rec.autos.simulators

NR2003: AI still inconsistent from one track to the next

Mar

NR2003: AI still inconsistent from one track to the next

by Mar » Wed, 19 Feb 2003 20:46:35


> Why should it be so difficult for someone to match Bob's best lap times?  He
> will admit he's not an alien racer.  I am not a newbie by a long shot,
> either.  I would be embarrassed if I couldn't get a decent time fairly
> quickly at the easiest track in the sim after YEARS of sim racing.  As I
> said somewhere else, I can beat the AI at 100% at every original track in
> GPL...to me that should be a hell of a lot harder than keeping up with the
> AI at a few ovals.

I may not have your level of experience, but I've been simming years,
too.  Got my first T2 while I was playing the original NASCAR Racing.
I know when I first downloaded the ASCAR Rockingham UK track I was
struggling to get 37s laps, now I do consistent mid-high 35s in race
conditions having worn my own groove into the tarmac and played with
the setup a bit.  Can't stop it getting a *severe* push about 18 laps
into a run, though...  I digress.  There's no reason to expect that
you won't run competitively given a bit of time, but to essentially
match Bob's best time within 10 minutes of driving the sim, with a
setup that will be geared towards his driving style?  I am *not*
saying that you didn't do it, I am saying that the game may have
fundamentally changed since that setup was completed.  Have you tried
a bit of tinkering yourself?  There must be a load more speed in the
car - I'd be stunned if you've reached your ultimate speed in 10
minutes.

Even with years of experience, each iteration of these sims gets a bit
closer to reality and we have to make adjustments ourselves.

I've probably put in about an hour with the michigan "official" demo
and find I'm just about getting the hang of the "feel" of this new
sim.

Regards,

Mark

Peter Ive

NR2003: AI still inconsistent from one track to the next

by Peter Ive » Thu, 20 Feb 2003 03:57:37





>> Why should it be so difficult for someone to match Bob's best lap times?  
>He
>> will admit he's not an alien racer.  I am not a newbie by a long shot,
>> either.  I would be embarrassed if I couldn't get a decent time fairly
>> quickly at the easiest track in the sim after YEARS of sim racing.  As I
>> said somewhere else, I can beat the AI at 100% at every original track in
>> GPL...to me that should be a hell of a lot harder than keeping up with the
>> AI at a few ovals.

<snip>

I have always found that decent setups, no matter who's driving style
they are geared towards, have allowed me to drive faster laps straight
away.  As yet I haven't got hold of NR2003, being in the UK, so I cannot
give a definitive response though I doubt very much whether this has
changed much.  Therefore, why does it seem implausible that Marc is
capable of matching Bob's time after just 10 minutes?

In my view, a good setup will help a good driver go quicker.

And to further test my theory I am going to d/l some fast setups for
NR2002 which, as yet, I'm still using the default fast setups.  :)
--
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 -21.77

Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Thu, 20 Feb 2003 08:07:05

You are correct Peter, but once again I will reiterate that this isn't about
Bob's set-ups!!!!!!!!!!!!!!!!

Take the fast set-ups in the game if you want.  Can you do equally well at
California and Michigan?  At Texas and Atlanta?  You should be able to.

The fact that Bob's Rockingham set-up gives a top 2 and his Michigan a
bottom 2 is just an example of the problems with AI calibration in my
opinion.  This was a known and widespread problem in past versions.  Jay
says it is fixed in 2003.  I just haven't seen any evidence of it myself.
And rather than blather on about my own skills or set-ups I am using Bon
Stanley's as an illustration because he was a beta tester, he knows more
about setting up a NR2003 car than 99.5% of us and his set-ups are readily
available for someone to take them and prove me wrong.

So for no one has, but I have a lot of unrelated suggestions that aren't
helping to prove wrong my original point (which I honestly hope is proven
completely wrong!!!!!!!!!!)

Thanks,

Marc






> >> Why should it be so difficult for someone to match Bob's best lap
times?
> >He
> >> will admit he's not an alien racer.  I am not a newbie by a long shot,
> >> either.  I would be embarrassed if I couldn't get a decent time fairly
> >> quickly at the easiest track in the sim after YEARS of sim racing.  As
I
> >> said somewhere else, I can beat the AI at 100% at every original track
in
> >> GPL...to me that should be a hell of a lot harder than keeping up with
the
> >> AI at a few ovals.

> <snip>

> > There's no reason to expect that
> >you won't run competitively given a bit of time, but to essentially
> >match Bob's best time within 10 minutes of driving the sim, with a
> >setup that will be geared towards his driving style?  I am *not*
> >saying that you didn't do it, I am saying that the game may have
> >fundamentally changed since that setup was completed.  Have you tried
> >a bit of tinkering yourself?  There must be a load more speed in the
> >car - I'd be stunned if you've reached your ultimate speed in 10
> >minutes.

> I have always found that decent setups, no matter who's driving style
> they are geared towards, have allowed me to drive faster laps straight
> away.  As yet I haven't got hold of NR2003, being in the UK, so I cannot
> give a definitive response though I doubt very much whether this has
> changed much.  Therefore, why does it seem implausible that Marc is
> capable of matching Bob's time after just 10 minutes?

> In my view, a good setup will help a good driver go quicker.

> And to further test my theory I am going to d/l some fast setups for
> NR2002 which, as yet, I'm still using the default fast setups.  :)
> --
> 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 -21.77

Peter Ive

NR2003: AI still inconsistent from one track to the next

by Peter Ive » Thu, 20 Feb 2003 11:19:12



Funnily enough, after downloading some random race setups for n2k2,
created by different people,  I was actually running at about the same
laptimes as with the default fast setups after 5 minutes or so (My
fastest laps being +/- 1/10th of a second compared to my previous
fastest).  Not sure what that says about my previous statement about a
good driver going faster with any good setup.  I was expecting to have
been quicker than before at all tracks I got them for.  Maybe I'm just
becoming a bit ***these days.  I bet if I went back to joys...  Nah!!

Though I've yet to get hold of NR2003 I too would expect to be
competitive at all tracks, let alone those that are similar.  I think
your best bet for a definitive answer is going to be to get hold of some
more setups, do a bit of tweaking yourself, or wait until some fastlaps
start to appear on the net, like has been said previously.  It would
seem that you're unlikely to get any corroboration of your findings at
this stage.  Who knows, maybe others have also noticed this, but are not
quite as sure of their own abilities as you are and so are not prepared
to stick their necks out.  When's it out over here in the UK again?  :)

<snip>

--
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 -21.77

Dave Boyl

NR2003: AI still inconsistent from one track to the next

by Dave Boyl » Thu, 20 Feb 2003 14:39:37

You keep saying that his Michigan setup gives bottom 2 time. Well, If he's
doing that time without a draft, I'm guessing he's capable of 37.0s (or
better) in a good draft. How does that time stack up against the AI?

You keep quoting Bob's times listed in his setups. But AFAIK, Bob isn't the
one with the problem, you are. So why don't you base your arguments off of
your own race results? I haven't seen anything that you've written that
supports your underlying argument. Unless your argument happens to be that
you can't Hot Lap with the AI equally at every track...

db


> You are correct Peter, but once again I will reiterate that this isn't
about
> Bob's set-ups!!!!!!!!!!!!!!!!

> Take the fast set-ups in the game if you want.  Can you do equally well at
> California and Michigan?  At Texas and Atlanta?  You should be able to.

> The fact that Bob's Rockingham set-up gives a top 2 and his Michigan a
> bottom 2 is just an example of the problems with AI calibration in my
> opinion.  This was a known and widespread problem in past versions.  Jay
> says it is fixed in 2003.  I just haven't seen any evidence of it myself.
> And rather than blather on about my own skills or set-ups I am using Bon
> Stanley's as an illustration because he was a beta tester, he knows more
> about setting up a NR2003 car than 99.5% of us and his set-ups are readily
> available for someone to take them and prove me wrong.

> So for no one has, but I have a lot of unrelated suggestions that aren't
> helping to prove wrong my original point (which I honestly hope is proven
> completely wrong!!!!!!!!!!)

> Thanks,

> Marc







> > >> Why should it be so difficult for someone to match Bob's best lap
> times?
> > >He
> > >> will admit he's not an alien racer.  I am not a newbie by a long
shot,
> > >> either.  I would be embarrassed if I couldn't get a decent time
fairly
> > >> quickly at the easiest track in the sim after YEARS of sim racing.
As
> I
> > >> said somewhere else, I can beat the AI at 100% at every original
track
> in
> > >> GPL...to me that should be a hell of a lot harder than keeping up
with
> the
> > >> AI at a few ovals.

> > <snip>

> > > There's no reason to expect that
> > >you won't run competitively given a bit of time, but to essentially
> > >match Bob's best time within 10 minutes of driving the sim, with a
> > >setup that will be geared towards his driving style?  I am *not*
> > >saying that you didn't do it, I am saying that the game may have
> > >fundamentally changed since that setup was completed.  Have you tried
> > >a bit of tinkering yourself?  There must be a load more speed in the
> > >car - I'd be stunned if you've reached your ultimate speed in 10
> > >minutes.

> > I have always found that decent setups, no matter who's driving style
> > they are geared towards, have allowed me to drive faster laps straight
> > away.  As yet I haven't got hold of NR2003, being in the UK, so I cannot
> > give a definitive response though I doubt very much whether this has
> > changed much.  Therefore, why does it seem implausible that Marc is
> > capable of matching Bob's time after just 10 minutes?

> > In my view, a good setup will help a good driver go quicker.

> > And to further test my theory I am going to d/l some fast setups for
> > NR2002 which, as yet, I'm still using the default fast setups.  :)
> > --
> > 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 -21.77

Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Fri, 21 Feb 2003 09:17:39

As I said, use the fast set-ups from the game, use your own set-ups, who
cares?  The results are the same.  That is, two ostensibly similar tracks
have AI that are quite different in speed.  Draft is irrelevant, the results
are the same in practise or race.  We are not talking testing alone on the
track--all normal modes have a full field of cars to potentially draft with.
And I always use a full field.

Marc


> You keep saying that his Michigan setup gives bottom 2 time. Well, If he's
> doing that time without a draft, I'm guessing he's capable of 37.0s (or
> better) in a good draft. How does that time stack up against the AI?

> You keep quoting Bob's times listed in his setups. But AFAIK, Bob isn't
the
> one with the problem, you are. So why don't you base your arguments off of
> your own race results? I haven't seen anything that you've written that
> supports your underlying argument. Unless your argument happens to be that
> you can't Hot Lap with the AI equally at every track...

> db



> > You are correct Peter, but once again I will reiterate that this isn't
> about
> > Bob's set-ups!!!!!!!!!!!!!!!!

> > Take the fast set-ups in the game if you want.  Can you do equally well
at
> > California and Michigan?  At Texas and Atlanta?  You should be able to.

> > The fact that Bob's Rockingham set-up gives a top 2 and his Michigan a
> > bottom 2 is just an example of the problems with AI calibration in my
> > opinion.  This was a known and widespread problem in past versions.  Jay
> > says it is fixed in 2003.  I just haven't seen any evidence of it
myself.
> > And rather than blather on about my own skills or set-ups I am using Bon
> > Stanley's as an illustration because he was a beta tester, he knows more
> > about setting up a NR2003 car than 99.5% of us and his set-ups are
readily
> > available for someone to take them and prove me wrong.

> > So for no one has, but I have a lot of unrelated suggestions that aren't
> > helping to prove wrong my original point (which I honestly hope is
proven
> > completely wrong!!!!!!!!!!)

> > Thanks,

> > Marc







> > > >> Why should it be so difficult for someone to match Bob's best lap
> > times?
> > > >He
> > > >> will admit he's not an alien racer.  I am not a newbie by a long
> shot,
> > > >> either.  I would be embarrassed if I couldn't get a decent time
> fairly
> > > >> quickly at the easiest track in the sim after YEARS of sim racing.
> As
> > I
> > > >> said somewhere else, I can beat the AI at 100% at every original
> track
> > in
> > > >> GPL...to me that should be a hell of a lot harder than keeping up
> with
> > the
> > > >> AI at a few ovals.

> > > <snip>

> > > > There's no reason to expect that
> > > >you won't run competitively given a bit of time, but to essentially
> > > >match Bob's best time within 10 minutes of driving the sim, with a
> > > >setup that will be geared towards his driving style?  I am *not*
> > > >saying that you didn't do it, I am saying that the game may have
> > > >fundamentally changed since that setup was completed.  Have you tried
> > > >a bit of tinkering yourself?  There must be a load more speed in the
> > > >car - I'd be stunned if you've reached your ultimate speed in 10
> > > >minutes.

> > > I have always found that decent setups, no matter who's driving style
> > > they are geared towards, have allowed me to drive faster laps straight
> > > away.  As yet I haven't got hold of NR2003, being in the UK, so I
cannot
> > > give a definitive response though I doubt very much whether this has
> > > changed much.  Therefore, why does it seem implausible that Marc is
> > > capable of matching Bob's time after just 10 minutes?

> > > In my view, a good setup will help a good driver go quicker.

> > > And to further test my theory I am going to d/l some fast setups for
> > > NR2002 which, as yet, I'm still using the default fast setups.  :)
> > > --
> > > 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 -21.77

Dave Boyl

NR2003: AI still inconsistent from one track to the next

by Dave Boyl » Fri, 21 Feb 2003 09:56:44

Results? What results? I don't remember you giving any race results...

Wait, you were quoting Bob's listed times as a baseline and using that as
the foundation of your position. From what I understand, his listed times

under similar conditions to even make the argument that the AI Hot Lap
inconsistently from track to track (as compared to Bob's times). But that's
not what I would consider a complete analysis of the AI speeds.

Who knows, you may be 100% correct. But you haven't given any info that
would make me draw the same conclusion that you have. You seem to think
Michigan is overly tough. Go run a 50 lap race there and give a report. IMO,
to honestly judge the AI, you need to run a full fuel run race against them.
Hotlapping against them really doesn't give a full representation of the AI
throughout an entire run.

db


> As I said, use the fast set-ups from the game, use your own set-ups, who
> cares?  The results are the same.  That is, two ostensibly similar tracks
> have AI that are quite different in speed.  Draft is irrelevant, the
results
> are the same in practise or race.  We are not talking testing alone on the
> track--all normal modes have a full field of cars to potentially draft
with.
> And I always use a full field.

> Marc



> > You keep saying that his Michigan setup gives bottom 2 time. Well, If
he's
> > doing that time without a draft, I'm guessing he's capable of 37.0s (or
> > better) in a good draft. How does that time stack up against the AI?

> > You keep quoting Bob's times listed in his setups. But AFAIK, Bob isn't
> the
> > one with the problem, you are. So why don't you base your arguments off
of
> > your own race results? I haven't seen anything that you've written that
> > supports your underlying argument. Unless your argument happens to be
that
> > you can't Hot Lap with the AI equally at every track...

> > db



> > > You are correct Peter, but once again I will reiterate that this isn't
> > about
> > > Bob's set-ups!!!!!!!!!!!!!!!!

> > > Take the fast set-ups in the game if you want.  Can you do equally
well
> at
> > > California and Michigan?  At Texas and Atlanta?  You should be able
to.

> > > The fact that Bob's Rockingham set-up gives a top 2 and his Michigan a
> > > bottom 2 is just an example of the problems with AI calibration in my
> > > opinion.  This was a known and widespread problem in past versions.
Jay
> > > says it is fixed in 2003.  I just haven't seen any evidence of it
> myself.
> > > And rather than blather on about my own skills or set-ups I am using
Bon
> > > Stanley's as an illustration because he was a beta tester, he knows
more
> > > about setting up a NR2003 car than 99.5% of us and his set-ups are
> readily
> > > available for someone to take them and prove me wrong.

> > > So for no one has, but I have a lot of unrelated suggestions that
aren't
> > > helping to prove wrong my original point (which I honestly hope is
> proven
> > > completely wrong!!!!!!!!!!)

> > > Thanks,

> > > Marc








> > > > >> Why should it be so difficult for someone to match Bob's best lap
> > > times?
> > > > >He
> > > > >> will admit he's not an alien racer.  I am not a newbie by a long
> > shot,
> > > > >> either.  I would be embarrassed if I couldn't get a decent time
> > fairly
> > > > >> quickly at the easiest track in the sim after YEARS of sim
racing.
> > As
> > > I
> > > > >> said somewhere else, I can beat the AI at 100% at every original
> > track
> > > in
> > > > >> GPL...to me that should be a hell of a lot harder than keeping up
> > with
> > > the
> > > > >> AI at a few ovals.

> > > > <snip>

> > > > > There's no reason to expect that
> > > > >you won't run competitively given a bit of time, but to essentially
> > > > >match Bob's best time within 10 minutes of driving the sim, with a
> > > > >setup that will be geared towards his driving style?  I am *not*
> > > > >saying that you didn't do it, I am saying that the game may have
> > > > >fundamentally changed since that setup was completed.  Have you
tried
> > > > >a bit of tinkering yourself?  There must be a load more speed in
the
> > > > >car - I'd be stunned if you've reached your ultimate speed in 10
> > > > >minutes.

> > > > I have always found that decent setups, no matter who's driving
style
> > > > they are geared towards, have allowed me to drive faster laps
straight
> > > > away.  As yet I haven't got hold of NR2003, being in the UK, so I
> cannot
> > > > give a definitive response though I doubt very much whether this has
> > > > changed much.  Therefore, why does it seem implausible that Marc is
> > > > capable of matching Bob's time after just 10 minutes?

> > > > In my view, a good setup will help a good driver go quicker.

> > > > And to further test my theory I am going to d/l some fast setups for
> > > > NR2002 which, as yet, I'm still using the default fast setups.  :)
> > > > --
> > > > 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 -21.77

Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Fri, 21 Feb 2003 13:10:02

Why should the AI be able to hot lap any better than the human player?

You keep dwelling on particular points in my posts and ignoring the main
point (the subject line of the thread).  My experience so far is that with
Bob's set-ups, with the fast set-ups, in practice, in a race, regardless,
the same results occur: certain tracks AI at 100% are very, very tough while
at a sister track they are not.  This seems unrealistically so to me.  I am
waiting for someone to provide some set-ups or some replays or just some
hard factual feedback to suggest I am wonky.

Why don't you just tell us what your experience is in this regard?  With
details please...

Also, here is a quote from another old hat here in another post about wacky
AI speeds at the SS tracks posted a short while ago:

As for the AI, since making it is an almost completely new task for each new
track, I find it easy to imagine that the AI, as it has always been, is a
little more in line with real-life speeds on some tracks than on others. And
when the AI gives us a good draft, the player car will probably follow into
exaggerated speeds to an extent. I only need to look at the speeds my wife's
little grasshopper car reaches on the highway when she gets a good draft <g>

Achim

As I have been noting all along, this isn't a new issue.  The AI has always
been imbalanced in the past versions and it would be a surprise if it were
NOT imbalanced in this version.  Not the other way around.  Although I hope
it is the first time they have got it right.

Marc

"Dave Boyle" <Like.I.Need.E...@More.Spam> wrote in message

news:gJV4a.1289609$TJ.181541@post-02.news.easynews.com...
> Results? What results? I don't remember you giving any race results...

> Wait, you were quoting Bob's listed times as a baseline and using that as
> the foundation of your position. From what I understand, his listed times
> are all @ 70*, Clear, No Wind, No Draft. You have to at least run the AI
> under similar conditions to even make the argument that the AI Hot Lap
> inconsistently from track to track (as compared to Bob's times). But
that's
> not what I would consider a complete analysis of the AI speeds.

> Who knows, you may be 100% correct. But you haven't given any info that
> would make me draw the same conclusion that you have. You seem to think
> Michigan is overly tough. Go run a 50 lap race there and give a report.
IMO,
> to honestly judge the AI, you need to run a full fuel run race against
them.
> Hotlapping against them really doesn't give a full representation of the
AI
> throughout an entire run.

> db

> "Marc Collins" <marc_coll...@sympatico.ca> wrote in message
> news:E8V4a.6558$iu1.566335@news20.bellglobal.com...
> > As I said, use the fast set-ups from the game, use your own set-ups, who
> > cares?  The results are the same.  That is, two ostensibly similar
tracks
> > have AI that are quite different in speed.  Draft is irrelevant, the
> results
> > are the same in practise or race.  We are not talking testing alone on
the
> > track--all normal modes have a full field of cars to potentially draft
> with.
> > And I always use a full field.

> > Marc

> > "Dave Boyle" <Like.I.Need.E...@More.Spam> wrote in message
> > news:tME4a.4575728$6N5.610268@post-03.news.easynews.com...
> > > You keep saying that his Michigan setup gives bottom 2 time. Well, If
> he's
> > > doing that time without a draft, I'm guessing he's capable of 37.0s
(or
> > > better) in a good draft. How does that time stack up against the AI?

> > > You keep quoting Bob's times listed in his setups. But AFAIK, Bob
isn't
> > the
> > > one with the problem, you are. So why don't you base your arguments
off
> of
> > > your own race results? I haven't seen anything that you've written
that
> > > supports your underlying argument. Unless your argument happens to be
> that
> > > you can't Hot Lap with the AI equally at every track...

> > > db

> > > "Marc Collins" <marc_coll...@sympatico.ca> wrote in message
> > > news:r0z4a.3555$mZ2.539022@news20.bellglobal.com...
> > > > You are correct Peter, but once again I will reiterate that this
isn't
> > > about
> > > > Bob's set-ups!!!!!!!!!!!!!!!!

> > > > Take the fast set-ups in the game if you want.  Can you do equally
> well
> > at
> > > > California and Michigan?  At Texas and Atlanta?  You should be able
> to.

> > > > The fact that Bob's Rockingham set-up gives a top 2 and his Michigan
a
> > > > bottom 2 is just an example of the problems with AI calibration in
my
> > > > opinion.  This was a known and widespread problem in past versions.
> Jay
> > > > says it is fixed in 2003.  I just haven't seen any evidence of it
> > myself.
> > > > And rather than blather on about my own skills or set-ups I am using
> Bon
> > > > Stanley's as an illustration because he was a beta tester, he knows
> more
> > > > about setting up a NR2003 car than 99.5% of us and his set-ups are
> > readily
> > > > available for someone to take them and prove me wrong.

> > > > So for no one has, but I have a lot of unrelated suggestions that
> aren't
> > > > helping to prove wrong my original point (which I honestly hope is
> > proven
> > > > completely wrong!!!!!!!!!!)

> > > > Thanks,

> > > > Marc

> > > > "Peter Ives" <p...@pgives.ALL_STRESSfreeserve.co.uk> wrote in
message
> > > > news:NdJ4fCAhIoU+Ewg+@pgives.ALL_STRESSfreeserve.co.uk...
> > > > > In article <8c277338.0302180346.636d8...@posting.google.com>, Mark
> > > > > <the.davis...@btinternet.com> writes
> > > > > >"Marc Collins" <marc_coll...@sympatico.ca> wrote in message
> news:<3Zd
> > > > > >4a.1458$sv3.290...@news20.bellglobal.com>...

> > > > > >> Why should it be so difficult for someone to match Bob's best
lap
> > > > times?
> > > > > >He
> > > > > >> will admit he's not an alien racer.  I am not a newbie by a
long
> > > shot,
> > > > > >> either.  I would be embarrassed if I couldn't get a decent time
> > > fairly
> > > > > >> quickly at the easiest track in the sim after YEARS of sim
> racing.
> > > As
> > > > I
> > > > > >> said somewhere else, I can beat the AI at 100% at every
original
> > > track
> > > > in
> > > > > >> GPL...to me that should be a hell of a lot harder than keeping
up
> > > with
> > > > the
> > > > > >> AI at a few ovals.

> > > > > <snip>

> > > > > > There's no reason to expect that
> > > > > >you won't run competitively given a bit of time, but to
essentially
> > > > > >match Bob's best time within 10 minutes of driving the sim, with
a
> > > > > >setup that will be geared towards his driving style?  I am *not*
> > > > > >saying that you didn't do it, I am saying that the game may have
> > > > > >fundamentally changed since that setup was completed.  Have you
> tried
> > > > > >a bit of tinkering yourself?  There must be a load more speed in
> the
> > > > > >car - I'd be stunned if you've reached your ultimate speed in 10
> > > > > >minutes.

> > > > > I have always found that decent setups, no matter who's driving
> style
> > > > > they are geared towards, have allowed me to drive faster laps
> straight
> > > > > away.  As yet I haven't got hold of NR2003, being in the UK, so I
> > cannot
> > > > > give a definitive response though I doubt very much whether this
has
> > > > > changed much.  Therefore, why does it seem implausible that Marc
is
> > > > > capable of matching Bob's time after just 10 minutes?

> > > > > In my view, a good setup will help a good driver go quicker.

> > > > > And to further test my theory I am going to d/l some fast setups
for
> > > > > NR2002 which, as yet, I'm still using the default fast setups.  :)
> > > > > --
> > > > > 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 -21.77

Dave Boyl

NR2003: AI still inconsistent from one track to the next

by Dave Boyl » Fri, 21 Feb 2003 16:08:45


Well, they shouldn't. But I haven't had any experience with them having
significantly faster blaps than me yet. Using Michigan (a track you seem to
have a problem with) My fastest race lap is now 37.132 from my second race
there (still using Bob's setup).  I think the fastest AI lap in that race
was 37.0. But even so, do you consider a blap as the most important factor
to judge the AI by? I certainly don't...

Wait just a minute there...
If I recall correctly, I'm the *ONLY* person in this thread that has given a
race analysis. *Including* you. Since you are the one that claims there is a
dramatic problem, why don't you pick a track that you think is way too fast,
run a race and then report the results?

Like I said earlier, I agree that the AI has had *many* issues in past
releases. But I have seen *nothing* in NR03 that approaches the problems of
the past. I think you have simply made up your mind to find a fault and are
doing everything possible not to contradict your predetermined conclusion.

Jay Taylo

NR2003: AI still inconsistent from one track to the next

by Jay Taylo » Fri, 21 Feb 2003 16:52:34

Marc,
Its really pointless for me to add more, you have an firm opinion on how to
judge the AI's balance in the game, and their consistancy from track to
track. Unfortunately  the game was balanced completly using race or qual
session data, which is different then the methode you insist on.  We can
continue to disagree on what difference this makes, but I've stated over and
over, that you really need to generate data from races with the AI.  Compare
their speeds in the draft to yours, compare their speed leading the draft to
yours, and compare their corner and straight speeds for both conditions.
You also need to look at their speeds over a full tire/fuel run, and how
they match up with the players on average.  This was how the track by track
balancing was done.  Qual speeds are different because you just get data
from the beta team on what kind of qual speeds they are running for each
track, and then adjust the AI's qual modifier until you get them to average
around the point your looking for.

I havn't seen you give any feedback of you verse the AI in any races, just
your hot laps compared to their practice sessions.  If you've given some,
and I missed it I apologize.   I'm just figuring your interaction with the
AI is mainly to compare fast laps, which really isn't the focus of the AI in
NR2003.  IN the end your fast laps( or maybe just Bobs posted laps) are
inconsistant verse the AI ( in practice) at Cali and Mich.  Its very
possible there is a slight difference in difficulty which gets skewed even
further when you just look at their practice data, and you dont take into
account how many different driving styles people have.
I do not claim by any means that the AI are perfect, but I do feel they are
much more evenly balanced then they have been in previous efforts.
At this point all I can do is agree to disagree with you.

Jay Taylor

Uncle Feste

NR2003: AI still inconsistent from one track to the next

by Uncle Feste » Fri, 21 Feb 2003 20:52:01


> Why should the AI be able to hot lap any better than the human player?

> You keep dwelling on particular points in my posts and ignoring the main
> point (the subject line of the thread).  My experience so far is that with
> Bob's set-ups, with the fast set-ups, in practice, in a race, regardless,
> the same results occur: certain tracks AI at 100% are very, very tough while
> at a sister track they are not.  This seems unrealistically so to me.  I am
> waiting for someone to provide some set-ups or some replays or just some
> hard factual feedback to suggest I am wonky.

Seeing how 90+ percent of us run it only online, it's unlikely we have,
or ever will, notice any such discrepencies in the AI.  Just lighten up
& enjoy.  See you online.

--

Fester

Abandon the search for Truth; settle for a good fantasy.

Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Fri, 21 Feb 2003 23:09:48

I already made the point that thanks to wonky AI in the past most people
have migrated to online only.  It doesn't help answer the question or solve
any problems.

Marc



> > Why should the AI be able to hot lap any better than the human player?

> > You keep dwelling on particular points in my posts and ignoring the main
> > point (the subject line of the thread).  My experience so far is that
with
> > Bob's set-ups, with the fast set-ups, in practice, in a race,
regardless,
> > the same results occur: certain tracks AI at 100% are very, very tough
while
> > at a sister track they are not.  This seems unrealistically so to me.  I
am
> > waiting for someone to provide some set-ups or some replays or just some
> > hard factual feedback to suggest I am wonky.

> Seeing how 90+ percent of us run it only online, it's unlikely we have,
> or ever will, notice any such discrepencies in the AI.  Just lighten up
> & enjoy.  See you online.

> --

> Fester

> Abandon the search for Truth; settle for a good fantasy.

Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Fri, 21 Feb 2003 23:23:20

Jay:

For argument's sake let's say I agree with you 100%.  Can you then tell me
why Practise Mode is so different?  Why wouldn't Papy program practise mode
so the cars are doing the same things they do in a race, since if you are
playing the game properly, you try to practise first to ensure you are good
enough to not cause mayhem in a "real" race?  Or, practise until you are
fast enough it is worth racing (assuming driving around 400 laps in 43rd
isn't anyone's idea of fun).

Instead, they ignore incidents on the track and pile drive through other
cars or the human player (as though they were waiting for a yellow flag so
they could see what was in front of them) and according to what you are
saying, have access to some draft effect or grip or speed that they don't
during the race or that the human player doesn't have...??  What would be
the point of this?

It is my belief that the problem in the first paragraph was "solved" with
the adaptive AI and Auto AI.  I have no idea what the design issue is in the
second paragraph, but it was there in N3, N4 and NR2002.  Needless to say, I
wasn't impressed when I saw it again in NR2003.

Marc

PS, you say "I haven't seen you give any feedback of you verse the AI in any
races, just your hot laps compared to their practice sessions."  If you read
even just the message you were replying to you will see that I am not
talking about just practise mode any longer.  I noticed the problem there
first when I originated the thread, but subsequently confirmed that in my
experience the trend is the same in races.  Of course you are correct that
lots and lots of testing would need to be done to corroborate this, which is
impossible for any one person with an outside life, hence my talking about
it here and attempting to get more opinions and experience.  Instead, Uncle
Fester seems to have said what I already suspected...


Dave Boyl

NR2003: AI still inconsistent from one track to the next

by Dave Boyl » Sun, 23 Feb 2003 04:24:25

There is much more to racing than simply getting one clean lap in a good
draft. Negotiating traffic, managing tire wear and (if you use cautions or
run long enough races for pit stops) strategy/tactics all play a major role
in a race. So to make a broad statement about the AI based on best practice
laps isn't necessarily proper. It's the same thing in real life, the guys
who post the best practice times don't necessarily do the best in the race.

You mentioned race mode in your last message, but I still don't recall any
race info. At what tracks do you find the AI too difficult to race and which
to easy? What are your race results at these tracks?

db


Marc Collin

NR2003: AI still inconsistent from one track to the next

by Marc Collin » Mon, 24 Feb 2003 02:20:17

All those things are there in Practise mode, too, assuming you practise to
be better in a race (which I do--I never hot lap), so can you answer my
question?

Also, you should look at the later thread "N2003 Tona/Dega" by Jason Moyer
for some interesting follow-up to the issues here.

Marc


> There is much more to racing than simply getting one clean lap in a good
> draft. Negotiating traffic, managing tire wear and (if you use cautions or
> run long enough races for pit stops) strategy/tactics all play a major
role
> in a race. So to make a broad statement about the AI based on best
practice
> laps isn't necessarily proper. It's the same thing in real life, the guys
> who post the best practice times don't necessarily do the best in the
race.

> You mentioned race mode in your last message, but I still don't recall any
> race info. At what tracks do you find the AI too difficult to race and
which
> to easy? What are your race results at these tracks?

> db



> > Jay:

> > For argument's sake let's say I agree with you 100%.  Can you then tell
me
> > why Practise Mode is so different?  Why wouldn't Papy program practise
> mode
> > so the cars are doing the same things they do in a race, since if you
are
> > playing the game properly, you try to practise first to ensure you are
> good
> > enough to not cause mayhem in a "real" race?  Or, practise until you are
> > fast enough it is worth racing (assuming driving around 400 laps in 43rd
> > isn't anyone's idea of fun).

> > Instead, they ignore incidents on the track and pile drive through other
> > cars or the human player (as though they were waiting for a yellow flag
so
> > they could see what was in front of them) and according to what you are
> > saying, have access to some draft effect or grip or speed that they
don't
> > during the race or that the human player doesn't have...??  What would
be
> > the point of this?

> > It is my belief that the problem in the first paragraph was "solved"
with
> > the adaptive AI and Auto AI.  I have no idea what the design issue is in
> the
> > second paragraph, but it was there in N3, N4 and NR2002.  Needless to
say,
> I
> > wasn't impressed when I saw it again in NR2003.

> > Marc

> > PS, you say "I haven't seen you give any feedback of you verse the AI in
> any
> > races, just your hot laps compared to their practice sessions."  If you
> read
> > even just the message you were replying to you will see that I am not
> > talking about just practise mode any longer.  I noticed the problem
there
> > first when I originated the thread, but subsequently confirmed that in
my
> > experience the trend is the same in races.  Of course you are correct
that
> > lots and lots of testing would need to be done to corroborate this,
which
> is
> > impossible for any one person with an outside life, hence my talking
about
> > it here and attempting to get more opinions and experience.  Instead,
> Uncle
> > Fester seems to have said what I already suspected...


rec.autos.simulators is a usenet newsgroup formed in December, 1993. As this group was always unmoderated there may be some spam or off topic articles included. Some links do point back to racesimcentral.net as we could not validate the original address. Please report any pages that you believe warrant deletion from this archive (include the link in your email). RaceSimCentral.net is in no way responsible and does not endorse any of the content herein.