First rule of Fight Club: Don’t talk about fight club.  Corollary: In any group of men the most annoying person will use that line during a period of silence.

But what about your game?  How do you teach it to fight?  What is the philosophy behind fighting games?


If you understand what a Markov Chain is then this article will be an easy read, if you don’t know what a Markov Chain is, then just read the words.  SInce this is a computer science kind of paper many of the algorithms are presented in words and if you read the words it is worthwhile.  Other than the Matrix presentations, most of the paper uses simple algebra for the documentation.

Think about what it takes for the avatar to fight another avatar:

  • Distance to opponent coded in terms of three real-valued features based on unit-variance Gaussians placed at 1, 3 and 5 meters,
  • 4 binary features coding which of the four sides of the agents are blocked,
  • 6 binary features coding the previous action of the opponent,
  • 2 binary features coding the physical situation of the opponent player (in air and knocked).

The paper could easily be read in a first game design class for the students to be better prepared to design games.  If you are just interested in game design then this is another reason to read it.