SDI Scritto:
-------------------------------------------------------
> @PaZa8523:
>
> 1) Values are between 0 and 16384 I'd say. You can
> use higher values, but I'm not sure if that will
> still increase the chance. With 100% chance it's
> still not the case that the driver will make
> errors all the time, just more often.
> 2) Not entirely, the variation is only used at the
> start to randomize the power. The error chance is
> about drivers making driving errors (or so it
> seems). Best to experiment with it and see for
> yourself.
>
> AI returning immediately to the track is already
> possible I think, using desc66 (spin recovery
> range). Or you want that to be driver specific
> too?
Point 2. I make you an example:
Australian GP, Clark curve. Many times AI make mistakes in this curve.
But unfortunately, AI stops at the side of the track waiting all cars to pass instead to move parallel to the track and return on it!
It's very annoying..