[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Bug-gnubg] GNUBg Position ID question
From: |
M. J. Mannon |
Subject: |
[Bug-gnubg] GNUBg Position ID question |
Date: |
Tue, 15 Jan 2013 11:31:20 -0800 (PST) |
Michael Petch is right (as usual).
I researched this extensively a couple of years ago. Attached are the positions
I tested then. Below is what I concluded:
After downloading the latest version of GNU Backgammon, I played a game, and
ran the first 7 or 8 consecutive moves through my BackgammonDriver.exe program.
The results are attached as a text file. At the top, I have inserted board
positions from the GNUBG command Edit->Copy as->Position as ASCII. At the
bottom, I have placed the positions as reconstructed in my program from the
GNUBG IDs. The positions match exactly.
My program checks the MatchID dice owner first. It then decodes the PositionID
extracting checkers for the player who does NOT own the dice, followed by the
checkers of the dice owner. As you will see when you have time to examine the
file, this interpretation allows for perfect duplication of the GNUBG
positions. No hassles. No need to reverse the board.
In the test, Player X (bottom) gets stuck on the bar. This makes it
particularly obvious that when the dice owner changes, the PositionID is
reconstructed from the opposite player's perspectcive. You can see this because
the PositionID changes radically when O gets the dice, even when X has not
moved a single checker! By following the action through the double/take that
happens early in the game, you will see that the PositionID does not change
when the turn owner changes (and the dice owner is held constant).
I hope this information is useful to you.
- Mike
GnuBg-MusicLover_0p_2011-01-25, Game 1.txt
Description: Text document
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Bug-gnubg] GNUBg Position ID question,
M. J. Mannon <=