Our ReVolt Pub

Actually, players and the AI run the same car totally different sometimes. Ratings set by your playing would be off for the AIs, and ratings by the AI behaviour leave them helpless against tricks that players can do with the cars. Maybe the latter is better. Or maybe there’s no point in elaborating the ratings into anything more complex than just based on top speed.

It would nice to have some kind of test feature where you can select some cars and a track for the AI to drive and then based on these settings calculate the outcome of the race (like doing a race without all the visual stuff and timer limitations).

That would certainly help a lot figure out how well certain cars perform against other cars (and in what environment).

@ Citywalker: It's a fine line to walk. I see both sides. If the cars are updated based off of user performance, then AI-driven vehicles with bad AI parameters will be overrated. But if cars are updated based off of AI performance with bad AI parameters, the user-driven vehicle has a chance of being underrated. But it's undeniable that anything semi-close to the right car ratings with two extra ratings to choose from is going to be a heck of a lot better than having five ratings and trying to squeeze cars with speeds of 40 mph on up to 100 mph all into the Pro category. However, relying on top speed alone just isn't good enough for all types of cars. It at least requires an in-game lap test of some sort to be anywhere close.

I have yet to try out any cars with the "expert AI" parameters, but is that a fixed set of AI parameters that makes them handle well or are they individually tuned? If it is the same for all cars, then I could just make that the AI for all vehicles and then test them out with the AI. Then both problems would be solved. I doubt that they all would be the same, but just throwing out an idea.

@ Kenny: Yes, that would be nice. It would be a big time-saver. But for a game like Re-Volt, it might be more time-consuming to create that type of program than it would to go ahead and test all of the cars by hand, idk. Something like that would probably either be scarily precise or horribly inaccurate.

This is what I do. Run a lap of Nhood1. If you can get a lap time of under 40 seconds without effort then the car is Super-Pro. If not, then it's then Pro. It's easier to guage between the other classes.

I tried letting the AI run some cars for rating purposes, and it’s no use, at least with the old AI. In any track, if they collide with something, they may end up between two adjacent track zones which means that they remain stuck there for a totally random time, which messes up all attempts of measuring the lap times.
And V1.2 AI is still somewhat messed up (sudden excessive countersteer in some curves at sometimes), so it doesn’t qualify for tests in my eyes.
And anyway, if measuring by AI runs, the rating would change as soon as someone changes the car’s AI (after all, it _is_ possible to tune the AI now).
This also answers your question, Mandolin: Expert AIs are tuned for each specific car.
And Skarma, player handling depends on the player. Everyone would get different results.
So in my opinion, the best (and currently only?) way to rate cars is to measure their in-game top speed and acceleration (speed at 5 seconds?), using e.g. Test Track 6 (straight concrete). This is reliable enough to use (doesn’t depend on player or AI), and takes into account that some cars recover better (good acceleration) with same top speed.
Handling performance is not reliable to use, because how do you measure it, independently from specific players? And it is a complex thing that can be exploited by players against AI in various ways, no matter how you rate it.

Just my 52 cents.

@ Skarma: That's actually the exact way I tested my smaller collection of cars for Super Pro earlier. Same idea, same track, only I considered 42 seconds or less a Super Pro car and 35 or less an above Super Pro car, but I think you've been playing since I was a little bitty thing if you've been playing since close to the release date of Re-Volt, so you would most likely be a little better of a racer than me. As Citywalker said, it really just depends on the player. But I think you're right, especially after Citywalker's test (thanks!!!)... user tests would be better.

And I did make a few exceptions for cars I couldn't quite get a good time on because of handling, but that I figured would be possible to get the right times on. You can almost feel just from racing a car around a little, regardless of time or number of laps or anything, what rating it would fall in.

The only problem with using the test tracks is that some vehicles that just handle downright horribly do not deserve to be in the Super Pro class or higher. Yeah, for that one track a car with horrible handling might be able to win, but it is possible that it wouldn't easily be able to beat even a Semi Pro or Advanced car or lower otherwise. In my opinion, if you can win Pro races with a car like that, you deserve to win!

Mandolin @ Feb 21 2013, 02:41 AM wrote: @ Skarma: That's actually the exact way I tested my smaller collection  of cars for Super Pro earlier.   Same idea, same track, only I considered 42 seconds or less a Super Pro car and 35 or less an above Super Pro car, but I think you've been playing since I was a little bitty thing if you've been playing since close to the release date of Re-Volt, so you would most likely be a little better of a racer than me. As Citywalker said, it really just depends on the player. But I think you're right, especially after Citywalker's test (thanks!!!)... user tests would be better.

And I did make a few exceptions for cars I couldn't quite get a good time on because of handling, but that I figured would be possible to get the right times on. You can almost feel just from racing a car around a little, regardless of time or number of laps or anything, what rating it would fall in.

The only problem with using the test tracks is that some vehicles that just handle downright horribly do not deserve to be in the Super Pro class or higher. Yeah, for that one track a car with horrible handling might be able to win, but it is possible that it wouldn't easily be able to beat even a Semi Pro or Advanced car or lower otherwise. In my opinion, if you can win Pro races with a car like that, you deserve to win!

The Super Pro Car Will Done In Next Released Version?
Edit:The Super Pro Car Can Be Unlocked By Beating Titanum Cup And Win All Track In Insane Tracks And Collect All Star In Insane Tracks

My working of basing on the 40 second laptime comes from one of Dolo's videos where he recorded a lap time of 40.512 using Cougar on Nhood1, that is ridiculous. I think he has since beaten that with Toyeca but I've never come across any mentions or anyone ever getting even so much as 0.001 seconds UNDER 40 seconds in any of the stock Pro cars, without cheating and NOT counting any of the clockworks or the Probe UFO. Considering that Toyeca is Pro class, that's what should be capable of a car when it's really put to the test. Anything that can do that without putting in effort is Super-Pro. Makes logical sense, no?

As someone who messes around with parameters quite alot, it's pretty amazing what you can actually fit into each class. I made a car that does 47 MPH top speed but it's Semi-Pro for a good reason.

To me yes. My best time (today, at least, when I actually looked at my lap time and gave it about 15-20 tries) with Toyeca is about 43.6 on Nhood1. And that's running the track many times and working off of the ghost car. Without putting extra effort in, I tend to get about 46 or so on average. So for me, 42 and below on the first lap or two without the extra effort would be significantly better than a Pro car. It's just going to be different for each person like if somebody else were to say they would help with this (*HINT HINT* lol) they would not use 42 seconds unless they were about on my level. It would either be more time or less time.

@ Huki/Jig (if you happen to look at this): Would you consider putting any of this into 1.2? Haven't really heard much about it from y'all and that really kinda determines whether or not it's even worth talking about. Lol!

Mandolin @ Feb 22 2013, 04:15 AM wrote:@ Huki/Jig (if you happen to look at this): Would you consider putting any of this into 1.2? Haven't really heard much about it from y'all and that really kinda determines whether or not it's even worth talking about. Lol!

Yep, we (I...) actually consider doing something about it. As far as I'm concerned I will likely see things that way: an extra Super-Pro rating officially supported, and maybe an extra rating unofficially added for cars that goes ridiculously fast as well.

Text Removed
Edit:Is Continue Working?

Platform @ Feb 23 2013, 05:28 AM wrote:Next Re-Volt 1.2 Alpha Build Release Will Soon?

This kind of question tends to be useless...

Yep, we (I...) actually consider doing something about it.  As far as I'm concerned I will likely see things that way: an extra Super-Pro rating officially supported, and maybe an extra rating unofficially added for cars that goes ridiculously fast as well.

Sweet!!!

So you think the feature would use a new line in the parameter code to maintain backward compatibility? or some other method? I could go ahead and start testing as much as time allows me. As I said, it will potentially take me a long time to do all of it just because of school and playing at concerts and other stuff being on my plate. So if ANYBODY wants to help out with the testing (it doesn't have to be a lot, anything is helpful), it would make the process of getting Super Pro + cars into Re-Volt a lot shorter.

@ Jig: Over half a month since last reply... I know that the collision thing is a big deal right now and that it takes priority over this, but I could go ahead and be working as much as I can on this if I knew for sure how this is going to be implemented (a new parameter line for Super Pro+ cars? If so, what will that line be so I can make appropriate parameter adjustments? Would there be two extra ratings for sure?).

@ ZipperRules (you approve all the car submissions on RVZT, right?): If the new car ratings get going pretty well, many of the old files would have to be replaced by the new, tested cars, and that in and of itself would take a fairly lengthy amount of time to do. Would you be able to do that if I send you all the files incrementally? And are there any cars on some sort of queque to be released on RVZT? I know you usually keep just ten new cars on the site about every week so they all get decent showing time on the sidebar. If this gets started, the cars on the waiting list would have to be updated as well, and maybe for newly submitted cars until people get used to using the Super Pro and extra rating for their faster creations. Is all of that ok?

If you really do go about doing this, leave all of my own cars to me, I'll deal with them.

Ok, THANKS VERY MUCH. I will. That'll take care of well over 100 of the cars, which would cut a chunk out of the testing time.

@ Jig: Another week's gone by (and nearly a month since last reply ) I really wanna start on this cauze I may not get the chance once college kicks in all the way for me. And I'm not sure if anyone else would do all of the work for this... so, any news for how the programming side of things would work? New param line? One or two new classes?

I don't mean to be pushy in any way lol I just don't want to run out of time to work!

Umm... jigebren are you ok? You haven't really posted much all month, just one reply to some other topic that I saw. I tried PMing and haven't gotten a response yet... it is kind of now or never for me. I won't be able to do this if I don't start soon.

Jig wasn't active here for much longer, so don't worry. He's just busy or don't have time, that's all.

Same for Huki I guess, I hope they are ok.

Hey sorry for the delayed post. I did talk to Jigebren but things are starting to get hectic for me with graduation coming up in less than a month. I haven't really had time to do much of anything with Re-Volt and I don't think I will be able to for a while. Sorry guys. Maybe some other people could pick up this project. It can be a reality within Re-Volt limitations, all it needs is a few people who are willing to change all the car parameters on RVZT (assuming that the guys over there don't mind; I haven't heard an okay or a nay for that).

I say better wait until your graduation and other stuff in real life are over and do this beautiful job yourself, the situation here will be practically the same than it is now.

Since Huki & Jig are dissappeared and noone knows what the next Patch is going to be, we are all kind of stopped right now. Waiting...

Phantom @ Apr 27 2013, 12:33 AM wrote: I say better wait until your graduation and other stuff in real life are over and do this beautiful job yourself, the situation here will be practically the same than it is now.

Since Huki & Jig are dissappeared and none knows what the next Version is going to be, we are all kind of stopped right now. Waiting...

Next Version Soon To Release New 5 Super Pro Cars?

Platform @ Apr 27 2013, 06:51 AM wrote: Next Version Soon To Release New 5 Super Pro Cars?

You don't understand anything, do you?

Next Patch Will Released At:rv1.2a13.XXXX Oncoming

Mandolin @ Mar 12 2013, 05:41 AM wrote:... I could go ahead and be working as much as I can on this if I knew for sure how this is going to be implemented (a new parameter line for Super Pro+ cars? If so, what will that line be so I can make appropriate parameter adjustments? Would there be two extra ratings for sure?).

@ ZipperRules (you approve all the car submissions on RVZT, right?): If the new car ratings get going pretty well, many of the old files would have to be replaced by the new, tested cars, and that in and of itself would take a fairly lengthy amount of time to do. Would you be able to do that if I send you all the files incrementally? And are there any cars on some sort of queque to be released on RVZT? I know you usually keep just ten new cars on the site about every week so they all get decent showing time on the sidebar. If this gets started, the cars on the waiting list would have to be updated as well, and maybe for newly submitted cars until people get used to using the Super Pro and extra rating for their faster creations. Is all of that ok?

Mandolin @ Mar 20 2013, 04:53 AM wrote:And I'm not sure if anyone else would do all of the work for this... so, any news for how the programming side of things would work? New param line? One or two new classes?

I don't mean to be pushy in any way lol I just don't want to run out of time to work!

Mandolin @ Apr 26 2013, 10:24 PM wrote:Maybe some other people could pick up this project. It can be a reality within Re-Volt limitations, all it needs is a few people who are willing to change all the car parameters on RVZT (assuming that the guys over there don't mind; I haven't heard an okay or a nay for that).

zipperrulez wrote:...he wants is to add two new classes of cars above Pro, and he wants to download every single RVZT car and reclass them.

Disregard his idea, or tell him flat out "No", there's no way that's ever going to happen.

ZR

My personal point of view is this;
The idea of implementing other classes above "Rating 4" would be great.
My understanding suggests that if RV reads the Rating value then it just needs to be told to allow classes above 4. No real additional editing to the param files needed. This could allow for more than just two extra classes.
Then any new car can have a class rating based on a generally agreed class per speed scale, any older cars can be altered/edited by individuals as and when required.

Of cause this is only realistic if there is a persuasive majority interest in this project area, and only if the project managers/developers have actual free time outside of other personal commitments, which do tend to take priority.

And with this, Manmountain just wrote down all that needs to be done. This and nothing more.

My understanding suggests that if RV reads the Rating value then it just needs to be told to allow classes above 4. No real additional editing to the param files needed. This could allow for more than just two extra classes.

Because most people will use their own ratings anyway, based on their own preferences, their own skills and their own racing set of downloaded cars.

...he wants is to add two new classes of cars above Pro, and he wants to download every single RVZT car and reclass them.

Disregard his idea, or tell him flat out "No", there's no way that's ever going to happen.

ZR

Thank you for the reply, ZR (and thanks MM for forwarding it). I understand your reasoning because if I were to go through with this plan then it would indeed be a lot of work for you also, and it might mess up the flow of RVZT (idk how your dates work, but it might reset them all if there's no way to change them). I guess that settles that. No work for me!

I don't recommend ONLY having the extra ratings allowed in Re-Volt because, for one, it would cause compatibility issues with 1.1. Also, this would practically make it mandatory for everybody to do rating changes manually if they want to use old cars and new cars together (all the new cars in the new classes would not be able to race against any of the old cars in the regular classes without all the tedious work of changing ratings). That pretty much defeats the purpose of having the new ratings unless EVERYBODY wants to change their own ratings (which I'm sure they don't because nobody has volunteered to help with this except for Skarma who said he would do his own cars). I think this idea may have come really close to being a reality, but at this point I just can't see a way around the obstacles. Thanks everybody for supporting the idea and maybe someone else could figure out some way to do this.

If Re-volt v1.2 gets a minimum bug final release then everybody would be using it without a second thought as we all use at least the 1207 patch not just the original release. So that would get rid of the compatibilty error.
As for individuals changing thier own set of well loved and used cars, I would think this wouldn't be that much of an hassle.
The point of the different class ratings is to give a selective range of cars their own class so that only that selection gets chosen when Re-Volt selects the cars to match your own car selection. So, as everyone prefers their own choice of cars they wish to race against, their choice of class settings my differ from anyone else's and so individual parameter/class rating changes will occur naturally.

As for interest in helping, I would be glad to help changing class settings for all the cars that might exsist (currently 2500 at RVZT), but the problems arise of:
How many classes ?
What definition, guide or rule decides which car goes into which class ?
Would there be a majority agreement on this ?

I think these three questions would create a massive community debate which would not be decided any year soon.
So as long as v1.2 had the facility for more classes, individuals would probably prefer to class their own collection of cars as they would like IMHO.

The fact is we all wan't a defined item we all can use together, but we liked to define it from our own individual point of view.

Manmountain @ May 3 2013, 03:19 PM wrote: How many classes ?
What definition, guide or rule decides which car goes into which class ?
Would there be a majority agreement on this ?

Personally the only setup that makes sense to me is this (a little bit of that has been mentioned here already):
- 2 additional classes (perhaps 3 if we want to have general satisfaction)
- in one of those classes go all the cars that are a little bit better than the current Pro class (by little bit I mean a difference like with Rookie to Arcade, Arcarde to Semi-Pro and Semi-Pro to Pro)
- in the other class go all the cars that are overpowered (with an insanely good handling/acceleration/top speed)
- in the last class (if there is one) go all the other cars (that are undriveable or are simply not suitable to race against other cars or that the user wants to be seperated from all other cars)

The only questionable thing here is how we define are car to be better than another car? I guess the most important factor is top speed here, then one needs to analyze the handling (how difficult it is to control in a curvy track), then the acceleration and lastly the mass.

If someone has a better suggestion, feel free to post it or simply discuss.

However I think modifying all the cars on RVTZ to go along with new classes is kind of pointless but I guess thats not up to me to decide.
Last but not least, this discussion is also pointless if we don't have any promise from the devs to implement additional classes.

What we can do now is support something called "user ratings": these are unnamed ratings that would simply be displayed as User 1 (for a Rating of 5), User 2 (for Rating 6), etc.

Updating existing cars is not really important. When enough cars of a rating are not available, re-volt automatically selects cars from the previous rating - this is already the case. It means at first when you only have one or two Rating 5 cars installed, you'd mostly be racing against Pro cars (the current behavior). As you add more Rating 5 cars to the game they will gradually take precedence over Pros.

If compatibility with old versions is required, the solution is to include two Rating lines, like this, This would ensure that old versions use a Rating of 4 (and skip the second line), while the new v1.2 would give preference to the second line. Compatibility is optional and left to the choice of car authors, but note that old versions could crash if they detect cars with Rating above 4.

For now I'm thinking about only allowing 2 extra ratings (i.e., Rating 5 and 6), so it should be similar to jigebren's idea (the only change is that both ratings remain unofficial). An interesting idea about individual car packs was discussed, but I'm not sure if should have anything to do with the Ratings variable...

Kenny @ May 3 2013, 08:18 PM wrote:Personally the only setup that makes sense to me is this (a little bit of that has been mentioned here already):
- 2 additional classes (perhaps 3 if we want to have general satisfaction)
- in one of those classes go all the cars that are a little bit better than the current Pro class (by little bit I mean a difference like with Rookie to Arcade, Arcarde to Semi-Pro and Semi-Pro to Pro)
- in the other class go all the cars that are overpowered (with an insanely good handling/acceleration/top speed)
- in the last class (if there is one) go all the other cars (that are undriveable or are simply not suitable to race against other cars or that the user wants to be seperated from all other cars)

For the "undrivable" cars, you can simply stop them from being selected for AI, by setting CPUSelectable parameter to FALSE.

Kenny @ May 3 2013, 03:48 PM wrote:- in one of those classes go all the cars that are a little bit better than the current Pro class (by little bit I mean a difference like with Rookie to Arcade, Arcarde to Semi-Pro and Semi-Pro to Pro)

Stock cars have speeds from 29 to 43 mph split throuh 5 classes, so to define more classes on similar increments would suggest a request for at least 6 more classes upto speeds of 65+mph plus 1 class for ultimate speed cars 70+mph.

Kenny @ May 3 2013, 03:48 PM wrote:- in the other class go all the cars that are overpowered (with an insanely good handling/acceleration/top speed)

If your wanting a fair challenge from your AI opponents, depending on which tracks you race on, 'Top Speed' would be a critical factor no matter how good the handling.

Kenny @ May 3 2013, 03:48 PM wrote:- in the last class (if there is one) go all the other cars (that are undriveable or are simply not suitable to race against other cars or that the user wants to be seperated from all other cars)

Why would you want a class for cars that sound 'Not Fit For Purpose', you just would not want them at all.

Kenny @ May 3 2013, 03:48 PM wrote:However I think modifying all the cars on RVTZ to go along with new classes is kind of pointless but I guess thats not up to me to decide.

Yes it is. Along with everyone else, it's called democracy. Then the majority decision wins.... or we get a recount.

PS This is my 1000th post

This is the reply I got from Jigebren a while back...

...Sorry for not answering faster, but I didn't have time to delve into this issue yet. From experience I just know that we have to avoid to give any detail before it's done, otherwise any changes thereafter will be heavily criticized...
Since it seems to be a matter of time for you, I can just tell you how I see the things for now - but this has no legal value.

I think we'll simply support two extra values (5 and 6) for the rating, though I'm not sure 6 will be officially supported. Anyway, any value above the max value should be safely capped, so even value like 7 or more could be safely used.

The only issue I can see so far would be the compatibility with older re-volt versions. I have planned for quite some time to add an escape char like the ";)" that would do the opposite, make the line skipped by v1.2 only. That way we could have a parameter line dedicated for v1.2 only and another line for older version.
But this is less easier than for the ";)" support, so I can't tell for now how it'll work. Maybe any line that contains ";(" will be skipped by v1.2, but we still have to check if this can actually be done efficiently, or if there's a better solution...

So I guess since then the ";)" has been tested to work maybe? If so, that is great! And if everybody is happy with not having RVZT cars fixed for the ratings, then I guess that isn't a real problem. If that could work for everybody, we might have two new ratings a lot sooner that we thought.

Edit: and congrats MM on 1,000th post

Manmountain wrote:Stock cars have speeds from 29 to 43 mph split throuh 5 classes, so to define more classes on similar increments would suggest a request for at least 6 more classes upto speeds of 65+mph plus 1 class for ultimate speed cars 70+mph.

I didn't say we need to add 6 more classes, just that we use the already existing classes as a template on how we set the requirements for new classes.
If the numbers you posted are true than that would mean that the next class past Pro would contain cars with a top speed of 46 mph (+2.8 mph to be exact) which sounds like a reasonable limit to me.
And since there are totally overpowered cars it doesn't make much sense to me to go through the hassle to make classes for each of them, just put them all in one class. Anything above 60 mph is pretty much undriveable anyway (imo).

Manmountain wrote:If your wanting a fair challenge from your AI opponents, depending on which tracks you race on, 'Top Speed' would be a critical factor no matter how good the handling.

Well if you take the AI into account (which I didn't) then the whole thing becomes a new dimension. Because as we know the AI is able to drive very good with cars that we can't handle very well and vice versa.
So either we focus the car's class on how good the AI performs with the car or how the user goes along with it (I guess we are better off with the first one).

Manmountain wrote:Why would you want a class for cars that sound 'Not Fit For Purpose', you just would not want them at all.

I was thinking about cars like Trolley or some other ridiculous custom cars (I can't think of any now but I'm sure there are some).
Besides I was primarily meaning for that class to be one where it is up to the user to decide what goes in it, 'Not Fit For Purpose' was just a suggestion.

Kenny @ May 3 2013, 04:54 PM wrote:

Manmountain wrote:If your wanting a fair challenge from your AI opponents, depending on which tracks you race on, 'Top Speed' would be a critical factor no matter how good the handling.

Well if you take the AI into account (which I didn't) then the whole thing becomes a new dimension. Because as we know the AI is able to drive very good with cars that we can't handle very well and vice versa.
So either we focus the car's class on how good the AI performs with the car or how the user goes along with it (I guess we are better off with the first one).

Not to be too critical (I'm loving this thred right now).
Defining the class is to allow Re-Volt to select the appropriate AI to challenge your car selection, so not taking the AI into account makes this debate a bit useless.

Good AI performance is dependent on quality and flow of tracks used. And who would define and create this database of results ?

Kenny @ May 3 2013, 03:48 PM wrote:2 additional classes (perhaps 3 if we want to have general satisfaction)
- in one of those classes go all the cars that are a little bit better than the current Pro class (by little bit I mean a difference like with Rookie to Arcade, Arcarde to Semi-Pro and Semi-Pro to Pro)
- in the other class go all the cars that are overpowered (with an insanely good handling/acceleration/top speed)
- in the last class (if there is one) go all the other cars (that are undriveable or are simply not suitable to race against other cars or that the user wants to be seperated from all other cars)

This perfectly summarizes how I see it so far, except for the last class for which the CPUSelectable is better suited, just as Huki said above.

For the ";(" support I once mentioned to Mandolin, it may not be that needed in fact since you can do as Huki suggested, put the v1.1 line before the v1.2-only line.

Guys, you should leave the AI out of this, and handling in general also.

Kenny wrote: So either we focus the car's class on how good the AI performs with the car or how the user goes along with it (I guess we are better off with the first one).

1) I told in an earlier post that you can’t reliably measure the lap times of AI. I know, I tried. If a car bumps something and ends up in indecision between two track zones, it stays there for a _random_ time. Not even tuning them to Expert AI helps there, it comes from the game’s coding. (Even though two different cars, both tuned to Expert AI, run along the racing line with equal smoothness, basically the only difference coming from their top speed and acceleration, in pack racing they can still bump into each other and end up in indecision between two track zones for a random time.) So that’s not a reliable criterion.
2) Every user races differently with the same car. That’s not a reliable criterion.

Bottom line: AI lap times cannot be measured reliably. User handling cannot be measured reliably. Only solid performance like top speed in straight line and acceleration in straight line can be measured reliably i.e. so that it doesn’t depend on some random effect or on the player.

Oh, and I think Huki’s system of User 5 and User 6 ratings (May 3 2013, 05:06 PM) would work best, due to the reasons stated in his same post.

Guys, you should leave the AI out of this, and handling in general also.

Since there will be many people uploading their own cars with the new ratings, perhaps that actually would be best. The dev speed test in 1.2 should do an accurate job of measuring in-game speed and acceleration and would be the quickest way to determine what rating a car should have. We should come up with official guidelines for which speed/acceleration cominations belong to each class. Then, of course, the car creators are still free to deviate a little based on the car's handling if they want to. Also, maybe three classes would be better than two. I didn't think about the fact that the original classes are so close together. The higher classes wouldn't have to have quite so much space in between classes (the AI and the user tend to start losing some conrol over the car the faster they go anyway, so the win usually goes to the best driver as opposed to the fastest car), but two classes still might not be enough to have the right amount of speed/acceleration differentiation among cars in a given class. I think six is way too many, but maybe three.