Author: Robert Hyatt
Date: 18:26:07 09/02/04
Go up one level in this thread
On September 02, 2004 at 16:46:49, Stuart Cracraft wrote: >On September 02, 2004 at 14:09:22, Andrew Platt wrote: > >>On September 02, 2004 at 09:45:25, Michael Henderson wrote: >> >>>>>With pruning by evaluation latest movei get it only in 10 plies but I think to >>>>>change my pruning to make it see it faster. >>>>> >>>>>Uri >>> >>>is recapture extension also playing a role here?...if that applies. >> >>Having shaken some bugs out I'm now trying to understand where else my search >>needs to be improved to get this in a reasonable time and re-capture seems to be >>a key item. I played around with Crafty a bit and disabled re-capture extensions >>(I left mate threat extensions in). Without re-capture things look like this: >> >> 9 7.09 -0.88 1. Kf1 a5 2. Rxh5 gxh5 3. Rxh5 Re1+ >> 4. Kxe1 Nxh5 5. Qg5+ Kf8 6. Qh6+ Ke8 >> 7. Qh8+ Bf8 8. Qxh5 >> 9 8.23 +1 1. Qxf4!! >> 9 8.56 +3 1. Qxf4!! >> 9 9.24 +M 1. Qxf4!! >> 9 41.38 10.88 1. Qxf4 Be5 2. dxe5 Rxe5 3. Bxe5 Qb6 >> 4. Rd1 Re7 5. Rd6 >> 9-> 41.38 10.88 1. Qxf4 Be5 2. dxe5 Rxe5 3. Bxe5 Qb6 >> 4. Rd1 Re7 5. Rd6 >> 10 51.88 +1 1. Qxf4!! >> 10 1:05 +3 1. Qxf4!! >> 10 1:22 +M 1. Qxf4!! >> 10 6:22 Mat06 1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6 >> 4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8# >> 10-> 6:22 Mat06 1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6 >> 4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8# >> >> >>compared to this with both extensions (sorry about the switch from native output >>to xboard output): >> >> 9 -88 659 4040010 1. Kf1 a5 2. Rxh5 gxh5 3. Rxh5 Re1+ 4. Kxe1 N >>xh5 5. Qg5+ Kf8 6. Qh6+ Ke8 7. Qh8+ Bf8 8. Qxh5 >> 9 -77 742 4591270 1. Qxf4!! >> 9 23 779 4824187 1. Qxf4!! >> 9 223 862 5373878 1. Qxf4!! >> 9 32756 6672 48787620 1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6 4. Rxh >>6 Qh2+ 5. Kxh2 Kf8 6. Rh8# >> 9 32756 6672 48788236 1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6 4. Rxh >>6 Qh2+ 5. Kxh2 Kf8 6. Rh8# >> >> >>Andy. > >What is the nature of Crafty's recapture extension? Pretty simple. 1. to[ply] == to[ply-1] otherwise it can't be a recapture since it isn't to the same square. 2. val(captured[ply]) == val(captured[ply-1]) so that the pieces being captured have the same value (ie bishop for knight, rook for rook, etc.) 3. no recapture at last ply, so that a long series of captures does not get extended every ply. For example rxr rxr rxr rxr. I want to extend twice. Not 3 times. > >Mine is: extend by 1 ply, unless already extended at this node, >if the last move restored the material difference between the >two sides to what it was prior to the start of the search. That last is a problem. I gave an example of how in an earlier post...
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.