Author: Dave Gomboc
Date: 22:56:08 10/08/00
Go up one level in this thread
On October 05, 2000 at 00:03:34, Robert Hyatt wrote: >On October 04, 2000 at 11:44:15, Ricardo Gibert wrote: > >>On October 04, 2000 at 10:22:10, Robert Hyatt wrote: >> >>>On October 04, 2000 at 07:54:00, Ricardo Gibert wrote: >>> >>>>On October 03, 2000 at 18:18:25, Robert Hyatt wrote: >>>> >>>>>On October 03, 2000 at 12:34:11, Uri Blass wrote: >>>>> >>>>>>On October 03, 2000 at 11:58:05, Robert Hyatt wrote: >>>>>> >>>>>>>On October 02, 2000 at 13:54:23, Mike Adams wrote: >>>>>>> >>>>>>>>The move that my freind said crafty considered was the pawn takes knight move >>>>>>>>toward the end. Knight takes bishop earlier wreked the positon and i dont think >>>>>>>>pulsar would get itself into that kind of postion again with the current >>>>>>>>evaluate. >>>>>>> >>>>>>> >>>>>>>taking the knight isn't particularly bad. NxB opens a file against your own >>>>>>>king. PxN (I assume that is the knight you are talking about taking) is a >>>>>>>reasonable move for white. He wants that open file. >>>>>> >>>>>>Mike did not mean to that knight. >>>>>> >>>>>>He meant to taking the knight by gxf3 in the following line >>>>>> >>>>>>[D]r1bq1rk1/ppp2p2/2n4p/3pP1pn/3P4/2P2NB1/P1P1B1PP/R2QK2R b KQ - 0 14 >>>>>> >>>>>>Nxg3 hxg3 g4???? Qd2 gxf3???? >>>>>> >>>>>>avoiding gxf3 is a simple tactics. >>>>>>Crafty17.11 can see at depth 7 that gxf3 is bad. >>>>>> >>>>>>The bigger problem is to avoid g4 and crafty17.11 needs depth 9 for it. >>>>>> >>>>>>Fritz5.32 can avoid gxf3 at depth 5 and it can avoid g4 also at depth 5. >>>>>> >>>>>>Hiarcs4 can avoid gxf3 at depth 3. >>>>>>Faile 1.4.4 can avoid gxf3 at depth 6. >>>>>>CometB02 can avoid gxf3 at depth 5. >>>>>> >>>>>>All these programs that are not the best programs need few seconds or less than >>>>>>1 second on a slow hardware to avoid gxf3. >>>>>> >>>>>>I think that you need to improve the tactical ability of your program and >>>>>>working on positional knowledge is less important. >>>>>> >>>>>>Chess is a tactical game. >>>>>> >>>>>>Uri >>>>> >>>>> >>>>>Yes... but the tactical ability improves automatically. You should try it on >>>>>my quad xeon. It fries tactical things there. :) >>>> >>>>Crafty is a good program, but perhaps the way you should be looking at this is >>>>maybe you would write a better program if you were running on slower hardware. >>>>Fast hardware can cover up a lot of deficiencies of a program as you have >>>>essentially pointed out. >>> >>>Why design for obsolete hardware? Within a year or two everybody will be on >>>hardware faster than my quad xeon (which is probably about the speed of a >>>1.5ghz machine or so). With only so many hours available for working on Crafty, >>>spending time on the parts that new hardware won't fix makes more sense, IMHO. >> >>I'm sure there is some truth to what you are saying, but if Crafty needs to >>search to a greater depth to find the same tactics, isn't that a sign of some >>type of inefficiency? > > >Not to me. The issue is whether it finds it or not when playing a game. Not >whether it needs a 10 ply or 12 ply to find it... I don't buy this -- if Crafty requires a 12-ply search where other programs require a 10-ply search, then positions will also exist where Crafty requires a 16-ply search where other programs require only a 14-ply search. Ceteris paribus, of course. Dave
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.