Believable Stupidity In Game AI

Gamasutra is running a feature written by Mick West, co-founder of Neversoft, about creating game AI that is dumb enough to defeat, yet intelligent enough that its "mistakes" are similar to those a real player would make, thus preserving the illusion that the AI is not just throwing the game. "The simplest way to introduce stupidity into AI is to reduce the amount of computation that it's allowed to perform. Chess AI generally performs billions of calculations when deciding what move to make. ... The problem with this approach is that it decreases the realism of the AI player. When you reduce the amount of computation, the AI will begin to make incredibly stupid mistakes — mistakes that are so stupid, no human would ever make them. The artificial nature of the game will then become apparent, which destroys the illusion of playing against a real opponent. ... By reducing the amount of computation, we create an AI opponent that is trying to win, but has been crippled in a way that leads to unrealistic gameplay."

Comments

Cheating AI

[Anonymous Coward]

Actually I don't have much issue with the quality of AI's used in games, but I don't like that so many of them cheat. A few games of gotten it right, but many AI's can track my movements through walls. In order to defeat the AI, I have to figure out how it works and so I'm constantly aware it has superhuman abilities and I find that very distracting from the realism.

Re:Cheating AI

[Vectronic]

I think that's largely because AI is generally controlled as a group entity, so there is one main master pool of data that they get information from.

When what should be done, is that each AI is IAI or something, individual artificial intelligence, which can be done with object masking, and an analysis of what the AI can see in it's perspective camera, or it's range to a sound, instead of all players and sounds being a dot on a grid, with no regard for obstructions and range.

Little more on topic, I don't really mind figuring out how the AI is working, what I dislike, is like hard-coded faults, usually with waypointed bots in FPS type games, where they will always get stuck at the same spot on that same path. Because then I abuse it, i'll lead them there, wait for them to get stuck, and kill them... lotsa fun for 15 minutes, but the game gets really boring quickly, however I actually like the superhuman AI as long as they still have to abide by the rules I do (not shooting through walls I can't, etc), makes for great practice.

I generally don't play games for realism, but rather for the lack of it, I can't go out collecting coins from trees, or shooting my neighbours "really"... excluding racing/flying simulators, but usually they don't have much problems with realistic AI because of how many variables there are to "fuck with", most, if not all of which can happen in reality, sudden gust of wind, punctured tire, blown engine, etc, perhaps thats what humanoid AI games need, is more variables to be more realistic. Different eyesights, hearing, reaction times, strength, etc, etc, then slightly randomized variations on them during the same match, so that even a hard-coded fault in the AI wouldn't come to the exact same result, humans don't play by constants, why should AI.

Now that i'm rambling, I'll end with the fact that most games are multi-player now, so they spend more time working on the human interaction with the game, and the AI is just tossed in afterwards, probably carried over from v1.0, just so they can say it has that option, expecting people to want to play people. As a side note, maybe thats the logic behind some of them, make shitty AI, to try and force more people to buy+play the game so the game is useful.

Re:Cheating AI

[Nick Ives]

FC2 was notorious for the enemies ability to see you through anything, they clearly didn't even attempt to solve the perfect aim / x-ray vision problem.

The best shooter in this regard is Crysis. The enemy AI can only see you over long distances if they happen to look in your direction through either binoculars or a scope and if you can't see them they can't see you, even through bushes.

AI leaps and bounds?

[Gizzmonic]

Year after year, I read these kind of articles that report how game AI increases in leaps and bounds...and I still don't see it. Bad guys in GTA still seem to rush towards grenades, Halo/Gears of War enemies are either completely impulsive or avoidant. I'm not knocking the programmers...I think game AI must be very difficult to achieve, and even harder to detect for the layman (such as myself).

Does anyone have an example of really good AI in action games (or any non-RPG, non-RTS games)?

Competing goals

[olclops]

If you want an AI to make human like mistakes, you have to have at least a roughly human cognitive model. The simplest way to do this, it seems to me, is to give the AI competing goals. Rather than just have the AI "try to win", and then cripple its ability to do that effectively, you could give it multiple goals to strive toward, and then give it some degree of randomness in which goal it chooses to pursue. Victory vs. pain-avoidance, attack vs. finding time to recover, etc.

Re:Deep Blue

[Evanisincontrol]

This has been mentioned before, but I think it's worth bringing up that Kasparov wasn't facing Deep Blue under fair conditions. Even ignoring accusations of cheating and whatnot, there's two very important facts to consider:

• Normally when a grandmaster plays in a chess tournament, they are givens months of advance warning. The grandmaster will study hundreds of game transcripts, studying their opponent's style and looking for weaknesses. (The opponent will also be given the grandmaster's game transcripts). Deep blue was given Kasparov's transcripts, but Kasparov was given nothing.
• IBM modified Deep Blue after Kasparov won his first match. This meant that, after finally learning how Deep Blue played (on the fly!), Kasparov basically had to play a brand new opponent.

If this were a karate match, this would be the equivalent of the master having to fight someone he's never met before, but that person has studied the master's every move. Then, after the master wins the fight, he has to fight someone else who has studied his every move and acts differently. Not a fair fight.

Re:Deep Blue

[josh61980]

I read about in college. Kasparov did ask for the transcripts. However the Big Blue team told him no. Basically saying it would give Kasparov an unfair advantage because he could study Big Blue's moves and deconstruct the algorithm it used.