Advertisement

learning technique for fighting games

Started by August 26, 2005 12:49 AM
5 comments, last by GameDev.net 19 years, 2 months ago
I would like to know which are the simple as well as advanced techniques to implement "learning" in a fighting game.
N-grams are a good start.
Advertisement
There's always Virtual Fighter 4, which comes with an interesting mode where you can train your own NPC by beating the crap out of it. It'll start imitating your moves and run variations on them and such. Its actually quite interesting, and may present you with a few ideas.
I'm having a look at the PS2 version of Virtual fighter 4. It is a pretty neat fighting game. The AI of the game is very interesting. The oponent always learns from your capabilities and adjusts to your moves.
I was currently reading on pattern recognition used in games. A lot of people use it in their AI implementation. It seems like a good technique to be used for implementing learning in fighting games on a smaller extent. Is it advisable to use it on a smaller extent ?
'Pattern recognition' is very vague so it's hard to give you a good answer. It does sound like whatever you're reading might be helpful though.
Advertisement
Its a pretty good idea, but i think it would be better to have it get random generated strengths and weanesses (like smarter, technological, or brute and strong) and have it do a pattern recognition (like you said0 but influenced by these values, to make the opponents more individual.

This topic is closed to new replies.

Advertisement