Author: Severi Salminen
Date: 09:39:58 01/26/01
Go up one level in this thread
>- Where do you perform this test? at the beginning of search? After nullmoves. And before beginning the normal search. First you test this limited razoring condition. If succesful you have D=2 instead of 3. Then you test for extended futility condition at every move(like normal futility but wider margin (about rook)). Heinz suggests not to use limited razoring or extended futility pruning if we have just extended (if we are in check, at least) for additional safety. >- Does this transform to a futility pruning case when you at at D=1? No, it transforms to extended futility pruning, described above. If you are at D=1 you use normal futility pruning. So basically limited razoring limits the depth of "bad" pre-pre-frontier nodes (D=3) to 2. If we have made this reduction we can use extended futility pruning to cut even more. >- In what cases this will be reducing the tree size that null-move did not >reduced already? I don't know whether there should be any relation netween those two techniques. But this surely makes trees a little smaller at high search depths. Try it out! I have not made any specific tests between LR and nullmoves. And like I stated in another thread: NPS goes down but speed goes up. Severi
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.