How does A.I. work in sports games

[darkchrono]

How does Artificial Intelligence work in sports games. How do programmers fit Artificial Intelligence into a game.

On my college football game that I like to play I was having trouble with the option that the computer team was running (believe it was Texas). Then as the game wore on I noticed how some of my defenders (not controlled by me) began to react to it to try to shut it down.

When the qb started trying to run the option my computerized Defensive tackle would release off his block and run down the line to the hole that the qb was running through earlier in the game when the option was working. This caused the quarterback to slow down a little bit and try to run further down the line where he could find another hole. Because he wasn't able to hit the hole so quickly my other defenders were able to catch up with him and stop him before he was able to get to many yards.

This was all done through artificial intelligence and the computer adjusting to shut something down. So when they create sports games how do they put something like that in the game (what do they put in to make it look like the computer is actually thinking even though we are not able to do that yet with computers)?

[sdm42393]

Games like Madden Football (http://en.wikipedia.org/wiki/Madden_NFL), Earl Weaver Baseball (http://en.wikipedia.org/wiki/Earl_Weaver_Baseball) and Tony La Russa Baseball (http://en.wikipedia.org/wiki/Tony_La_Russa_Baseball) all based their AI on an attempt to duplicate on the computer the coaching or managerial style of the selected celebrity. Madden, Weaver and La Russa all did extensive work with these game development teams to maximize the accuracy of the games. Later sports titles allowed users to "tune" variables in the AI to produce a player-defined managerial or coaching strategy.

[Yar]

AI can sometimes be a joke in sports games. If on NHL 08 I can win 37-0 with Columbus, there's a problem.

[darkchrono]

AI can sometimes be a joke in sports games. If on NHL 08 I can win 37-0 with Columbus, there's a problem.

How hard of level do you play on? At least for the well made sports games the computer on the harder levels is just as tough as any human player (accept for those guys who play it 50 hours a week online).

[demondude]

I don't know much about artificial intelligence, I know that it gives number values to determine what the best course of action is, but in sports games it's probably a lot more complicated.
For instance in a naughts and crosses game, the computer would put its marker on every square, if the square would lead to a loss, it would give it (-1) But if the square would win it for the computer it would give (+1), if it has no real effect on the outcome it will be (0), the computer would go for the highest value. (+1)
That's what it says in my science encyclopedia anyway, if that's the basis, then it sort of branches out in different directions I guess, but I have no idea how it works in a more complex game.

[Captain Maxx Power]

It really depends on the type of AI that's implemented. The most common form is what's known as a Finite State Machine. Basically each player has a series of States they can be in, Transitions that define which States they can switch to, and Events which determine when these Transitions occur. A State could be something like "Get the ball", "Dribble the ball", "Pass the ball", "Get into range of ball" etc. Internally these would be programmed as behaviours that are switched between if certain variables (the aforementioned Events) are true. Scaling the AI is therefore a case of slowing down the decision making of the computer by implementing a time delay between Transitions.

That's the basic gist of it. I could go into more complicated detail.

[Mirage]

Please do.

Also, it's fun when I start to predict an AIs actions. Makes me feel intelligent.

[darkchrono]

Yeah I'd be interested in finding out more about it as well. The example I gave about how my Defensive Tackle adjusted to try to shut the play down (how would the programmers go about inputting something like that to where players can learn how to solve a problem and adjust (or at least make it look like that)).