@monkeyking1969 said:
@alexglass said:
@monkeyking1969 said:
So how does 600% more AI look in a racing games? Umm, turn left if teh rule in NASCAR I hear...
No, really what does 600% more AI look like? Does the car accidentally turn 1 degree less in a turn slightly moving its line around the turn by a foot? Okay, well you don't need sophisticated AI for that a random number generator should have the driver apply less/more break or gas or turn of the wheel and get the same results. Car AI sucks, but ist not because there isn't enough processing power in a console.
Well in this case, the game's NPC A.I. Intelligently adapts to yours and your friends' driving style so that instead of static, permanent A.I. or rubber band A.I. so many racers suffer from, it will incorporate opponents that drive theirs cars more like you and your friends would.
You just answered your own question. Car A.I. sucks. So Forza tries to do something about it by giving it dynamic A.I. that doesn't suck and will change based on how you and other players improve over the lifetime of the game. So basically it adds a more realistic driving and learning curve, not just for you, but for the game's NPCs as well.
What's there to complain about?
I'm sorry this all seems like hookum. An AI that watches/studies other players who are expert at driving virtual cars or even expert at driving virtual cars are indistinguishable from a good AI that doesn't watch because a tracks are mostly fixed paths, and cars (even with variations) are a KNOWABLE set of variables. All this "600% better AI" is in reality is accumulated ghost data turned into a path...but that path is indistinguishable from teh path of an AI that isn't watching.
That's what there is to complain about - it is hookum. They are collecting data that when processed and spit back via the cloud doesn't change that AI at all.
You got the first part right, then somehow you ended up concluding a human gamer playing a game = a set of variables.
An A.I. that isn't watching is either designed by a programmer, a group of progammers, or pooled from data somewhere and it's static. It's a set of static rules that often leaves a lot to be desired in racing games.
In this case the A.I. design comes from actual gamers playing the game.
And yeah, that's actually what the cloud does in this case. The cloud processes the data, then your A.I. in your game changes from those static variables created by programmers to match the variables created by you or your friends. Which get updated over time and are now dynamic.
And to really put it in reality, what this does is actually introduce human error and habits into the A.I. The idea and hope behind it is that if for example you have a player that always tenses up during a particular turn and misses that turn consistently, you might have some of the A.I. drivers in your game mimmicking that. Rather than follow static, proper racing lines, they might similarly start running off the road. Or if it's a child playing, the A.I. may attempt to mimmick loss of concentration or going too fast through turns and spinning out. Things of that nature, etc.
And considering a ghost is just a straight re-run of a player's race, this is quite a bit different. Because it actually adapts their habits into those set of variables, into the game's A.I., rather than just replaying their ghost.
And one more thing. The path of a 5 year old's Forza A.I. will be very distinguishable for the standard path usually found in A.I. set initially by developers who are amazing at the game.
Log in to comment