rec.autos.simulators

I've searched this forum and others, I know it has to do with the
ai_dlongpad_scale and ai_dlat_pad values but some threads say you increase
it to stop getting rear ended and others say you decrease it.  What was the
end consensus to stop the AI from being so rammy?

Thanks
Rudeboy

I've tried upping the drafting distance a smidge (.3-.5) at several tracks,
and I like the difference so far.  For one, it seems easier to follow the AI
around the track without getting smacked from behind should I brake a little
early.  Plus, I don't have the AI shoving their noses down low every single
pseudo-chance they get.  They're still close enough to pounce when I make a
real mistake (iow not just sliding up a tiny bit) or am a bit slow coming
off a turn.  It also keeps them from tripping over each other and ending up
three wide at bad times or in bad places.  Just a cleaner race overall.

Alan

Alan, can you provide some more information on the changes you have
made to allow better racing.  Is it just the drafting distance that
needs chagning?  The changes your describe sound ideal.

On Tue, 02 Mar 2004 04:38:59 GMT, "Alan Le"

ai_drafting_distance = 1.6
ai_dlongpad_scale = 1000
ai_dlat_pad = 1.2

I'll try them out tonight and report back.

Also all the old posts say go to www.THEGREENFLAG.com but it's no longer
around.  Is there a new site or is he gone?

Indeed.  When running in a tight line, the AI would often try to go low if I
let up a little.  Problem was that the opening usually isn't even a car
wide!  Now, they're not close enough to get there right away, giving me time
to either close it off or drift up high while also allowing the AI to
realize it's not a good idea (if I continue just a bit off the line).

In the ini, I've only adjusted ai_drafting_distance, and so far it has
worked.  As I'm going through the championship season, I just increase the
value by .3 or .5 depending on how bad I think the AI is (.5 being for
really bad AI).  Also, I dropped all driver ratings by 10, low and high, for
every single car and upped the overall AI strength from 90% to 95%.  That
was done first with no success, so I don't know whether or not it's really
necessary.  Plus, I'm better now at following the AI around the track thanks
to setup and ability improvements.

If you find yourself getting nailed or dive bombed in braking zones due to
braking a lot earlier than the AI, dlongpad might help.  The description
sounds like lowering it would result in the AI spreading out in braking
zones.  I'm thinking dlat changes how close the AI gets when side by side.
To be honest though, I've never tried either.  I might try dlat at tracks
where the AI likes to squeeze the outside car against the wall.  It's a PITA
when you've got a limited FOV while on the outside down a straight with just
a few inches of room on either side.  Maybe I just suck. :)

Alan

   Hmmm that sounds like the AI Jan & Goy always came up with in the old
offline TPTCC leagues at the Pits...

dave henrie

Stephan Jay Gould running up to the Simpson's house yelling e***dly :  "I
have the results!  I have the results!"
Lisa : "What are they?"
SJG: "Inconclusive."
Lisa: "Why were you running?"
SJG : "I have to use your bathroom."

Anyway, inconclusive.  Seemed better but hard to tell.  I was running at
LasVegas last night. Appeared better than at the Rock but might be due to
the nature of the track and not the track.ini changes.  Or maybe the Placebo
effect.  First corner still a fifty-fifty thing getting out alive but strung
out fairly quickly and not really any incidents after the first couple of
laps.

Rudeboy

Ya I know. I was just so traumatized by being run over at the Rock I didn't
want to go back ;-)

But I  ran at I-70 last night and it seemed to be a bit better.  Again might
just be placebo.  Will continue testing.

Has anyone checked this to confirm that it doesn't happen.  I just
adjusted the drafting distance, as a test, at Bristol to 1.6.  A bit
extreme I know.  But it meant that as soon as the leader caught up with
the first backmarker he just got stuck behind him for the next 30+ laps,
after which time I gave up watching.
--
Peter Ives (AKA Pete Ivington)
Remove ALL_STRESS before replying via email
If you know what's good for you, don't listen to me :)
GPLRank Joystick -50.63 Wheel -25.01

I've never set it really high, but AI passing does seem to diminish as the
drafting distance goes up.  It also seems very track dependent.  I've been
finding an increase of .3 works at most tracks for me.  It's a good starting
point at least.

I ran Bristol a few days ago, and I set it to 1.45.  That was probably on
the high side, but the AI was still right there with a lot of good, clean
racing.  There wasn't constant passing, but it was definitely happening.
Plus, the passing attempts I did see looked more "normal."  IOW, the AI
wasn't constantly jumping out to try passes that obviously weren't going to
work.  There were still plenty of times where I just sat behind slower cars
simply because they were duking it out and using up the track.

Alan

What I was hoping, obviously, was that they affect how more likely
drivers are to make a mistake.   I tried setting the 'driver
consistency' rating to 0.25 to see if the drivers were going to be all
over the place but, although there appeared to initially be a small
flurry of early slip ups at Bristol, it wasn't long before all cars were
sticking to the racing line without error.

PS:

Any chance someone could send me the original Mark Martin .car file.  I
foolishly made a small change whilst looking at it in notepad and saved
the damned thing.  Only realised what a big mistake that was when I
scrolled further down the file to discover that the file wasn't just
plain text.  :(  Now MM don't want to race anymore.  See my sig for how
to email me.
--
Peter Ives (AKA Pete Ivington)
Remove ALL_STRESS before replying via email
If you know what's good for you, don't listen to me :)
GPLRank Joystick -50.63 Wheel -25.01

          John DiFool