >>>Thanks for you answer. All more or less points, I could imagine. I thought, you >>>would give perhaps one totally different point (after all, you also use very >>>small TTs compared to other programs in your engine, if we can trust what we >>>heard here - 32 or 64 Mb in Leiden this year). >> >>To be precise: 30 * 2^20 byte. :-) >>But I don't see how that connects to a "totally different" point. >> >>>This first point, I don't get however. What has searching when needed to do with >>>clearing of TTs? > >It was just speculation. As you confirmed, you use less space for TTs than other >authors under similar conditions. I think most engines don't clear TTs afer each >move. So, you can see perhaps, how I came to my speculation. You do it >different. > >In a game of CM posted at the WB-forum (my engine was the opponent), I >recognized another thing, that also might be related. It was some sort of >fortress pos (that saved the skin for my engine). Also evalutions were posted. >CM seemed rather clever, and the score got more drawish soon (far away from 50 >moves rule, and most probably far away from a point where repetions could seem >forced from the engine point of view). My only explaination was, that you pull >the score to draw in positions, where you have a rather big material advantage, >and nothing happens to improve the pos. We speak about endgame positions, the >stage where cutoffs from TTs are most efficient. But making the score dependent >from the path does not work well together with the typical usage of TTs. But you >seem to have found a good solution for this. Perhaps, you can see, how I came to >my speculation. It seems, you do some things differently than most (all?) other >engines in a very clever way. All those points might be connected to TT, and >reasons why to clear it. Ok, I understand now why you were prepared to expect something weird. All engines are different, but some are more different than others. :-) However, the example you give has nothing to do with TT (or cleverness :-). IIRC Uri mentioned it about 2 years ago: the evaluation of The King decreases with increasing fifty_counter. This encourages it to take action after 10 moves rather than after 49 moves. It is a rather simple and old rule that applies only to the evaluation. Because of this it will fail occasionally due to perpetual postponement. Also because of this it hardly depends on the path, hence it rarely threatens TT consistancy. And never reduces hit rate. ... Johan
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.