Wow! The end of the book is here. No more late nights (for a short while at
least). I can go out and party. See my friends. I wonder how old they’ll
look. Maybe my girlfriend will even start talking to me again!
Seriously though, I hope you’ve enjoyed reading this book, have learned
from it, and are ready to start implementing the techniques in your own
games (if you haven’t started already). If you’d like to discuss any of the
topics in this book or anything else related to AI, please visit the forum at
my web site, www.ai-junkie.com.
I’d like to end the book with a few guidelines I recommend you revisit
from time to time… to keep it real, as rappers are fond of saying.
There is very rarely just one correct way of creating a good game AI
solution. Experiment with various methods as much as time permits
before committing yourself to a design.
Playtest often, and listen to your playtesters. If possible, watch them
play. Make sure you take a notepad and pen, as you’ll be using them
During your learning curve you will find yourself drawn — inevita-
bly, like a moth to a flame — to one or two AI techniques that really
make your whiskers curl. Do not fall into the trap of obsessing over
such techniques and contriving problems to apply them to. This is
akin to finding a hammer and then walking around looking for stuff
to hit with it.
Do at least one brainstorming session devoted to AI with everyone
else on your team, not just the game designer/producer (yes, even the
artists). This will result in several new and possibly exciting ideas
for you to mull over.
The design of game AI is an iterative process. There’s no way you’re
going to get it right the first time. It’s just not possible to consider all
the intricacies of anything more complex than the simplest of prob
lems, so do not become disheartened when your first attempt per
forms poorly. Persevere, learn from your mistakes, and keep
repeating the design cycle until you get it right.