Author: Bob Durrett
Date: 09:22:34 12/21/03
Go up one level in this thread
On December 21, 2003 at 12:04:35, Jaime Benito de Valle Ruiz wrote: >On December 21, 2003 at 10:08:43, Bob Durrett wrote: > >>On December 21, 2003 at 08:53:39, Jaime Benito de Valle Ruiz wrote: >> >><snip> >> >>>because the final static >>>position of the PV does not match the score is showing anyway >> >> >>The idea that the score should match the final position of the PV is clearly >>incorrect, IMHO. > >Well... for my engine I can assure you that it is! I don't mean to pry, but I am curious to know the criteria your software uses to select the PV line to be displayed. Do you somehow identify a critical path leading to a leaf node where the leaf node chosen is no worse than any other evaluated leaf node? How can you be satisfied that the PV line displayed meets your criteria? As a user, can I "believe" your PV? Generally, I assume that the first move in any PV is most reliable and that the next move is less reliable, etceteras, due to some sort of horizon effect. For example, the evaluation of a leaf node position must be done without the normal searching process. [This is true in spite of the fact that the position evaluation is necessarily sequential because that's the nature of microprocessors, ignoring multithreading.] I would expect that leaf node evaluations would be less reliable than the evaluations of internal nodes since branches are evaluated. Bob D. >For Shredder's, if it gives a >positive score for itself, but the PV goes kamikaze and looses its queen and >another couple of heavy pieces, ending up in a hopeless position that not even a >silly Qsearch can make up, the positive score makes little sense... I think. >Have a look at the example I posted: The line shows that Shredder expects the >other player to change his queen for a pawn (????) and no compensation at all, >and the score seems close to zero! Obviously the PV is not showing what Shredder >really expects. > >The theory that Shredder is extracting the PV from the hash tables seems >completely convincing to me, especially because an old version of my engine >,that used a similar method, usually exhibited similar stupid PVs once in a >while... but could beat me. > > > >> >>My interpretation of the score has been that it is either an evaluation of the >>position being analyzed or it represents an evaluation of the position that >>would occur immediately after the recommended move is played. If the engine >>"thinks" that the move is best, then the evaluation before and after the >>recommended move is played would be the same, it seems to me. >> >>I agree that Shredder's "PV" is flawed and that is irritating. However, it >>would be a bad mistake to assume that the moves of Fritz's PV are "good." >> >>Disclaimer: I am not a chess programmer but am a user of the programs. >> >>Bob D. >> >> >><snip>> Jaime Benito
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.