Author: Bob Durrett
Date: 17:37:03 04/07/04
Go up one level in this thread
On April 07, 2004 at 18:15:00, Alan Grotier wrote: > > Thanks.You answered my question.Very interesting. > > But here is a thought:having multiple engine analysis of a position > assumes one is capable of determining which amongst them has the better > understanding of the position. Well, you "kept the conversation alive" by offering another thought! : ) With Shredder or Fritz you can, in infinite analysis mode [allowing sufficient time for the findings to be sufficiently credible], set the program to not only show the move it considers best but also second best, third best, etc. Also, the evaluation score is shown for each. If you find that several moves produce similar position evaluation scores, then an exhaustive analysis would have you look at all of them. Essentially, what would do is enter these moves into the scoresheet and then let the engine examine each in turn. This process can produce a small tree of analysis lines. It is important to continue this process several moves in order to overcome the "horizon effect" problems. Sometimes it happens that an engine sees a move quite differently after examining subsequent moves. This just scratches the surface of the proceedures/methods you can use to improve the analyses. With multiple engines, what you are accomplishing [essentially] is identifying additional moves to be examined. Sometimes one engine, such as Junior, will find moves not "seen" by the other engines, assuming you are having only a few moves displayed at one time. [It is also possible to set the program to display ALL of the legal moves but you may have to wait almost forever for the search depths to reach reasonable depths.] Generally, using multi-engine you would use several engines to examine each move. Using too many engines can cause this task to become so large as to be overwhelming and not worth the effort. The database software serves several useful purposes. [I will leave it to Pawlak to elaborate if he wishes to do so.] Not least of these useful tasks which can be performed using database software is identification of moves played by other people. A thorough analysis would consider each of these too. Obviously, openings and early middlegames are what I'm talking about. Advanced methods can be used for endgames, which is probably more than you wanted to know. Bob D. > > I know I don't.I finally do when I play the game.And usually lose to > computers. > > You know,I have a better "hit" rate when I play humans.........Alan Pugilism is bad. You should not hit people.
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.