Author: John Merlino
Date: 13:20:31 01/20/03
Go up one level in this thread
On January 20, 2003 at 15:24:15, Robert Hyatt wrote: >On January 20, 2003 at 13:58:19, John Merlino wrote: > >>On January 20, 2003 at 13:11:48, Eiko Bleicher wrote: >> >>>Hi, >>> >>>AFAIK the endgame tablebases like the Nalimov Tablebases use one byte per >>>position. Thus, the max. length results in somewhere about 126 moves to win / >>>loose. >>> >>>Now the question is this: How do these databases handle the positions that are >>>mate in 130 or 140 or .... >>> >>>Any help would be greatly appreciated. >>> >>>Eiko >> >>First of all, it's an unsigned byte (8 bits), so it can handle mates up to 255 >>(or 254, if the 255 value is used for something else, about which I am not >>sure). I'm not exactly sure how Nalimov handles the 255 value. > >It can't go that far. Because it must have mate in N _and_ mated in N. For >the 1-byte format, 126 is the max... > > >> >>But, in Chessmaster 9000, which also uses an unsigned byte to store this value, >>all mates greater than or equal to 254 have the same value -- 254. Therefore, if >>more than one move has a value of "Mate in 254", then any one of these moves may >>be chosen as "best play". This may result in less than optimum play for two >>reasons: >> >>1) One of these moves may be Mate in 254, and the other(s) may be Mate in much >>more than 254. This isn't necessarily a bad thing, except for... >>2) One of these moves may be a mate that will not be subject to the 50-move >>rule, and the other(s) all will be draws with best play under the 50-move rule. >> >>jm > >If you do that, how do you possibly handle "mate in N" and "mated in N"??? Ok, I remember the main difference now between Nalimov and deKoning (in this case). Johan does not store "mated in N", but rather just a generic "loss" value. So, although I'm not exactly sure of the actual values used, I would guess that 1-254 are "Mate in N", 0 is "draw", and 255 is "loss". To make sure, I checked Johan's FAQ on his EGDB generator, and it says: - Nalimov's format has KxyKz and KzKxy merged. Eg in KRBKQ the wins for white (with RB) are stored as well as the losses for white. In Thompson's and De Koning's white can only win or not win. jm
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.