Author: Peter Fendrich
Date: 14:37:35 08/12/04
Go up one level in this thread
On August 12, 2004 at 11:35:59, Tord Romstad wrote: >On August 12, 2004 at 10:52:36, martin fierz wrote: > --snip-- >A brief summary of the pruning systems discussed, sorted by the type of >node where they are useful: > > > +-----------------------+-----------------------------+ > | Fail high nodes | Fail low nodes | >+----------------------+-----------------------+-----------------------------+ >| | * Recursive null move | * Reduced-depth pre-search | >| Big remaining depth | * ProbCut | of moves late in the move | >| | * MultiCut | in the move list | >+----------------------+-----------------------+-----------------------------+ >| | * Futility pruning | * Static null move pruning | >| Small remaining depth| * Razoring | | >+----------------------+-----------------------+-----------------------------+ > > >Additions to the table are welcome, particulary for the upper right hand >corner. > >Tord A neat table! Of course Hash and EGTB probes will fit in all 4 squares and probably doesn't add anything useful here. Maybe the same goes for "static knowledge pruning" like mate, stalemate, 3-repetition and certain known end games. A third row could be Quiescense: +----------------------+-----------------------+-----------------------------+ | Quiescense | * Evaluate pruning(?) | * SEE pruning | | | | | +----------------------+-----------------------+-----------------------------+ Not sure what to call the FH node thing. If a stand pat would FH I don't even need to try the captures. Similar tables could be done for extensions and reductions... /Peter
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.