January 23, 2018, 07:44:42 AM
bigger smaller reset 800px Wide width Full width Reset * *

NRTracks

 
Welcome, Guest. Please login or register.

Login with username, password and session length
« previous next »
Pages: [1] Go Down Print
Author Topic: Grip Settings  (Read 14634 times)
Lyndon
Administrator
Sr. Member
*****
Posts: 147



WWW
« on: February 10, 2011, 01:46:30 PM »

I ran some tests awhile back to test grip characteristics (using my AAA test track that I created for just that very purpose i.e. the track has few sections and therefore it is very easy to change all track properties). I was going to publish the results in a spreadsheet but I didn't notate very well and it's somewhat confusing (even for me at this point). So I'll just write a little synopsis for now and summarize what I think I found.  think

I conducted the tests using 100% AI and very basic values in the Track.ini. The only values I changed were the grip numbers (for each test iteration). The major changes were to the track--I changed the surfaces from 'all asphalt' to 'all concrete' to 'all paint', etc. Results weren't all that surprising but I wasn't looking for surprises; I wanted numbers, something to tell me how much difference I could expect and what it would take to equalize those numbers.

In a nutshell: asphalt is fastest.  Cheesy  Here are basic results with individual grip properties set at 1.00:

  • surface -- avg. time -- dif (vs. asphalt)
  • asphalt -- 27.797 -- 1.00
  • concrete -- 28.365 -- .98
  • paint -- 29.167 -- .95
  • gravel -- 33.194 -- .837
  • sand -- 33.249 -- .836
  • dirt -- 34.065 -- .816
  • grass -- 36.280 -- .766

Don't think that you can apply those percentages to increased grip and normalize the results though. For instance, concrete is 98% as fast as asphalt so increasing the grip by 2% (grip x 1.02) should give similar results right? No, it just doesn't work. As further experiments proved you would have to increase concrete grip by closer to 4% (1.04) to get speeds comparable to a standard asphalt grip of 1.00. Paint results are more disparate. Raising the paint grip by 5% (1.05) only gets paint track times to about 97% of asphalt--you have to raise paint by about 10% to get speeds roughly equal to asphalt. As for other surfaces? Well why would you want to? Gravel, dirt you're dealing with dust as well as grip problems. But just for the sake of argument, dirt is about 18% slower than asphalt. Yet increasing dirt by a full 24% (1.24) still leaves it 10% slower than asphalt. I didn't test this further to see what it would take--it just seems futile. You get the idea though--there's some kind of logarithmic (or some other math-sounding term) thing going on here.  blink

Maybe at some point I will repeat the tests, possibly changing criteria and maybe even make a comprehensible spreadsheet.  But this gives a comic balloon envelope to what I found. Feel free to dispute my results or add your own observations--I'm merely on a curiosity-driven mission here. I seem to have no problem sharing my ignorance, do I?
« Last Edit: February 10, 2011, 01:59:33 PM by Lyndon » Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #1 on: February 11, 2011, 03:57:42 AM »

Nice try, but it's a lot more complicated than that.
(I'm currently without NR2003 access, so my observations/opinions are based on my experience developing ai for various tracks.)

First, I would say the test(s) should be conducted on both an oval and RC.
Keeping all other variables constant, more grip can decrease laptimes up to a point, where the increased drag from superior grip actually slows the car down.
As a general rule in the real world, concrete has more grip than asphalt. Concrete is faster than asphalt in corners, if not everywhere, but concrete at higher grip values "may" be slower on straights (because of the drag) than asphalt. I say may be slower because it depends on the range of values you tested with. My experience has been that concrete is normally faster than asphalt up to a point, where the extra grip has more drag than is beneficial. The same applies to asphalt, i.e., asphalt with a higher grip value is faster than asphalt with a lower grip value, up to a point. When that point is reached, laptimes can still be reduced, but the car setup has to be freed up to take advantage of the added grip without incurring the wrath of overwhelming drag.
I've found the "normal" grip values to be 1.00 - 1.075. Anything higher requires freeing the setup to keep from slowing down, but doesn't really provide a noticeable decrease in laptimes.
You have to get into the 1.5 - 2.1 range to see a noticeable decrease in laptimes, as compared to the "normal" range. This higher range will produce faster times than the normal range using the normal range setups, but to really get substantially faster you have to free up the setups.
These normal and high ranges are just that, a range of values that yield substantial results. Lower laptimes may be reached with slightly higher values in each range, but generally, for the slight extra speed, I've found the "feel" of the car goes beyond what the physics of the car would allow. (Of course that's my interpolation of what a race car (whatever series) can do.)

Re: your specific results...
At a value of 1.00 I would expect concrete to be faster than asphalt. I would want to make sure the race lp, fast setup, and the rest of the track.ini was optimal. For example, parameters, such as slip angle, can affect the cars performance. If the car slides too much, the ai sometimes catch the slide which causes the car to "dig in", so to speak. If the surface was asphalt, it would dig in less than concrete, thereby not slowing as much, hence, a faster laptime. With the fairly small differences in results, the fast setup might favor the looser asphalt over the grip-ier concrete; a setup for the concrete might make the asphalt too loose, and tip the results in the concrete's favor.

When I'm able to get back into it, I'm gonna check this out.
Thanks Lyndon, I don't have enough things on my plate. (Did Radt have anything to do with this?)

Just MHO.
14 angel
Logged
Lyndon
Administrator
Sr. Member
*****
Posts: 147



WWW
« Reply #2 on: February 11, 2011, 06:11:46 PM »

I'll be here--waiting.  Cheesy Ok. I was painting with a pretty wide brush and you've piqued my interest by adding other factors. I look forward to seeing where this goes. As for Radt... Is that smelly old reprobate still around?
« Last Edit: February 11, 2011, 06:25:39 PM by Lyndon » Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #3 on: March 16, 2011, 09:18:29 PM »

I had a little trouble getting started, I'm not attuned to ovals. In fact, I spent about 30 minutes just trying to get some speed on the default asphalt ptf. I tried a multitude of setps, some of yours and some from California (which is fairly close banking-wise and where I used to be rather proficient). I gave up from frustration.
I started thinking of other tracks that could make this easier,such as Talladega, where you could just mat it. But that could have an effect on the straightaway top speed due to possible drag from increased grip. I decided there were too many variables no matter what the track, oval or RC. First and foremost, a human is telling the ai how to drive, with the race.lp and the exe. Second and related to the first, the human must ring out every thousandth from the track to attain a baseline from which the ai can be developed to run the fastest and most consistent. That would reduce the number of anomalies that could influence the results, IMHO.

I didn't do that. I'm not used to driving ovals, and to get good enough would take more time than I'm willing to spend.
So, I loaded the concrete ptf. Right off the bat I was better because, as I pointed out earlier, concrete physically has more grip. I practiced long enough to get consistent, but I was using my own setup. I checked your posted ai times to see where I was. I was running 100% ai asphalt laptimes on the concrete. I thought an ai comparison was in order, so I loaded up race mode with only a !00% ai guy. I didn't want any interference from traffic. I switched to the fast setup, which I could now drive.
To my amazement, the ai was only doing 30's. We ran some laps and I saved the results.
Went back to asphalt ptf. Same general results, me vs. ai (now doing 31's), but the laptimes were slightly slower. Exactly what I expected.

To quantify the results, i.e., how much difference between asphalt and concrete, I say it's track dependent. The difference on a RC is going to be larger than the difference on an oval. I suspect the difference on a highbanked oval won't be the same as a flat oval, a SS vs. a short track, etc. I just know from experience, if you want a little more grip, go with concrete. A value from out of nowhere, concrete is probably worth asphalt at + 0.005 - 0.01. (I'll have to try that sometime.)

The question I have, how come your 100% ai is faster than mine.  yikes
Huh? think
 
 cheers


* A v C comparison.jpg (121.29 KB, 587x847 - viewed 712 times.)
Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #4 on: March 17, 2011, 02:18:34 AM »

I went back to the track, I don't remember for what reason. Anyway, I was looking at the tack.ini (for what I don't remember) and I see concrete grip is 1.10. ohmy Lyndon, you weenie. Tongue I just spent all that time testing and it only confirms that concrete at 1.10 is faster than asphalt at 1.00. Duh.
I re-tested the concrete at 1.00, and now it's not as fast as asphalt. I know this isn't right. Now, to figure out why.
For one, since concrete has more grip, the tires slide less so tire heat is less. If the tire heat is correct for asphalt, the concrete version is being crippled with too cool tires. I lowered the tire heat to 0.95 (a guess) and loaded the asphalt version. Laptimes were reduced but still still a tad quicker than concrete laptimes. Tire temps on the RF were closer to the concrete temps. I didn't record exact temps for either version, I was just searching for an anomaly that would explain why asphalt would appear to do better than concrete.
This type of variable was what I was alluding to in my first post. Sorting all the variables is why it takes so many laps (both player and ai) to get good ai.

I'll play with this some more, but I don't forsee any useful "formula"(?) being derived from the findings.

This still doesn't explain your 100% ai being faster than my 100% ai. I think (know?) you must have used a different ptf (other than the 5- 10 banked) or track.ini than what I found in the track folder. We are talking about AAA tester?

 cheers
Logged
Lyndon
Administrator
Sr. Member
*****
Posts: 147



WWW
« Reply #5 on: March 18, 2011, 12:01:32 PM »

My AI are faster? Hmmm... sensible and hard-charging drivers? Seriously though, I don't know. I never jacked with the ratings or rosters of the original Cup, GNS, or CTS series... just the PTA. Do you have some forgotten papy.ini tweak or some such thing present? I'll have to check my notes but it seems I used the 5-10 banking config (since I struggle with the flat ovals). And all of my testing was done with the AAA test track.

Sorry about the 1.10 concrete grip--I can't believe that I didn't reset all the values to 1.00 since those were the only values I was changing. Well, maybe I can believe it. What I can't believe is that you didn't check it first, you knowing me.  Tongue

Yeah I know I could have changed a lot of variables and re-checked... it was mostly a "all other things being equal" test (although I did water the entire track down before testing  biggrin). All my "life" experiences (as a sometimes 'crete man in the trades) have indicated that concrete is supposed to be slicker than asphalt.  So you still haven't convinced me... roflmao

Edit #1--I went and looked at my upload and for some reason I had 4 different INI files included. I reset all of them before uploading except for the default.  blush
« Last Edit: March 18, 2011, 01:20:13 PM by Lyndon » Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #6 on: March 18, 2011, 03:58:27 PM »

From "The E#ngineers Handbook".


* engineershandbook.jpg (32.17 KB, 638x222 - viewed 665 times.)
Logged
Lyndon
Administrator
Sr. Member
*****
Posts: 147



WWW
« Reply #7 on: March 20, 2011, 07:35:25 AM »

You're gonna use science against my tennis shoes and Radt's boots? Seems kinda unfair since Radt is scientifically challenged. biggrin
Logged
Radt Dastard
Member
*
Posts: 3



« Reply #8 on: March 20, 2011, 07:38:48 AM »

That's right--I defy scientific explanation.   crazy Say, how come that Waldrip fella says drivers at concrete tracks always sez "my tires feel like basketballs?"
« Last Edit: March 20, 2011, 12:29:28 PM by Radt Dastard » Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #9 on: March 20, 2011, 06:57:08 PM »

Say, how come that Waldrip fella says drivers at concrete tracks always sez "my tires feel like basketballs?"

Remember, you're quoting a guy that says "B___ity, B___ity, B___ity". Okay, it was different (?) for a race or two a couple of years ago, but I had really hoped that would have been retired this year. I guess $$$ overrides good taste or common sense. dry
Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #10 on: February 13, 2013, 02:23:57 AM »

Grip level value for NR2003
(Do I want to go sim racing or play a video game)



Applied grip level via the track.ini

Asphalt is asphalt.
For the same series (mod), why should you need track_asphalt_grip = 1.00 at one track and track_asphalt_grip = 1.10 at another track?

You shouldn't, and you don't.

Race cars tend to have a high HP to weight ratio. Without traction control you can't mash the throttle at will without getting wheelspin
   and a potential loss of control. Without antilock brakes you can't jump on the brakes without locking it up and losing control in a slide.
Race cars, no matter the downforce available, have some degree of drift in high speed corners. They do not track like they are on a rail.
These are real world physics/circumstances.
This is a racing sim. It's not reasonable to expect to be able to jump into any car (series) and immediately drive at peak performance,
   nor should it be likely.

I'm not a proponent of the idea that certain track types require higher grip values, especially not the 1.10+++ values being used for every
   type track (and series) lately.
Learn car control and put in plenty of practice time. Learning how to adjust a setup wouldn't be a bad idea either. These are the traits
   of a sim racer. Disregard for these practices indicates someone who's looking for a video game experience, not a bad thing, it's just not
   sim racing.

The reason to increase the grip level for the same series is:
   1) The setup is inadequate. The grip level is raised to compensate for a crappy handling car to reach a preconceived level of performance.
   2) You're looking to be able to apply more throttle more quickly, or just generally have an easier time negotiating the track. This is not sim
         racing, it's a video game.

Reasons for using a constant grip level (for the same series) for all tracks:
   1) Loose is fast. When a setup proves (by laptimes) to be as fast as is possible on a specific track, and you raise the grip level
         significantly (e.g. from 1.00 to 1.10), the setup is no longer as fast as is possible. If the setup is freed up the laptimes will come down.
         The more outrageous the grip level, the more the adjustments will excede "normal" ranges. When the grip level is 1.2 or more, the setup
         has less and less effect and at 2.0 the setup essentially becomes moot.
   2) The more outrageous the grip level, the more the track.ini ai adjustments will be outside the usual range, making it more difficult to get
         the ai to act/drive like a real person.

   ***   ***   ***   ***   ***   ***

What grip values are appropiate?

Asphalt grip?
For the stock car series (cup, gns, cts), 1.03 is more than adequate.***(see below).  This also works for pta.
Gtp requires 1.05, 1.03 just doesn't provide the added grip of the aero-downforce of that series (even though the modified exe should have/is
   supposed to have provided it).

Concrete grip?
Concrete and paint have their own physical properties (coefficient of friction). When the grip value is the same as asphalt, concrete will have
  slightly more grip and paint will have less.
When used for higher grip situations to enhance a particular situation or remedy an ai problem, the usual concrete grip value will
   be 0.005 - 0.01 greater than asphalt.

Paint grip?
When running on a paint surface in a straight line or on a narrow surface (such as a paint line), the grip difference from asphalt is not noticeble.
   However, when running on a lane width or more of a paint surface-type while the car has significant lateral G-loading, the car will slide far
   more than it would on asphalt or concrete.
What grip value to use? It depends on the intended application. For lines, the same value as asphalt. For raised curbs, the same value as asphalt
   or up to 0.01 more than asphalt.

Special circumstances.
The original OWR with the modified exe provided an adjustable front wing which creates downforce on the front of the car and in conjunction with
   the rear wing balances out the overall downforce on the car.
OWR using the pta exe, OWR07, requires grip values far beyond what should be considered "normal" because there is little front downforce supplied by the exe.
   The downforce of open wheel cars far exceeds what the pta exe provides, so very high mechanical grip must be applied to compensate. However,
   just raising the asphalt grip will not provide the car response that would be associated with open wheelers. Super-high asphalt grip does not allow
   the car to drift in the corners, even with a very loose setup. Super-high asphalt grip on the straights induces too much drag, reducing top speed.
   OWR07 requires the track to be a combination of surfaces, each with their own grip value, and each surface placed to
   produce a specific grip level for specific track conditions.
Asphalt: Generally the lowest surface-type grip value, 1.50. Use on the straights.
Concrete: Used for braking zones, high speed kinks, and VERY slow corners, 1.7 - 1.8.
Paint: Used for faster corners since it will allow the car to drift, 1.9 - 2.1.

The other surface types.
Invisible: This can be used for special circumstances, but this surface produces no tire sounds.
Rumble: For raised curbs. Rumbling sound conjoined with this surface type.
Grass: <1.00, the physics properties as supplied by Papyrus include an irregular surface, too high a value exaggerates the irregular surface to the
           point of being "bumpy"
Sand: <1.00, can be used for reduced grip hard surface but it produces dirt plumes.
Gravel: @ 0.75, almost stuck if momentum is lost, but there's just enough traction to get out (eventually)


*** Tire Heat:
If there isn't enough tire heat ([ track ] or [ track_(series)] track_tire_heat = x.xx), the available grip won't be utilized.
You'll want the tire temp (on the hottest tire) to be 200 - 210F* the maximum amount of time during a lap. Long, fast straights raise tire temps
   dramatically, so it's okay if the temp exceeds 220 as long as it drops back to 210 or less fairly quickly (or at least before you really need the
   tires to work at full capacity). Very long straights raise the tire heat, the optimal tire heat value may have to be reduced to compensate for
   this circumstance.
* RC temperature range. Other chassis types run higher tire temps, use the highest "white" (the next stage is yellow) temp -10 degrees as the
   upper range target and -10 more degrees for the lower range target.



This is my opinion.  Tongue
Your opinion may vary.  Sad


« Last Edit: December 14, 2017, 12:27:08 AM by fortine_oo » Logged
Amtronic
Member
*
Posts: 11


« Reply #11 on: March 16, 2013, 09:59:29 AM »

I have a wrench to throw into your works.
What if a driver uses Traction Control in the Options - Controls window?
In my case, I must have this option enabled since I use a joystick. Yes I have
a wheel/pedal combo, but it takes up a lot of space that I need. So for checking
a track layout and just general ease of use, I use the joystick since the footprint
is so much smaller. Some people even use keyboard control. The end result is
that the throttle control is not linear but off/on and traction control is mandatory.
Now you have the hard-coded Traction Control function bumping against the
traction settings so carefully set. From practical experience, there isn't that much
difference between 1.0 and 1.5 settings for a given track surface type. And the
subtle difference between asphalt at 1.0 and concrete at 1.0 are nonexistent.

 cheers

Logged
fortine_oo
Legacy
Sr. Member
*****
Posts: 188



« Reply #12 on: March 22, 2013, 02:10:50 AM »

I have a wrench to throw into your works.
What if a driver uses Traction Control...

I fail to see how 'traction control' would alter or conflict with the "grip level" thesis.

IMO, traction control is tantamount to video game. Driver's aids are "designed to be a disadvantage to the experienced racer" *.
   [* NASCAR Racing 2003 Season manual]
I wouldn't delineate between high grip values and traction control, both are outside the realm of realism for racing series that don't allow
   traction control (or anti-lock brakes), therefore their use is not within the definition of a simulation.


From practical experience, there isn't that much difference between 1.0 and 1.5 settings for a given track surface type. And the subtle difference between asphalt at 1.0 and concrete at 1.0 are nonexistent.

I'm not sure what "that much difference" is referring to. The ease of car control? Laptimes? Feel/Force-Feedback?
From my experience, a grip level value increase of 0.50 will most certainly decrease laptimes. It will increase car control to the point of
   lessening the overall effect of a specific car setup.
 
As for the difference between asphalt and concrete at the same value, that's a distinction that can only be recognized by it's application,
   and the distinction is further influenced by the grip level value, the type of track, and the car setup. The real difference can only be
   determined by the difference in laptimes. However, with prudent application, the difference is readily visible when head to head racing
   with ai; with small grip level differences, the ai do not reap the benefits of slightly better grip as definitively as the Player.
The difference is more noticeable on a roadcourse than it is on an oval. On an oval, if the concrete setup is not freed up from the asphalt
   setup, oval laptimes can actually be slower on concrete. On a RC, a racing surface that is entirely concrete can be slower than an
   asphalt (straights)/concrete (corners) combination. The combo surface scenario can also be true on an oval, but the overall benefit is
   dependent on the oval's configuration, i.e. the tighter the corners, the greater the benefit. These are a few examples, infinite variables
   yield infinite outcomes.
The difference between asphalt and concrete is most notable when hard braking is required. Since the coefficient of friction is higher
   for concrete, braking distance is shortened because greater brake force, enabled by greater tire to pavement friction, can be
   applied (before wheel lockup).
Logged
Pages: [1] Go Up Print 
« previous next »
Jump to:  

Powered by SMF 1.1.15 | SMF © 2006-2008, Simple Machines
Leviathan design by Bloc | XHTML | CSS