Author: Les Fernandez
Date: 22:05:02 09/26/05
Go up one level in this thread
On September 26, 2005 at 20:00:56, Dann Corbit wrote: >On September 26, 2005 at 19:10:39, Andres Valverde wrote: > >>Hi, >> >>I am building a book file for my little engine. It consist in a text files with >>lines like this : >> >>rr4k1/3nppbp/5np1/q1pPN3/8/2N3PP/PP3PK1/R1BQR3 b - - c0:d7e5 ce:50 >>rr4k1/3nppbp/3p1np1/q1pPP3/8/2N2NPP/PP3PK1/R1BQR3 b - - c0:d6e5 ce:50 >>rr4k1/3nppbp/3p1np1/q1pP4/4P3/2N2NPP/PP3PK1/R1BQR3 w - - c0:e4e5 ce:50 >>rr4k1/2qnbppp/p2pb3/1p2p1Pn/4P2P/1NN1BP2/PPP2Q2/1K1R1B1R w - - c0:f3f4 ce:50 >>rqb2rk1/pp2bppp/2nppn2/1N6/4PP2/2N1B3/PPP1B1PP/R2Q1RK1 w - - c0:a2a4 ce:50 >>rqb1kbnr/pp1ppppp/2n5/1N6/4P3/8/PPP2PPP/RNBQKB1R w KQkq - c0:f1d3 ce:50 >>rqb1kbnr/pp1ppppp/2n5/1N6/4P3/8/PPP2PPP/RNBQKB1R w KQkq - c0:c2c4 ce:33 >>..... >> >>The EPD lines are extracted from PGN files and the "ce:x" code is the % scored >>by side to play. The question is: what is the best opcode for the move? "bm:"? >>"cO" as a comment. I would like to make books that other people could read/write >>with any EPD tool. > >I would use pm (predicted move) unless there is a move that your engine cannot >calculate and you know to be superior -- in that case use bm. And also I would >suggest standard EPD records. Separate segments with semicolons and lose the >non-standard colons. > >IMO-YMMV. >So this: >rqb1kbnr/pp1ppppp/2n5/1N6/4P3/8/PPP2PPP/RNBQKB1R w KQkq - c0:c2c4 ce:33 >becomes this: >rqb1kbnr/pp1ppppp/2n5/1N6/4P3/8/PPP2PPP/RNBQKB1R w KQkq - pm c2c4; ce 33; > >I also think SAN is better than coordinate, but to each his tastes. Hi Dann, I am shocked that you haven't asked Andres for a copy of the book file when he is done <s>. Its not like you to pass on an opportunity to increase your database. You know who <s>
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.