Author: Roy Brunjes
Date: 10:40:58 04/21/02
Go up one level in this thread
On April 21, 2002 at 13:15:35, Jason Repa wrote: >On April 21, 2002 at 12:37:33, Roy Brunjes wrote: > >>On April 21, 2002 at 11:47:43, Jason Repa wrote: >> >>>On April 21, 2002 at 11:39:49, Torstein Hall wrote: >>> >>>>On April 21, 2002 at 11:34:47, Jason Repa wrote: >>>> >>>>>So what is the best way to configure the Fritz "deep-positional analysis" >>>>>functions ie; evaluation window level, variation length, depth, branching factor >>>>>in order to optimize it's strength? >>>> >>>>To be honest I'm not sure. But the more time you give it, the better the >>>>analysis, for sure. >>>>How many lines it should consider at each level must be different from position >>>>to position. And perhaps the evaluation window should be set a bit higher if you >>>>give Fritz a lot of time. >>>> >>>>Torstein >>> >>>Yes, you sir have a definite flare for stating the obvious!! Now that that is >>>out of the way, is there anyone who knows the Fritz program well that >>>understands what type of modifications to the search tree should be done to what >>>types of positions? >> >>No, it is not possible for today's programs to provide optimal analysis for >>*any* type of position. And the program authors cannot possibly account for >>every possible nuance of position ahead of time with their software. Fritz 7 >>happens to be pretty darn good at a wide range of positions, but the problem is, >>without extensive time (months certainly, possibly years) spent with the program >>and a solid background in chess, you will not know apriori that Fritz 7 stinks >>at analyzing positions of type X, where X might be only very slightly different >>from position Y where Fritz 7 is very good at analyzing things and that Fritz 7 >>is superb at positions of type Z -- say, an open tactical fight with no serious >>issues of compensation [a piece and pawn down, but with chances for a strong >>attack]. >> >>It is even more difficult to recognize such situations and then propose to >>tinker with engine settings to produce a better quality analysis in those >>positions where the program is not particularly good (using its default >>settings). Better add another year to your studies for those situations.... >> >>I implore you to re-read my post(and pay attention this time). I never suggested the possibility of some "gestalt" configuration setting for all positions. I was talking about SPECIFIC modifications to the default that would be beneficial for SPECIFIC types of positions. I will provide a very simple example(that even you should be able to grasp). When you are looking for a line in which white has a forced win, you are not going to want to do branching for both black and white(tying up processor time on unnecessary white alternative lines). You will want to set branching for black only. Understand now? Yes, I understand that you are rude and not worthy of my help. Piss off asswipe >> >> >> >> >>There is no
This page took 0.01 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.