Author: Andrew Dados
Date: 11:12:03 08/03/00
Go up one level in this thread
On August 03, 2000 at 07:13:38, Jouni Uski wrote: >On August 03, 2000 at 05:28:35, Martin Grabriel wrote: > >>I am putting into one file 500 positions from the following test suites: >> >>Aemis 1 + 2 = 200 positions >>ECM = 231 postions >>MATS = 23 positions >>MES (modern endgame studies) = 46 positions. >> >>total = 500 positions. >> >>Before I run the test on any of the chess engines, my question is : >>is there a possibility that there are repeated positions from the above >>cocktailing? >> >> >>By the way, i intend to use max time = 90secs for each position, which hopefully >>i can complete the run in less than 12 hours. any comments? > >Have You removed the normal 10-20% of incorrect positions as almost all >suites have? Usually there are multiple solutions, but occasionally solution >is simply wrong. > >Jouni You have some subset of ecm.epd and mes.epd suites - my files have: ecm.epd - 879 positions mes.epd - 1258 positions; (You can download those files from Dann Corbitts ftp) when I combine above testsuites (I use my own utility for that) I get: C:\Nut\EPD>perl duplic2_x.cgi aemis.epd : 100 lines, new : 100 dupes : 0 repl : 0 aemis2.epd : 100 lines, new : 100 dupes : 0 repl : 0 ecm.epd : 879 lines, new : 879 dupes : 0 repl : 0 mats.epd : 24 lines, new : 23 dupes : 1 repl : 0 mes.epd : 1258 lines, new : 1257 dupes : 1 repl : 0 unique total : 2359 Positions with errors: 0 positions without BM : 1 So if your mes file is subset of mine you have at most 2 duplicate positions. -Andrew-
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.