Awari Solved

Gerard Jendras sent in a submission about applying computing power to an ancient game. The game of Awari has been solved: with perfect play, the game always results in a draw. There is a Java applet to test your skills against.

Comments

Important Step?

[rockmuelle]

From the article:

"The research is an important step forward in a research area within Artificial Intelligence, to solve games with increasing complexity"

I don't quite understand why a big lookup table is an important step for AI. Humans don't play games by checking every possible move and picking the best one and never will.

The AI community really needs to stop looking for tricks that allow computers to solve problems in ways that humans never could and instead spend their time trying to understand how intelligence actually works.

Hint: scrap predicate logic (and in doing so the Turing machine) as the model for intelligence. Instead, define a model from which predicate logic can emerge (Reginald Cahill has more or less done this, but I'm not sure if he realizes it yet: Process Physics.).

-Chris

depends on what you call perfect

[Eric Seppanen]

There are lots of games where you can create a "perfect" player that can do as well as possible against other "perfect" players.

The thing that's interesting is making a program that plays as well as possible against imperfect players, as demonstrated by the RoShamBo Programming Competition.

Kind of Bummed - Just Brute Force

[zetetikos]

I'm kind of bummed that this solution is by enumerating every position, rather than some kind of huristic or mathmatical solution. I don't find brute force methods to be very elegant or interesting, although they do present their own chalenges from a resource management perspective. I'll be much more interested if they can analyse the information they have and come up with a computational approach that plays perfectly. It's likely that such a thing could then be generalized to solve many other types of problems.

Zetetikos