Human games

Posted by Jäger on 2004-01-30 14:38:30

World-readable

Bitscape, Yanthor, and I played four games on my new map Wednesday and Thursday nights and generally enjoyed ourselves. Yanthor had some comments about balencing the map that I'll contemplate and incorporate into a future release.

Yanthor's comments about balencing caused me to contemplate various mechanisms to test various ways a map might compensate for different features. It might be interesting to test a map by staging numerous AI games and statistically analyzing the outcome. If one position wins significantly more frequently than another, then it needs to be adjusted. Ideally this would be an automated process -- I click "go", go to work, and come back to find the results of two hundred AI games. The closest I could get, though, would be to put myself in observer mode and watch single games at 800% speed. (Or is 1600% the maximum?)