Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!agate!boulder!news.coop.net!news.den.mmc.com!alan
From: alan@den.mmc.com (Alan Popiel)
Subject: Re: Diplomacy
Message-ID: <1995Mar15.194758.28648@den.mmc.com>
Sender: news@den.mmc.com (News Admin)
Nntp-Posting-Host: po
Organization: Martin Marietta Astronautics, Denver
Date: Wed, 15 Mar 1995 19:47:58 GMT
Lines: 25

In article <3k4n68$8du@decaxp.harvard.edu> soss@husc7.harvard.edu (Michael Soss) writes:
>
>Have you ever played Diplomacy?  Now, that would be one incredible feat 
>to program that.  Of course, you'd need to account for alliances and 
>discussions with the computer players . . .
>
>Mike

A full-up AI opponent would indeed be quite a feat of programming.  

However, it is amazing how far you can get with very simple techniques.  
Ten or so years ago I wrote a Diplomacy move analyzer using nothing more 
than SQL and a relational database.  It was pretty good at developing two 
move look ahead sequences for capturing new supplies, while protecting 
existing supplies.  I think I even implemented simple alliances, but I 
don't recall now.  I did not pursue it because it took a lot of computing 
time, even on a VAX 780, and I needed to get some "real" work done.

Implementing the DISCUSSION aspect would be the real challenge, with all
the natural language considerations.
-- 
| R. Alan Popiel            | Internet: alan@den.mmc.com                    |
| Martin Marietta, SLS      |                                               |
| P.O. Box 179, M/S 3810    | All std disclaimers apply. Cute ones, too.    |
| Denver, CO 80201-0179 USA | Void where prohibited. Your mileage may vary. |
