- 917
- ZeratulSG
Okay, to everyone who thinks the reason for the poor AI is due to it being a low priority for PD, tell me then why they made it one of their big marketing points during the dev of GT4? Quoted from the official GT4 site:
"Boasting a fantastically advanced and lifelike artificial intelligence (AI), your computer-controlled rivals in Gran Turismo 4 compete with a fervor unmatched even by some of your human opponents. With real-time reactions to your maneuvers and course behaviors modeled after actual professional racers, the challenge is more genuine than ever."
Okay, so we all know this is a steaming pile of cow guts, but the point is they punted "improved AI!" during pre-launch, and they're still doing it, so even though they dropped the ball badly (and IMO again by keeping this type of stuff on the site when it so obviously isn't true), it is/was obviously a priority to them.
Now, to bring this a little more OT, I was racing the American GT Cup (can't remember the name) over the weekend in the Chevy Camaro LM Race Car (which was overkill) and while leading the pack by several seconds after lap 1 at Seattle I decided to try and figure out why I was so much quicker.
In theory I should NOT have been able to walk all over the rest of the grid, which included an Esperante, a 'Vette C2 Race Car, a couple of Ford GT race cars, and the Team Oreca Viper. So I tried altering my driving style to be more conservative, and to replicate what I understand of RL GT car racing (which is to be a whole lot more cautious than GT4 forces you to be). So, braking slightly earlier, feathering the throttle more through turns, and basically reducing tire squeal to an absolute minimum. All of a sudden my split times stayed absolutely constant through the rest of the race (and my laptimes dropped by almost 6 seconds!).
So my theory is that, in this Cup with these cars and this particular AI setting (IMO the AI difficulty is adjusted for different cars in different races, Lupo Cup anyone?), the AI attempts to replicate RL GT car driving style. The reason I'm so much faster is because GT allows us to push the envelope a lot further than we could/would in RL.
"Boasting a fantastically advanced and lifelike artificial intelligence (AI), your computer-controlled rivals in Gran Turismo 4 compete with a fervor unmatched even by some of your human opponents. With real-time reactions to your maneuvers and course behaviors modeled after actual professional racers, the challenge is more genuine than ever."
Okay, so we all know this is a steaming pile of cow guts, but the point is they punted "improved AI!" during pre-launch, and they're still doing it, so even though they dropped the ball badly (and IMO again by keeping this type of stuff on the site when it so obviously isn't true), it is/was obviously a priority to them.
Now, to bring this a little more OT, I was racing the American GT Cup (can't remember the name) over the weekend in the Chevy Camaro LM Race Car (which was overkill) and while leading the pack by several seconds after lap 1 at Seattle I decided to try and figure out why I was so much quicker.
In theory I should NOT have been able to walk all over the rest of the grid, which included an Esperante, a 'Vette C2 Race Car, a couple of Ford GT race cars, and the Team Oreca Viper. So I tried altering my driving style to be more conservative, and to replicate what I understand of RL GT car racing (which is to be a whole lot more cautious than GT4 forces you to be). So, braking slightly earlier, feathering the throttle more through turns, and basically reducing tire squeal to an absolute minimum. All of a sudden my split times stayed absolutely constant through the rest of the race (and my laptimes dropped by almost 6 seconds!).
So my theory is that, in this Cup with these cars and this particular AI setting (IMO the AI difficulty is adjusted for different cars in different races, Lupo Cup anyone?), the AI attempts to replicate RL GT car driving style. The reason I'm so much faster is because GT allows us to push the envelope a lot further than we could/would in RL.