Author: leonid
Date: 03:56:23 11/11/99
Go up one level in this thread
On November 11, 1999 at 05:23:24, Jeremiah Penery wrote: >On November 10, 1999 at 22:52:34, leonid wrote: > >>On November 10, 1999 at 21:14:09, Jeremiah Penery wrote: >> >>>On November 10, 1999 at 20:41:52, leonid wrote: >>> >>>>Hi! >>>> >>>>For some time I am trying to find at what speed usual logic goes in the chess >>>>game. Can somebody indicate the average time that game ask for solving the >>>>position in the middle game? To be more easy for me to undertand your number, >>>>please say it for the search done by frute force. No extensions. The most >>>>explicit will be the time of search executed 10 ply deep. If you could indicate >>>>some concrete position it will be appreciated. >>>> >>>>Thanks, >>>>Leonid. >>> >>>The time you ask for will be different for every position, and for every >>>program, especially since all of them do extensions. >> >>Real mess! > >Yes. :) > >> But how the raw speed of chess logic could be found? > >In the way you're suggesting, I don't know that it can. > >> By solving the >>mate containing positions I can not guess about the speed of my logic compared >>with some other. In mine, for solving the mate positions and for every other >>move, two distinct logics work in independent way. For mate solving logic speed >>was easely recognized. > >I would worry more about playing strength rather than speed. If you get 5 NPS >and depth 2 in a 40/2 game, but still beat everybody, is that not better than >getting 1,000,000 NPS and depth 15 but losing? :) > >Jeremiah I agree with you. Only when you write your engin you would like to know until what moment you should work on the raw speed of its certain part before going to the next chapter. The same is for soving the mate containing position. Once you found that your logic goes at the best speed possible, when seeing all the possibilities for finding the mate in the shortest number of moves possible, later you elaborate some kind of speedy logic. This speedy logic will look for mate in some partial but practical way for the most part of your game. Go back to some previous part of your logic is never desirable. With a "fresh" memory you do better your work. Leonid.
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.