Author: Fernando Villegas
Date: 09:00:36 03/19/00
Go up one level in this thread
On March 19, 2000 at 05:43:20, blass uri wrote: >On March 18, 2000 at 21:40:56, Fernando Villegas wrote: > >>Re-reading an old book by Eugene Znosko-Borovsky -How not to Play Chess- I not >>only realized that, in fact, I learned very well the lesson how not play it, but >>also I picked up this, that sounds to me like just another reason to make a >>program with dual engine, one for tactics and one for strategy. Let me quote: >>"When there is no clear forced win in sight, then you must do all that yopu >>posiibly can to streghtten your pieces, ie, your position..." >>In fact, that's the way strong players do the job: first looking for something >>tactical, then if nothing of the sort is found, looking for something quiet to >>improve his position. In each case they aply -probably unsconciously- different >>way of thinking. Programs, on the contrary, had merged in only one code -with >>excepcions I suppose: I understand that christophe's Tiger has a lot of modules >>to different situations- tactics and "knowledge" and so they does not optimize >>neither of those aspects. Or, like CSTAL, they fall in tactical deadly ravines >>due to lack of search because heavy amount of knowledge code and so some >>sloweness, or they miss the point even if they go very deep in the position, >>like happens to many fast searchers. So, I insist: why not a dual approach? You >>win time if tactics are the issue, not encumbered by positional stuff, but you >>get also better play if tactics are not relevant and so you go to the strategic >>module. Why not? >>fernando > >Because it is not so simple > >It is not clear that program can see more tactics by using a tactical engine for >half of the time relative to the original engine. > >It is not clear that using the positional engine for less time going to give >better positional game because it is possible to find good positional moves by >search. > >Uri Hi uri: Well, if it is not so clear that it is so, then also is not so clear it cannot be. Besides, I do not thing of just giving each of them half of the time. A more intelligent heuristuic to use time could be devised. Time allocation is a case for the entire game, not just a matter of move by move. Cheers fernando
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.