Are all Performance Points created equal? *UPDATE: cars + HP:KG ratio*

[bys]

Disclaimer- this is by no means a comprehensive study of the PP system, there are some flaws in the experiment and probably also in the write up! So please feel free to point them out and any comments and ideas are welcome. Also this is still on going, I will update if there in any interest.

If you just want the results (given a PP limit, which car is best?), scroll down to the red & blue lists, otherwise read on!

Let's start with the what - What are performance points? As the name suggests, they are numerical representations of the overall performance of a given car. The exact algorithm to compute the PP's is unknown, but it is most likely based on HP, weight, weight-distribution, down force and so on. However the specific relationship between these variables and PP's is not of interest here.

Rather, I wanted to investigate the "reliability" of the PP system -namely: Would two cars with identical PP's hold identical performance? And if not, which cars are more "efficient". Here we define efficiency as those cars which perform better than expected, for given PP's - this will make more sense later on.

How? Well the process is pretty simple, I chose a track (custom generated Toscana circuit*) and recorded lap times of cars with varying PP's. Then by plotting lap-time vs. PP's - not only can different cars of the same PP be compared, but we can also see the overall relationship between lap-time and PP's. I used the DFGT wheel, manual transmission, no driving aids apart from ABS (=1) and sport soft tires.

Now to get the issues out of the way -

• There is one big way in which this experiment can be improved - by using "stock" cars and thereby removing the variable of tuning. However, I actually started this as my tuning routine and after having these results at hand, decided to do the PP study. This could throw off the results - e.g. if my suspension and transmission settings are completely off, then it doesn't matter which PP the car is, the time would not be an accurate reflection. However, since tuning is done by myself only - I would assume the consistency in tuning somewhat alleviates this problem.
• Each car was run at least 5 times, thereby ensuring that I was "used" to the car and a representative lap time achieved.

Findings


(click for bigger version)

How to interpret the Graph - The green 'x' represents the lap-time achieved by a car at it's given PP rating. The blue 'o' represents power-to-weight ratio, where the bigger the circle, the more HP per KG. And finally the green line is the line of best fit and represents "average" PP to lap-time relationship. From here on we assume the sample size is large enough and the trend line accurately predicts performance. Following this we define:

• Any car on the line fits our predictions (the given PP yields the predicted lap time)
• Any car above the line is "expensive" (the given PP yields a slower lap time than predicted)
• Any car below the line is "efficient" (the given PP yields a faster lap time than predicted)

It can be seen that most cars are within a "close" range to the prediction. However it is interesting to look at the outliers - compare the BMW M3 with the R34 Nur Spec. If we account for the 0.5s error, we can still see that they manage the same time, even though the M3 has almost 70 PP lower. And I have found this to be the case online, even at 550 PP races I was performing better than most cars on track. (could also be due to my mad skills)

Looking at the power-to-weight markers, we can see the following:

• Higher the HP:KG ratio, higher the PP
• Higher the HP:KG ratio, lower the lap time

This generally holds true - an exception being the Countach 25th Anniversary edition, where judging by it's HP:KG ratio, it should have about 100 more PP.

Eventually the list below will be more fully populated, and can be used as a guide to picking an "efficient" car.

Efficient Cars

• F430 Scuderia
• M5
• RS8
• NOMAD DIABLO
• RE RX7
• BMW M3
• Syliva S15
• Countach 25th
• ZR1 RM
• EVORA
• NSX 91 RM

Expensive Cars

• coutach lp400
• Gallardo 560
• R34 NUR
• s2000
• Viper STR10 ACR
• fairlady z II
• EVO 10
• Ford GT
• EVO VI TME
• Corvette ZR1
• GATHERS EK9
• WRX STI 10
• Veyron (this track is definitely too tight for the Veyron)

Closing remarks - actually this turned out as I expected, and I believe it should be in this way. If PP did actually perfectly represent the cars performance, then GT5 would whittle down to "whose PP is biggest". Of course, driver skills still comes into play but some excitement is lost. I feel that PP gives a rough indication of car performance, but depending on driver styles and perhaps some unknown factors produces some outliers. (it could well be that Polyphony randomizes PP within say +/-5 points of the computed value)

* The track has a long front-straight, followed by progressively tightening curves with increasing slope, easing back out to the level front straight. I created this track to test all aspects of a car for tuning purposes.

Open Questions

1. Stock vs. Tuned PP, how does tuning effect performance when compared to predicted performance?
2. What factor determines PP, i.e. HP, weight or HP:weight?
3. Is tuning a car to max-potential and limiting down to a given PP the same as 'up tuning' to that PP?

Regarding the open questions, I welcome someone who wants to lend a helping hand and do some experiments. It would be pretty straight forward and we could get some idea on the behaviour. Send me a PM! (also need a test driver for my tunes )

 

[cactusOnR]

This is great.
Really impressed with your work.

Me and a friend did a very quick test if our own. I had a fully maxed light weight car and my friend had a high tuned heavy car.
Same PP but huge weight and BHP levels.
We found that on on a tight track like deep forest my little car was better. On a bigger track with long straights the high BHP heavy car left me standing.


So we found that the PP level is great for similar car types. But not so great when you compare the likes of a KA and an America muscle car.

Not as technical a your tests but we found it very interesting.

For tourneys on my site we often enforce a BHP and weight limits instead of PP.

 

[bys]

cactusOnR - thanks man, appreciated. Yeah that's a good point about the tracks, I've edited the post to include some info on it.

Also, speaking of HP and weight, that gives me an idea to plot the times also vs HP and weight, and even HP:weight ratio - to see if those factors alone influence lap times more than PP (although I doubt it).

 

[cactusOnR]

No worries mate.
I look forward to seeing more in this thread.

Good luck with the testing.

 

[Johnnypenso]

My findings are completely different than yours, although I haven't graphed the results and my methods are different. I've taken up tuning the last couple of months and have given some thought to opening a garage soon, although I'm still up in the air about it. I was shocked at how close lap times were once figured out how to tune the various, weight, chassis, tire etc. combinations.

My original thought was a 2.5% rule, meaning that no car would make it into my garage unless it was within 2.5% of my fastest time on a given track, same PP, tires etc. At Deep Forest Reverse for example at 450ppSS my fastest times are around 1:21.9 and the 2.5% rule would mean that a car has to be at least 1:24 or faster to make the cut and so far I have at least 20 cars that can make that time, a great many of them in the 1:22's. I'd say at least 75% of the cars I tuned were within that range. My findings were the same at 500, 550 and 600pp. I would say that most of the top ranked cars are within 2.5% of each other and there are dozens in each category that are that close.

The cars that are not close, are those that would not be close in real life, and regardless of how much I tune them, they aren't as fast as they should be given their pp levels. I expect a GT-R, NSX Type R, M3 GTR and a couple of dozen other "premium" level cars to be close at 550pp and they are. What isn't close for example is an old school American Muscle car like a 69' Camaro or a Chevelle etc.

Seems to me that PD got the pp system right for the real life top echelon of cars but fell short on the others whose handling isn't quite up to snuff. They don't make up in power what they lack in handling.

For me the big equalizer is tuning. I don't even drive a stock tune, I begin tuning immediately upon getting a car, so I can't even tell you how much tuning improves a car from the stock version, but I can tell you that so long as it's a good car in real life, I can probably tune it so it's very close to the max performance of a given PP level.

It's also true that some cars don't do as well when they fight out of their weight class. The RGT and BTR for example, and perhaps the Yellowbird and a host of other cars, can do quite well if they are kept near their natural PP levels. But try to get them to compete with the GT-R's and Ferraris and they can become a beast to drive and tune.

My conclusion was that PD made the assumption that a lot of guys would really enjoy tuning and squeezing the last bit of performance out of their cars and the PP system is based on the maximum potential of a car, not the base tune. Once you throw a good tune on a car, it's surprising how many cars are competitive at a given PP level.

 

[bys]

hmm interesting, when you say your findings are "completely different", what do you mean exactly? From what I gather you are saying that 75% of cars at a given PP level, are within 2.5% of lap-time of each other? I don't mean to have a go, I'm just trying to understand for myself.

For example, if you look at my graph and look at cars at 580 PP, the predicted time is 120s, 2.5% of 120s is 3s and you can see if you travel in the "slower" direction - all cars come in under 123s - thus satisfying your criteria of being accepted.

Regarding - cars not being close, that's what I'm trying to show. I haven't tried american muscle cars, but I'm sure they would be well above the line even at a stock 500pp or tuned 650pp. But, for me at least, it's more interesting to see which cars come in under the line.

You raise a good point about "stock" PP and tuned PP - i always wanted to test this also. I have done so with a couple cars, as with my tuning routine I usually start with stock and tune as I see fit. However the problem here is that if you simply leave the car stock, you would get faster anyway as you get used to the car.


/edit, ah yeah spec 2.0 released!!! not getting anything done today haha

 

[nomis3613]

Great testing, and great analysis 👍

2 What factor determines PP, i.e. HP, weight or HP:weight?

It's closest to HP:weight, I haven't looked for any trends but I doubt it will be linear, because...

...handling and aero also affect PP. For handling, you can see the PP change as ballast location is changed.
A theory to consider is perhaps mechanical grip and aerodynamic grip are treated separately. The "efficient" cars might have loads of mechanical grip, but can keep their PP down because they have less aero grip. Just a thought to consider

3 Is tuning a car to max-potential and limiting down to a given PP the same as 'up tuning' to that PP?

IMHO you want to tune a car at its final PP. One example is when tuning for a higher power level, traction will be more of an issue, so you might compromise other aspects to improve traction. But then when you reduce the power, the extra traction isn't needed so the tune isn't properly optimised.

 

[bys]

Great testing, and great analysis 👍

cheers bud! its something I did for personal interest and thought I'd share.

It's closest to HP:weight, I haven't looked for any trends but I doubt it will be linear, because...

...handling and aero also affect PP. For handling, you can see the PP change as ballast location is changed.
A theory to consider is perhaps mechanical grip and aerodynamic grip are treated separately. The "efficient" cars might have loads of mechanical grip, but can keep their PP down because they have less aero grip. Just a thought to consider

Yeah I agree, actually someone else raised a point about aero grip and PP - I'm going to test that in the weekend. Ill also pull info on HP, weight and see if it gives anything when related to PP.

And you're last point makes sense, and now it's handy that we have 3 tuning sheets. Can tune one to max, another to say 550PP and one to 500PP.

 

[Vander]

My main issue with performance points is that the suspension tunin has no effect on the value. If you take two completely identical cars and modify the suspension, lsd, and transmission settings, you can improve lap times greatly without affecting pp value. This, to me, is the biggest problem with pp it should be based on every aspect of modification or it is not a performance rating. Tires, for example, should effect pp probably more than anything, and yet they have no effect.

 

[savage388]

The pp system has major flaws. Even with only street cars there are "cheater" cars that are much faster tuned or not, for a given pp. Ive found that the 458, the evora, the premium nsx type r, and the racy viper ACR are very fast compared to cars of similar pp.

 

[danjee]

I've always what math equation PD uses to calculate PP...

 

[Motor City Hami]

Tires, for example, should effect pp probably more than anything, and yet they have no effect.

It's funny too that the in-game manual description for PP calls out weight, hp, aero and TIRES. Yet tires didn't make it into the PP calculation.

 

[danjee]

It's funny too that the in-game manual description for PP calls out weight, hp, aero and TIRES. Yet tires didn't make it into the PP calculation.

Does it take torque into account? Because would be a massive fail on PD's part if they left that out...

 

[XDesperado67]

Does it take torque into account? Because would be a massive fail on PD's part if they left that out...

Yes it does.

 

[Motor City Hami]

From the manual

6-15 Performance Points

Each car has a performance point (PP) rating that indiates its overall capabilities. In online races, it is possible to restrict entry based on PP ratings. In PP restrricted events, the closer they are to the PP limit, the more likely it is that cars with big differences in weight and engine power will be able to compete on the same playing field.

A car's PP rating will change to reflect changes in factors such as max power, weight, tires and aerodynamics. A car's PP rating can be checked in the garage or on the settings screen.

 

[Motor City Hami]

Yes it does.

Not sure I see torque in the manual?

 

[XDesperado67]

Not sure I see torque in the manual?

Obviously things changed between the time the manual was written and now. I believe it was update 1.6 or 1.7 that removed tires from the equation. As for torque that's easy to see. Take a car and modify it over a given PP. Then use various methods to reduce it to that PP.
As an example I have a car I'm tuning to 500PP.
My weight stays constant for all these and I have no aerodynamics and I have all engine upgrades except a turbo.
At 100% power limiter the car has 431hp, 420ft-lb torque and 511PP.
1st) applying 91.6% power limiter results in 395HP, 420ft-lb torque and 500PP.
2nd) removing the ECU and applying 98.1% power limiter results in 398HP, 408ft-lb torque and 500PP.
3rd) replacing the ECU I switch from the titanium racing exhaust to the sports exhaust and 96.9% power limiter resulting in 397HP, 412ft-lb torque and 500PP.
4th) same as 3 except I switch to standard exhaust manifold and 100% power limiter resulting in 398HP, 400ft-lb torque and 500PP.

As you can see torque definitely factors into the equation.
Another way to check this is to take two different cars and give them the same weight with ballast. At any given PP the car with greater torque will have less HP.

 

[bys]

The pp system has major flaws. Even with only street cars there are "cheater" cars that are much faster tuned or not, for a given pp. Ive found that the 458, the evora, the premium nsx type r, and the racy viper ACR are very fast compared to cars of similar pp.

Yeah there are a few cars that seem to appear quite often online, I guess these are "effecient" cars. I've also heard the Elise 111R is fast for it's PP - I'll probably update this weekend with all these cars just to quantify it a bit.

 

[VTiRoj]

PP used to take tyres into account, but later on PD decided it didn't want to anymore and made it so tyres don't change the PP.

 

[C-ZETA]

PP used to take tyres into account, but later on PD decided it didn't want to anymore and made it so tyres don't change the PP.

That only lasted for one patch. Too bad that was the case, would have made for a much better indicator to general closeness to the opponent.

 

[Motor City Hami]

Have we reached a consensus on wihich variable is most importat to build fast PP level driven cars? Which variables are most efficient or does it vary for each car?

Weight
HP/Torque
Aero

I have seen quite a few discussions where people think high aero in trade for lower HP is bad and not the way to be fast online. I havent' seen many discussions about weight vs. HP.

 

[MTLclllutch]

I have seen quite a few discussions where people think high aero in trade for lower HP is bad and not the way to be fast online. I havent' seen many discussions about weight vs. HP.

I think most people that go with max aero and say there is no downside to it use maxxed out cars with racing soft tires. I just dont see how max aero can help on tracks like tsukuba against having an extra gob of hp and torque.

 

[CSLACR]

Aero can depend on the track, but usually it works close to the same, whether high aero lower power or high power lower aero, the catch is, full aero is always easier to drive.

I just dont see how max aero can help on tracks like tsukuba against having an extra gob of hp and torque.

Since when do "gobs of power" help significantly on tight circuits?

The real trick with Aero is rear is worth more than front. Ideally in any combo with aero involved you find a good balance between hp and aero. For PP limited races, that is.

 

[Johnnypenso]

I don't think there is one general combination of hp/torque/downforce that applies to all cars, all tracks and all pp levels. But generally I find the gains in traction from aero begin to offset the loss of hp in pp limited situations in the 550range. At 600 pp on most tracks, for me max aero is the way to go for overall speed and consistency. There may be some situation where no aero can still work but not a lot, for me anyway. At lower PP levels, no aero is definitely the way to go for most cars.

 

[ampire]

If you want, I can run stock premium cars along trial mountain (best test circuit in my opinion because it has elevation, curves, straights, and most importantly short) all with racing soft tires and share lap times, maybe run a regression between performance points and lap time or performance points and power to weight ratio, etc. Other variables that I think could be used: wheelbase, peak horsepower to peak torque ratio (a measure of power-band efficiency), torque to weight ratio, cost in credits. I am guessing a multivariate model would be best.

I think racing softs is an appropriate tire because they improve consistency and the majority of people use it.

Some cars I think would be appropriate:
Subaru WRX (older premium model)
Mitsubishi Evo (older premium model)
BMW M3
BMW M5
GTR 07 and 09
Lotus Elise
Chevy Corvette ZR1
Chevy Corvette ZO6
Dodge Viper GTS
Dodge Viper ACR
Ford GT
Chevy Camaro
Lexus IS-F
Civic Type R
Skyline R34 Nur
Honda S2000
Honda RSX Type R
Mazda RX7 Spirit R

Any others you feel appropriate. I figure I'll list the best of 3 laps with each car. No oil change. I am undecided on the tire, maybe racing hards would be better than racing softs.

The other option is I do laps of Tsukuba Circuit. I am decent at these two tracks, these are the only two I am 100% consistent on.

 

[ampire]

I think most people that go with max aero and say there is no downside to it use maxxed out cars with racing soft tires. I just dont see how max aero can help on tracks like tsukuba against having an extra gob of hp and torque.

Tskuba you definitely don't want aero or extra gobs of HP. Wheel spin costs you precious time and none of the corners are highspeed enough to really help with aero except the last curve.

 

[esoxhntr]

actually, I find gobs of power to be nice on tsukuba. You're essentially drag racing between tight corners. i think it's why i do well in the f150 lightning on it in PP rooms.

I agree with johnnypenso, the aero vs hp debate gets cloudy at about 550pp on some tracks tracks. IMO, on HP tracks like Nurburgring the level is even higher - others disagree but I think that goes along with a point CSLACR made (easier to drive with more aero - never confuse easier to drive with faster.)

 

[CSLACR]

ampire - Using RS tires will increase the PP in which aero becomes more effective then power.
Because the aero isn't needed or helpful at lower speeds with super sticky tires.

 

[Motor City Hami]

Do tires increase the performance points?

 

[Rotary Junkie]

Do tires increase the performance points?

Not since the update immediately following the one that brought us the PP system. Basically, they did, then not long later they didn't.