Monday, May 22, 2006

"Good Enough" Really Is

Occasionally game designers and programmers use complicated AI when it isn't really needed. Many times this complex behavior can be simulated with a lot less code and processor use. So when should developers use complicated AI?

The answer is simple: only use complex AI when the player will notice and appreciate the result. Some examples would be first person shooters; sometimes the developers spend countless hours and energy developing amazingly intelligent AI when all it does in the end is run, jump, and shoot. Simpler systems do the same thing, but the programmers wanted to show off their skills instead. Examples? I don't have any. You can't tell the difference, so unless you developed the AI, you don't know anyway.

However, I do have some examples of games that were (and for some, still are) extremely popular and addictive, but utilized some of the simplest AI ever created. Think about Space Invaders, the first real video game ever invented. The AI was also the simplest AI ever invented; the aliens simply marched across, then down, across, then down, until they eventually overpower the player and the player loses. A simple design (the simplest, really), but it created the most popular game of its time (actually, it was the only game of its time!)


Space Invaders implemented extremely simple AI.

Another example would be Pacman. The AI was quite simple really; one ghost chased the player, one stayed close by, one tried to cut him off, and the last one moved randomly. The AI could have been much more complex, implementing reinforced learning, neural networks, or any other AI systems, but simple path finding did the trick. Even if those systems had been implemented, how much difference would it have made?


Simple path finding was used to make Pacman an ageless arcade favorite.

Can you think of other examples of simple AI? Or how about an example of complex AI that isn't really realized by players? Let us know about it; your opinion counts! Click the "comments so far" link below and tell the world what you think!

0 Comments:

Post a Comment

<< Home