Hi,
Exact format to be decided, but one approach will be through a "reconstruction syntax" that lets the user annotate degrees of freedom directly in the .lud descriptions. This will be the primary mechanism for generating the reconstructions that will be a major part of the DLP.
By running the random game generator several times and observing how often it threw up a "playable" game.
We've already implemented the equivalents of those metrics (see the "Game Evaluation" dialog) but they're not very useful for random game generation. You'd be waiting a looong time between rule sets that satisfy this test.
Regards,
Cameron
Quote:I am wondering what sort of control you are thinking of here.
Exact format to be decided, but one approach will be through a "reconstruction syntax" that lets the user annotate degrees of freedom directly in the .lud descriptions. This will be the primary mechanism for generating the reconstructions that will be a major part of the DLP.
Quote:I am curious how you arrived at that figure.
By running the random game generator several times and observing how often it threw up a "playable" game.
Quote:Maybe if you added the following criteria:
* A game that goes on for at least X moves or of the trials that end in less than X moves, each player wins at least Y% of the time.
I think that would filter out a lot of degenerate games.
We've already implemented the equivalents of those metrics (see the "Game Evaluation" dialog) but they're not very useful for random game generation. You'd be waiting a looong time between rule sets that satisfy this test.
Regards,
Cameron