Author: Robert Hyatt
Date: 19:22:52 12/15/98
Go up one level in this thread
On December 15, 1998 at 14:28:01, Dann Corbit wrote: >I ran my modified Crafty on my 945 megabyte PGN file. I set values that I >though should remove *all* moves from the output file. I have maxply set to 560 >and minplayed set to 100. I just wanted to get the game and move count. Yet >there are twenty five thousand positions in the resulting book!!! >Here is the output: > >parsed 111941117 moves (1496349 games). >found 0 errors during parsing. >discarded 0 moves (maxply=560). >discarded 74172434 moves (minplayed=100). >discarded 0 moves (win/lose=0.0%). >book contains 25126 unique positions. >deepest book line was 549 plies. >longest cluster of moves was 22. >time used: 127:34 cpu 127:35 elapsed. > >Does anyone have any conceivable explanation as to how this could be possible? >I use extract to throw out any two games with the same PGN hash also. It simply >boggles my tiny little mind. Yes.. you are misinterpreting the parameters. maxply 560 says cut off any game that goes beyond 560 plies. That does nothing as I doubt you have any games that go for 280 full moves. So that parameter didn't do a thing. Then you say minplay 100. That says discard any *move* that wasn't played at least 100 times. Suppose you have 200,000 games that start 1. e4... then you will get 1. e4 in the book since it was played 200,000 times. This "minplay" doesn't say "discard the entire game if the entire game wasn't played 100 times" it says discard any single position that wasn't seen 100 times. You could do what you want by setting minplay to something like 999999999 since you don't have a billion games, no move could be played that many times, even 1. e4...
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.