rec.autos.simulators

by **Marc Collin** » Mon, 10 Feb 2003 15:07:33

No one has replied to my earlier message about the AI ignoring incidents in
practice mode...it's like they only slow down based on a yellow flag, not on
what they are "seeing." Sorry, but after having this same issue since N3,
it sucks.

The next issue has also been around since N3...that is, the AI aren't based
on some realistically driveable model for each track. If you set them to
100%, at some tracks they are unbeatable (uncatchable) and at others
relatively easy to beat.

I downloaded Bob Stanley's set-ups. I have no end of respect for Bob and
his generous contributions to our community. I have used his set-ups for N4
and NR2002 extensively. Of course, he was also involved in testing and
developing NR2003. Fire up Michigan and after quite a while manage a
37.661, which puts me 37th in the field. I am not an alien, but I have been
around the block a few times and with a great set-up I should be able to at
least make the bottom of the top third of the field at an easy track like
Michigan. So I decide to go check out Bob's set-up notes, which always list
his best time: 37.660.

Within 10 minutes I was within 1/1000th of a second of the set-up guru's
best time, yet it is only good for 37th place???

Without labouring over this with details from many tracks, suffice it to say
that, once again, you cannot set all the tracks to 100% AI and enjoy a
realistic season. If Bob Stanley's best set-up for track A yields a 43rd
place and for track B a 15th and for track C a 1st, there is something
wrong. They are all ovals and Bob knows more than 99.9% of NR2003
purchasers out there about set-ups and can drive better than at least 98% of
them. It's the same old problem--the AI at 100% are too fast at most
tracks, but it is the inconsistency that is the major problem. If you could
set them to some value and then know that they would be consistently
competent (as a group), who would care what that value was? Instead of
fixing the problem, we get adaptive AI and auto-adjusting AI...neither of
which supplies a "realistic" simulation. Either the player car has a major
problem or the AI speed and intelligence is a major problem.

Oh yeah, did I also mention that I am annoyed that another N4 bug/feature is
still present: the non-existent full field option. Instead, if you go to a
track that inexplicably can't handle the full field, it sets the number of
cars to that track's upper limit (39, for example). Then, unless you
manually reset the field, you will get only 39 cars at every track from then
on. Very annoying.

The new physics are great, but after paying full price for another update,
even if it is the sentimental last one, it is aggravating to be reminded of
oversights from the days of N3 that are still not addressed.

Marc

--
****************************************************************************
Marc Collins

Your mouse has moved. Windows must be restarted for the change
to take effect. Reboot now?
****************************************************************************

by **Tom Pabs** » Mon, 10 Feb 2003 16:52:08

Marc....

I don't know if the AI slider was ever supposed to be set a one
setting....and have that be universally correct (compared to something, I'm
not sure what) for all tracks. To me, if its a "slider"....its meant to be
"moved." Seems like sort of a common sense thingy.....LOL. Since I don't
run in "season mode" I guess it would be frustrating if you are not allowed
to alter the AI settings for each track/race. Is that true?

Alternately, when I want to race against the AI at a track that I am fairly
competent at.....I try to find a "slider" setting that has the top of the
field running at or near what the real WC cars run (top of their field) at
the same track (taking into consideration any variables in
weather....depending on what data you are using for the real WC speeds/lap
times). Of course, this requires keeping a log of these settings for each
track....if you want to repeat a race at any time in the future. Not too
difficult.

I don't think the adaptive AI was meant for us *** simmers. Do you?

I can't comment on the AI "ignoring incidents on track" in practice
sessions....haven't run any yet. However, if that is universal to all
tracks, my guess is there's a setting out of whack in the "papy_ai.ini"
file....someone is sure to find it and post a correction (like JJ
Johnston....or someone with similar skills in hacking/improving Papy's AI
performance).

I don't think using Bob Stanley's setups as a "benchmark" of what the AI
"should" be running....is a reasonable proposition. While Bob's setups are
excellent, indeed.....they are not meant to be AI benchmarking
devices....are they? If you use one of Bob's setups and you are running at
lap speeds equivalent to the real WC drivers....then find a suitable AI
slider setting that matches up....and bingo....problem over! And by the
way, if you only took 10 minutes to be as fast as Bob's notes say he
is....then I would guess there's a lot more speed left in that
setup.....with further adjustments. You shouldn't have even warmed your
tires practically....in 10 minutes. Perhaps you were "fudging" a little bit
in your time statement? Or, maybe trying to impress us a little bit?

To explain the "inexplicably"....some WC tracks don't have enough pit stalls
for 42 cars.....so the field gets reduced to the number of pit stalls it
does have. That used to be a lot of WC tracks....now its just a few.
Having to manually reset to "full field, 42 cars" doesn't seem like such a
bummer.......to me at least.

Calling NR2003 "another update" is being a little unrealistic....IMHO. And,
I think its unfair to those people at Papyrus (who were not around when N3,
or N4 were produced) who worked very hard (again, IMHO) on getting as much
of the "goodies" packed into NR2003 as they did (and some cools stuff we'd
not even been asking for) because it was the "last" of the Papy NASCAR sims!
Fortunately, I think most of them will view your post, as I did.......as
being from a little too ***of a "POV"......I don't think they'll lose any
sleep over your being aggravated at their work.

I'm not trying to be a smart-ass here.....just trying to balance the scales
a little bit.....for the newbies to this sim racing community who might have
mistakenly read your post and thought it contained anything of merit. It
doesn't.

Regards,

Tom

****************************************************************************

> Marc Collins

> Your mouse has moved. Windows must be restarted for the change
> to take effect. Reboot now?

****************************************************************************

- Show quoted text -

by **jwilso** » Mon, 10 Feb 2003 23:22:18

not sure but wouldn't the auto setting on AI address this problem?

> Marc....

> I don't know if the AI slider was ever supposed to be set a one
> setting....and have that be universally correct (compared to something,
I'm
> not sure what) for all tracks. To me, if its a "slider"....its meant to
be
> "moved." Seems like sort of a common sense thingy.....LOL. Since I
don't
> run in "season mode" I guess it would be frustrating if you are not
allowed
> to alter the AI settings for each track/race. Is that true?

> Alternately, when I want to race against the AI at a track that I am
fairly
> competent at.....I try to find a "slider" setting that has the top of the
> field running at or near what the real WC cars run (top of their field) at
> the same track (taking into consideration any variables in
> weather....depending on what data you are using for the real WC speeds/lap
> times). Of course, this requires keeping a log of these settings for each
> track....if you want to repeat a race at any time in the future. Not too
> difficult.

> I don't think the adaptive AI was meant for us *** simmers. Do you?

> I can't comment on the AI "ignoring incidents on track" in practice
> sessions....haven't run any yet. However, if that is universal to all
> tracks, my guess is there's a setting out of whack in the "papy_ai.ini"
> file....someone is sure to find it and post a correction (like JJ
> Johnston....or someone with similar skills in hacking/improving Papy's AI
> performance).

> I don't think using Bob Stanley's setups as a "benchmark" of what the AI
> "should" be running....is a reasonable proposition. While Bob's setups
are
> excellent, indeed.....they are not meant to be AI benchmarking
> devices....are they? If you use one of Bob's setups and you are running
at
> lap speeds equivalent to the real WC drivers....then find a suitable AI
> slider setting that matches up....and bingo....problem over! And by the
> way, if you only took 10 minutes to be as fast as Bob's notes say he
> is....then I would guess there's a lot more speed left in that
> setup.....with further adjustments. You shouldn't have even warmed your
> tires practically....in 10 minutes. Perhaps you were "fudging" a little
bit
> in your time statement? Or, maybe trying to impress us a little bit?

> To explain the "inexplicably"....some WC tracks don't have enough pit
stalls
> for 42 cars.....so the field gets reduced to the number of pit stalls it
> does have. That used to be a lot of WC tracks....now its just a few.
> Having to manually reset to "full field, 42 cars" doesn't seem like such a
> bummer.......to me at least.

> Calling NR2003 "another update" is being a little unrealistic....IMHO.
And,
> I think its unfair to those people at Papyrus (who were not around when
N3,
> or N4 were produced) who worked very hard (again, IMHO) on getting as much
> of the "goodies" packed into NR2003 as they did (and some cools stuff we'd
> not even been asking for) because it was the "last" of the Papy NASCAR
sims!
> Fortunately, I think most of them will view your post, as I did.......as
> being from a little too ***of a "POV"......I don't think they'll lose
any
> sleep over your being aggravated at their work.

> I'm not trying to be a smart-ass here.....just trying to balance the
scales
> a little bit.....for the newbies to this sim racing community who might
have
> mistakenly read your post and thought it contained anything of merit. It
> doesn't.

> Regards,

> Tom

> > No one has replied to my earlier message about the AI ignoring incidents
> in
> > practice mode...it's like they only slow down based on a yellow flag,
not
> on
> > what they are "seeing." Sorry, but after having this same issue since
N3,
> > it sucks.

> > The next issue has also been around since N3...that is, the AI aren't
> based
> > on some realistically driveable model for each track. If you set them
to
> > 100%, at some tracks they are unbeatable (uncatchable) and at others
> > relatively easy to beat.

> > I downloaded Bob Stanley's set-ups. I have no end of respect for Bob
and
> > his generous contributions to our community. I have used his set-ups
for
> N4
> > and NR2002 extensively. Of course, he was also involved in testing and
> > developing NR2003. Fire up Michigan and after quite a while manage a
> > 37.661, which puts me 37th in the field. I am not an alien, but I have
> been
> > around the block a few times and with a great set-up I should be able to
> at
> > least make the bottom of the top third of the field at an easy track
like
> > Michigan. So I decide to go check out Bob's set-up notes, which always
> list
> > his best time: 37.660.

> > Within 10 minutes I was within 1/1000th of a second of the set-up guru's
> > best time, yet it is only good for 37th place???

> > Without labouring over this with details from many tracks, suffice it to
> say
> > that, once again, you cannot set all the tracks to 100% AI and enjoy a
> > realistic season. If Bob Stanley's best set-up for track A yields a
43rd
> > place and for track B a 15th and for track C a 1st, there is something
> > wrong. They are all ovals and Bob knows more than 99.9% of NR2003
> > purchasers out there about set-ups and can drive better than at least
98%
> of
> > them. It's the same old problem--the AI at 100% are too fast at most
> > tracks, but it is the inconsistency that is the major problem. If you
> could
> > set them to some value and then know that they would be consistently
> > competent (as a group), who would care what that value was? Instead of
> > fixing the problem, we get adaptive AI and auto-adjusting AI...neither
of
> > which supplies a "realistic" simulation. Either the player car has a
> major
> > problem or the AI speed and intelligence is a major problem.

> > Oh yeah, did I also mention that I am annoyed that another N4
bug/feature
> is
> > still present: the non-existent full field option. Instead, if you go
to
> a
> > track that inexplicably can't handle the full field, it sets the number
of
> > cars to that track's upper limit (39, for example). Then, unless you
> > manually reset the field, you will get only 39 cars at every track from
> then
> > on. Very annoying.

> > The new physics are great, but after paying full price for another
update,
> > even if it is the sentimental last one, it is aggravating to be reminded
> of
> > oversights from the days of N3 that are still not addressed.

> > Marc

> > --

****************************************************************************

> > Marc Collins

> > Your mouse has moved. Windows must be restarted for the change
> > to take effect. Reboot now?

****************************************************************************

- Show quoted text -

by **Jay Taylo** » Tue, 11 Feb 2003 02:13:03

Actually manualy setting the AI to 100% gives you very consistant results.
While tuning the settings in NR2003 myself and most of the beta team did
numerous runs at all the tracks and generated tons on data on fast and
average lap times. This was what was used to set the AI's speeds. I know
myself personally I've been able to get top 10 qual positions at almost
every track, and can definately qual in the top 20 everywhere. Poles are
few and far between, as they should be. Also notice the AI's qual speeds
can vary some from race to race, this is because of the drivers stats being
"rerolled" everytime you enter the track. One of the biggest factors is
the car set you use. The default one that comes with NR2003 is decent and
should give you good results, however the full set at the pitwagon site will
probably give you slightly more even and realistic results.

The goal of setting the 100% difficulty was to make it a real challenge for
even the best sim racers to qualify and race upfront. While nothing is
perfect, I feel that from all the feedback we got thru testing we are pretty
damn close this time around. The only two tracks that might be a little to
fast are the roadcourses. Admittedly the AI at these two tracks might be a
touch faster on average then compared to the rest of the tracks. Keep in
mind that ever person that drives NR2003 also has good and bad tracks.
Personally I struggle to run in the top 20 at Homestead, but I know others
can run in the top 10 there consistantly, so in that case it's just me being
off there.

Jay Taylor

by **Marc Collin** » Tue, 11 Feb 2003 02:11:58

Once again Tom, you are an ass. Your message was quite reasonable and
respectful until you couldn't stop yourself from saying my message contained
nothing of merit.

If you are not annoyed by bugs and oversights that were pointed out 3
versions ago, so be it. Shut the *&%# up about Sears Point/Infineon
accuracy, then, too, to be consistent, since the vast majority wouldn't
notice and could care less.

The point is that 100% AI *DOES NOT* represent true WC speeds. And how
closely it represents them varies dramatically from one track to the next.
So great orb of wisdom, tell us what someone--a newbie or old timers like
us--should set their AI to if they want the "realistic" experience that this
title is striving to provide. Looking forward to your reply.