Author: Robert Hyatt
Date: 13:50:41 10/13/03
Go up one level in this thread
On October 12, 2003 at 11:57:40, Tord Romstad wrote: >On October 12, 2003 at 09:47:39, Uri Blass wrote: > >>If you searches checks every where in the search then by definition >>you find the draw at ply 1 if your program has stalemate detection in its >>evaluation. >> >>[D]r7/8/8/8/8/PPPP4/2QP4/k6K b - - 0 1 >> >>If it does not detect the draw at depth 1 even with checks everywhere then we >>have different definition of everywhere so you should expalin your definition of >>everywhere. > >I haven't tried it, but I am 100% sure I wouldn't get a draw score at ply 1 in >this position. >My guess is that Gothmog would need many hours to return a draw score. > >My definition of "checks everywhere" is the checks are not limited to the first >ply of >qsearch, but can in theory be generated and searched at all deeper plies of the >qsearch. In this respect, I do use checks everywhere. I do not, however, >search *all* >checks everywhere. In particular, I try to prevent the qsearch from exploding >by >generating checks at deeper plies only if the number of replies to the previous >checks >is very limited. Beyond the first few plies, I *only* search checks if *all* >the previous >moves of the opponent in the qsearch have been single-reply-to-checks. There >are >also several further restrictions (for instance I don't search checks in >positions where >the side to move has a huge material advantage). > >Another reason I cannot solve your position is that I only include checks in the >qsearch in middle-game positions, and your position is classified as an endgame >position. There is another problem. In most cases, the check will be "unsafe". IE it hangs the rook. Of course the rook can't be captured without forcing a stalemate. But the check itself looks unsound using SEE or whatever, which makes it even harder to find unless you look at _all_ checks. And that _will_ explode the tree to unbearable space. > >>extending escape to checks everywhere is not enough to search checks everywhere >>and you need also to generate all possible checks everywhere. >> >>If after some checks by the black rook you stop generating rook checks then you >>do not generate checks everywhere. >> >>I do not believe that Diep or Hiarcs or Chessmaster extend checks everywhere by >>my definition. > >I also don't think so. Extending checks everywhere by your definition is >probably >not even possible; there are almost certainly positions where even a 1-ply >search >would not terminate within a reasonable amount of time. By definition, in fact. > >Tord
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.