Author: Robert Hyatt
Date: 10:31:35 02/16/05
Go up one level in this thread
On February 16, 2005 at 11:35:39, Uri Blass wrote: >Note first that Movei of today may make stupid mistakes in the middle game >regardless of time because of null move pruning and in case of zugzwang in the >middle game it may blunder. > >In thoery it also may blunder in endgames that are not pawn endgame because of >zugzwang but in that case you need more than one zugzwang in the same line. > >It is not in the top of the priority to fix it today and if somebody want to >encourage me to fix the problem in the future then showing me games when movei >lost because of missing a zugzwang may be productive. > > > > > >If you have an engine X that may do stupid tactical blunders regardless of the >time control that are only typical for it(I am not talking about null move >pruning that is a problem of a lot of engines) then please answer the following >question > >1)What is the reason that you did not fix the stupid blunders that X >does? _Every_ program will make stupid blunders. Caused by lack of depth, wrong evaluations, incorrect time usage, you name it. The question you have to answer is "is a change going to make it play better or worse?" Often you can fix a blunder, only to make the program play worse overall. So often, leaving a problem alone for the time being is the better approach, if the fix is going to be painfully complex and probably introduce even more bugs than before. > >a)It is not easy to fix the problems and you had other priorities. >b)It is easy to fix the problems but it is not easy to do it without >making it slower and the easy solution to fix the problem means >that it will be weaker. >c)refuse to reply >d)another reply Any of the above are plausible. > >2)Do you plan to fix the stupid blunders that only X does in a >future version of X (or part of them)? > >a)Not in the near future. >b)yes,it is one of the things that is planned. >c)refuse to answer. >d)another reply. > my long-term goal is to fix _all_ bugs. But reality sometimes influences the decision of "when" and "where". >3)In case that you fix the problems or part of them do you plan to >give information that you do it before the release of new version? > >a)yes >b)no >c)refuse to answer. >d)another reply. > >Uri I generally discuss things openly, or if I fix something nobody mentioned, the comments of my revision log always explains what was changed and why.
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.