OK.OK.OK. You got me. I'm the fool. Your right! I tried a few
dozen laps with my driver setting at 199-200 and got the same results
at 899-900 !!!! So, the hell with the driver settings!!! These have
absolutely no effect on calibrating N3. (oops)
However, to calibrate N3 you can still set up a test entry list of
three cars and get your lap times to match theirs by fooling with the
RELS line in the track's text file. I use the driver settings in the
AI cars to get an idea of slow or fast each setting effect laps times.
I noticed that the short, superspeedway, and road course setting have
no effect on Qualifying times....
Also, at the Daytona track I noticed that the AI was blowing out of
the pits at like 160mph!!!!! So I went into the pit.lp files and
adjusted the AI speeds there to a more reasonable speed (something
that the human controlled car could possibly keep up with. I also
noticed that at Tally I would lose AI cars in the back parts of the
track (especially turn 3) and they would get real speed brewing around
the front and be on my bumper by turn one. This kind of inconsistency
could have been fixed by tweaking the race.lp file. This may a
direction to head into if you don't want to mess with the RELS line.
Hey, ya learn something new everyday. Thanks for the info on the
driver settings....boy do I feel silly.
Hope whatever I said helped,
Gene
>Unfortunately, the power, aggression, drag, etc. values have no effect on a
>human controlled car--they only affect the AI. How does that influence your
>calibration process? I am not quite sure what it is achieving?
>Marc.
>> Hey Marc,
>> Here's how I've "calabrated" N3 (or any of the other Papy sims) for
>> REAL competetive results. What I do is create a driver entry list
>> called TEST with only three cars in it. One car is the car I'm
>> driving. I go in and set the aggression, power, drag, traction,
>> qualify, and track type settings so that there is only a one number
>> difference between the high and low values so there's no variables
>> between values. (Ex. Power: 500-501). Once I set my car, I set up
>> the second car, the AI car, EXACTLY the same as my car. Power, drag,
>> traction, etc. are all the same. I use the third car as a variable
>> car to get ideas of other settings and how they relate to my laps
>> times etc.
>> Once the settings are set, I go into whatever track I'm going to be
>> race on and start doing test laps in race mode. Make sure you set the
>> pace lap and yellow flags to off unless you like starting ALOT of
>> races in slow-motion and set the Opponent Strength to 100%. Also, set
>> the weather to constant (I've heard that variable weather conditions
>> only affect human controlled car and not the AI cars) and the mode to
>> realistic.
>> After about ten laps the AI cars and my car are pretty much up to
>> speed. I do about a total of 20-25 laps and then exit and look at the
>> last lap times and see how my car compares to the test car that is set
>> exactly like mine. If I'm beating the test car, I go into the track's
>> text file and bump up the RELS value until my last lap times are
>> similar to the computer test car's times. Once I've got this set,
>> then I go back to my real driver entry list (which I've tweaked to
>> match the real drivers as close as possible) and race in my
>> Championship Season!!!!
>> Hope this helps,
>> Gene
>> On Tue, 25 Apr 2000 02:20:15 GMT, "Marc Collins"
>> >I know this has been discussed before....
>> >For a variety of reasons, the AI in N3 perform relatively "better" and
>> >"worse" at various tracks. Even if we could magically get them to mimic
>the
>> >real life qualifying and racing speeds, the problem would still not be
>> >solved as your can cannot necessarily drive at the most realistic speed.
>> >Many have asked "why can't I keep up at Texas, but I can at Atlanta,"
>etc.
>> >What I am wondering is whether anyone has analysed the AI speeds relative
>to
>> >the human car and determined where (which tracks) the AI should be slowed
>> >down or sped up to make for more exciting racing. Can we as a group come
>to
>> >some agreement about these changes?
>> >I am happy to adjust the RELS value, but I would prefer to know what
>others
>> >have experienced, so I don't slow down the AI just because I haven't
>> >practised enough! Or, if I really can blow the doors off the AI, then
>maybe
>> >I would rather move on to another track rather than speed them up....but
>if
>> >most can beat them too easily, then I would rather speed them up for a
>more
>> >realistic challenge.
>> >So, is there any intelligence out there about this? Do people know where
>> >the AI are to fast and too slow and by how much? Please share if you do!
>> >--
>>***************************************************************************
>*
>> >Marc Collins
>> >Computers will never equal humans until they make mistakes and
>> >blame them on other computers.
>>***************************************************************************
>*
