Author: Georg v. Zimmermann
Date: 10:40:21 11/21/02
Go up one level in this thread
Preface: I just invented "nmt" = not much text. Patent pending. Hi Ed, that was the funniest post I have read in a while. Bit bored :-). Congratulations for finnishing your project. For a first program after a long pause this was an extremely interesting topic. I hope to see some analysis results soon. I really wonder what the biggest score difference etc is. Kind regards, Georg v. Zimmermann On November 21, 2002 at 08:33:41, Edward Seid wrote: >----- cross-posted on Winboard Forum ----- > >Perhaps the Subject is a bit over-enthusiastic. Let me explain. > >I've been thinking about FischerRandom lately. There are 960 possible starting >positions. If you turn off the opening book, and let your favorite engine >ponder the normal opening position, the evaluation will probably stay within +/- >0.25 (maybe even a smaller range). In other words, things are about equal. > >Have you ever wondered what the results would be if you let Crafty or some other >strong engine work on each of the 960 possible FischerRandom positions? Would >the evaluations stay close to 0.00, or are there some positions that are much >better for White (or Black?). > >Wouldn't it be neat if you could place an engine on auto-pilot to crunch through >each position for a fixed amount of time? Some engines already have this >feature. All you need is an EPD file of all the positions you want to crunch. > >Does such a file exist? A file with all 960 FischerRandom starting positions? >If there wasn't before, there is now :) > >First, I'd like to say that I'm not a programmer. You say C, I say "See what?" >You say bitboard, I say "I'm a bit bored too." But I've always been keen to >join the ranks of the elite that hang out in the computer chess forums. So this >is my first try at real programming (except for a little Basic on a TRS-80 in >high school). > >I found an algorithm to create the FischerRandom starting positions on >http://www.chessvariants.com. It's fairly straightforward and involves several >rolls of a 6-sided die. The EPD specifications can be found at >http://pgn.freeservers.com > >With these two reference points and "Active Server Pages for Dummies", I set out >to write a program to generate an EPD file of all 960 FR positions. > >My language of choice was VBScript because I can immediately use it to makeover >my website with ASP and it builds a good foundation for learning Visual Basic. > >The fruits of my labor are online to be shared with everyone. Visit my website >at http://www.winboard.info and click on the link to the EPD. The page uses >client-side scriping, so your browser should be VBScript-enabled. If you "View >Source", you can see the entire code in the Script section. Feedback is >welcomed. > >A helluva lot of work went into this little program. I had a lot of bugs, >including some faulty logic in coding the algorithm, but one by one, I solved >them all and have a great feeling of accomplishment. > >The next planned project will be an addition to my website... a searchable >database of Winboard engines. > >Let me know what you think. I'm glad I can communicate talk with you all, >programmer to programmer :) > >PS If anyone has a spare PC, you can try letting Crafty or Yace crunch through >the EPD. At 3 minutes/move, it will take 2880 minutes = 48 hours. Let me know >what you find.
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.