"Reverse engineering" on PP's.
- Thread starter mtsmtts
- 47 comments
- 6,071 views
- 376
- Nürburgring
- lukasspeed
It still does in GT6 just like in GT5. At least for the MR/FR cars I tested so far.
But the rule always was PP increases with more weight on the driven wheels, which probably has to do with better traction. So for FR/MR(/RR?) more rear weight increases PP while shifting to front decreases PP. On 4WD cars though shifting weight around doesn't change PP. I didn't check 4WD in GT6 so far though, but I'd be very surprised if it changed.
[Edit: Just went into game to confirm. For both the McLaren 12C and the Aston One-77 shifting weight back increases PP while shifting to the front decreases. You might have to shift upwards of 100kg ballast to see a change in PP though.]
Last edited:
Yes, peak numbers are what PP is based on from the looks of it, and there are extra decimal places involved we can't see.
Torque certainly factors in, look at the 95 and 97 NSX (same hp and weight, different torque, different pp).
Something you forget is the power "area" of the motor. We all know power is plotted on an X-Y graph, where a line represents the power output (Y-axis) at a given RPM (X-axis). The space under the power line has a defined area, and is representative of the total output of a motor across it's RPM range.
By upgrading a car and limiting it, there is more "area" within the powerband. Basically, the peak power is flattened, and this new (lower) "peak power" is given over a wider rpm range. These cars are effectively a bit faster, with the same HP and PP numbers. This is especially true if you manage to decrease power by an amount that only "flattens" the peak within your gear change difference while maintaining the full torque curve. It's similar to the difference between 300 "turbo" horsepower, and 300 "NA" horsepower, but taken to an extreme.
Quick Math: Lets say we're looking at 5000-8000RPM (space between the gears). Car has 250hp at 5000, 300hp at 8000, which isn't too out of the ordinary. That's an area of 825,000 (the unit is useless, as we're honestly just comparing area.)
Upgrade the car, and limit back down to 300hp. Car now is flat 300HP from 5000-8000RPM. That's an area of 900,000, about an 8% gain. Importantly, it's 50 extra horsepower on the low end of your powerband when switching gears, with no loss of peak output. I'm sure you can see how this can make you faster.
I've played with weight distribution, it does not appear to have a PP effect. Taking 100kg's from -50 to +50 makes no difference, but going from 0kg to 100kg certainly does.
It appears that the "cg" of the car factors in, but it also appears to be a set factor. This is why (some) cars with the same drivetrain, weights, and power have different PP values.
It also is apparent that drivetrain effects PP, and it appears to be a multiplier. Lets say FR=1.0. AWD/4WD appears to be a bit higher, say 1.05, while FF appears to be lower, say .95.
What this means, is at the same weight and power, a FF car has a lower PP than a FR, which is lower than AWD.
(In game, at the same PP, a FF car is lighter and has more power, and quite often ends up being faster, especially <550PP in my experience)
But, there are still cars that defy my logic. Look at a Suzuki Swift, there's and 05 and 07 version. Look the same, spec'd the same, 4pp difference right off the bat. Civic SiR-ii '93 and '91 is the same thing.
Torque certainly factors in, look at the 95 and 97 NSX (same hp and weight, different torque, different pp).
Something you forget is the power "area" of the motor. We all know power is plotted on an X-Y graph, where a line represents the power output (Y-axis) at a given RPM (X-axis). The space under the power line has a defined area, and is representative of the total output of a motor across it's RPM range.
By upgrading a car and limiting it, there is more "area" within the powerband. Basically, the peak power is flattened, and this new (lower) "peak power" is given over a wider rpm range. These cars are effectively a bit faster, with the same HP and PP numbers. This is especially true if you manage to decrease power by an amount that only "flattens" the peak within your gear change difference while maintaining the full torque curve. It's similar to the difference between 300 "turbo" horsepower, and 300 "NA" horsepower, but taken to an extreme.
Quick Math: Lets say we're looking at 5000-8000RPM (space between the gears). Car has 250hp at 5000, 300hp at 8000, which isn't too out of the ordinary. That's an area of 825,000 (the unit is useless, as we're honestly just comparing area.)
Upgrade the car, and limit back down to 300hp. Car now is flat 300HP from 5000-8000RPM. That's an area of 900,000, about an 8% gain. Importantly, it's 50 extra horsepower on the low end of your powerband when switching gears, with no loss of peak output. I'm sure you can see how this can make you faster.
I've played with weight distribution, it does not appear to have a PP effect. Taking 100kg's from -50 to +50 makes no difference, but going from 0kg to 100kg certainly does.
It appears that the "cg" of the car factors in, but it also appears to be a set factor. This is why (some) cars with the same drivetrain, weights, and power have different PP values.
It also is apparent that drivetrain effects PP, and it appears to be a multiplier. Lets say FR=1.0. AWD/4WD appears to be a bit higher, say 1.05, while FF appears to be lower, say .95.
What this means, is at the same weight and power, a FF car has a lower PP than a FR, which is lower than AWD.
(In game, at the same PP, a FF car is lighter and has more power, and quite often ends up being faster, especially <550PP in my experience)
But, there are still cars that defy my logic. Look at a Suzuki Swift, there's and 05 and 07 version. Look the same, spec'd the same, 4pp difference right off the bat. Civic SiR-ii '93 and '91 is the same thing.
Last edited:
- 963
- Toronto, Canada
- times_fade
Not past a certain extent.. I'd say once you reach the absolute limit of the suspension physics/achievable corner speed, you're talking about an exact relation of a 'predetermined' throttle, brake and steering input the physics engine would require. Can you and me, the driver, ever reach this kind of precision and give the exact driving input or so-called 'driving style' the physics engine needs to reach this 'absolute' top performance? Unlikely, this is the reason why different drivers achieve different times with the exact same cars, that limit is actually found at traction loss, and of course this is what racing is all about.. what can the driver do to over-drive the cars setup to achieve a lap time quicker than another person. The question then arises, how accurate can PD or any other racing game developer model the behaviour that comes after 'certain' traction.. of course tire behaviour is also a huge point.
I would love to see how long the physics equation is and all the variations/revisions that have occurred, or how many times they possibly completely started from scratch between GT5: Prologue and present. I'm assuming once PD has more freedom with processing power on PS4, the equation could be allowed to have more variables and complexity. It would also be nice to see the other developers physics models of course, just to see how different the approaches are.. out of the question I know.. just sayin
- 3,070
- Norway
Not to mention the fact that this would have to be calculated for each track, and for each weather situation, and for tyre wear on/off, and for each tyre compound etc etc.
- 2,823
- Massachusetts
- GBO-Possum
Clearly the PP calculation ignores a lot of stuff, such as tires.
That said, I'm impressed at how good an indicator it seems to be.
A mate and I often race with wildly different styles of car. Such as an American muscle car vs a European or Japanese car with excellent handling. We always start with equal PP values and equal tires, and our goal is to have the closest of close finishes.
Rarely do we have to tweak the PP much to overcome the differences between the cars. If there is a difference, it can usually be explained by the track characteristics and the types of demand the track makes on the car.
This is the sort of difference we get up to:-
That said, I'm impressed at how good an indicator it seems to be.
A mate and I often race with wildly different styles of car. Such as an American muscle car vs a European or Japanese car with excellent handling. We always start with equal PP values and equal tires, and our goal is to have the closest of close finishes.
Rarely do we have to tweak the PP much to overcome the differences between the cars. If there is a difference, it can usually be explained by the track characteristics and the types of demand the track makes on the car.
This is the sort of difference we get up to:-
nasanu
(Banned)
- 1,429
- Australia
- nasanu
- nasanu
- 287
- Minnesota
- Raven_WET01
I would like to point out that peak HP is not the measurement to go by for PP comparisons.
If you bolt on every upgrade and limit the power to hit a specific PP limit, you can wind up with way less peak HP than if you were to purchase one or two upgrades and limit the power slightly. The entire power curve is taken into account when PP is calculated.
To show the difference, I tested a LotusElise 111R at 476pp. The run is on SSRX at full throttle from the time you gain control to the 400m mark.
No upgrades: 203hp-21.2 seconds
Fully upgraded, power at 55%:192hp- 20.9 seconds
Clearly, the car producing max power throughout the entire rev range will accelerate faster, but it takes a roughly 5% peak output penalty. If you want the highest top speed, you'll want to limit the power as little as necessary to ensure peak HP is higher in the upper RPM range. So PP calculations aside, (as a general rule) reduce your power to meet the PP level as much as possible within the usable power band when quick acceleration is a priority, and reduce it as little as possible when top speed (or acceleration at higher speeds) is a priority. This can account for a vast difference in lap times between cars with similar PP levels.
If you bolt on every upgrade and limit the power to hit a specific PP limit, you can wind up with way less peak HP than if you were to purchase one or two upgrades and limit the power slightly. The entire power curve is taken into account when PP is calculated.
To show the difference, I tested a LotusElise 111R at 476pp. The run is on SSRX at full throttle from the time you gain control to the 400m mark.
No upgrades: 203hp-21.2 seconds
Fully upgraded, power at 55%:192hp- 20.9 seconds
Clearly, the car producing max power throughout the entire rev range will accelerate faster, but it takes a roughly 5% peak output penalty. If you want the highest top speed, you'll want to limit the power as little as necessary to ensure peak HP is higher in the upper RPM range. So PP calculations aside, (as a general rule) reduce your power to meet the PP level as much as possible within the usable power band when quick acceleration is a priority, and reduce it as little as possible when top speed (or acceleration at higher speeds) is a priority. This can account for a vast difference in lap times between cars with similar PP levels.
Last edited:
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
I believe your results have led you to an erroneous conclusion. There's a big difference between a drag race over 400 metres and racing on a circuit. On a circuit, you generally operate within range of 1000-1500 rpm and sometimes less, depending on the car and transmission, so limiting power outside of that range is counterproductive. The general rule of PP limiting is that limiting the flatness of the power curve outside of the normal operating RPM range on a circuit, will rob you of HP and increase laptimes.
A power limited engine with engine limiting that produces this type of flat spot is ideal:
View attachment 93512
- 287
- Minnesota
- Raven_WET01
Right right right... I should've said within reason instead of "as much as possible" in my conclusion. Please note that there was no launch involved, so the vehicle was in normal operating range the whole time (6-9k RPM).
Edit: further testing
To flatten the power curve from 6-9k RPM, a sports computer with 93.6% power limit is used to achieve 476pp. It ran less than .1 faster than the 55% power limit test.
Last edited:
- 10,903
- Sweden
- eran0004
Engine limiter versus ballast to reduce the performance points of the Daihatsu Copen
I tested two different ways of reducing the PP of a car, to see what effect it had on the power to weight ratio (kg/BHP).
In the chart below, the blue line is when the reduction is done by adding ballast, the dotted line is when reduction is done by limiting the engine and the red line is the values of the dotted line, flattened to a linear (because BHP is a decimal number, and in the settings menu it's only displayed in integers, so when a change is made from, say 50,9 BHP to 50,2 BHP they both are displayed as 50).
The horizontal axis is the decrease in performance points and the vertical axis is the increase in power to weight ratio.
What it shows is that for this car (Daihatsu Copen), when the PP is reduced by limiting the engine (power decrease), the changes to the power-to-weight ratio is pretty much the same, for every 1% the kg/BHP increases, one PP is dropped.
When the PP is reduced by adding more ballast, the power to weight change is at first not as high as for the power decrease, but once it reaches a certain point, the required change is bigger than for the power decrease.
This chart shows the absolute Power to weight ratios:
At -12 PP, the weight increase method has got a power to weight ratio that's nearly 1 kg/BHP better than that of the power decrease method. After that point, the weight increase curve climbs and gets closer to the power decrease curve.
Why is this happening?
My guess is that the power decrease curve only have an effect on one variable: power to weight. The weight increase curve has an effect on two variables: power to weight and handling. Up to a certain point, the PP system thinks that the weight increase impacts on handling so much that the car needs a better power to weight ratio than the car that has only dropped in power for both cars to perform equal. After a certain point it figures that the handling can't get much worse if you add more weight, so if you want to drop the PP more, you need to add so much weight that it impacts on the power to weight ratio, and for those reasons the kg/BHP needs to increase so much that it catches up with the car that has reduced power.
What does this mean?
Well, the car with added weight would probably have better acceleration than the car with reduced power. The car with reduced power should have better braking and cornering, because it's lighter. What method is the best depends on the track and your driving style, although I suspect that for most people the car with the better power to weight ratio is the better one.
At -12 PP, the car with reduced power did 0-100 km/h in 16.7 seconds, while the car with added weight did 0-100 km/h in 16.1 seconds. In a race situation that's a difference of half a second only in the start.
Implications for other cars are a bit sketchy. The Copen is a lightweight car to begin with, and possibly the effect of added weight is greater for light cars than for heavy cars. Next car I'm going to test is the Audi A3, which is 500 kg heavier than the Copen.
Audi A3
Now this result seemed a bit odd at first. They are the opposite of the Copen: Adding weight gives a worse power to weight ratio than reducing power.
I would guess that this is because now we're dealing with aerodynamic drag to a much higher degree than what the Copens had to deal with. Because a high powered, heavy car would have an advantage over a low powered, lightweight car when it comes to high speed performance, it seems to me like the PP system in this range values power higher than weight. While it's true that the lighter car may have better acceleration, power is what's needed to reach those high speeds, and thus reducing power has a bigger impact on PP than adding weight. The car with reduced power has slightly better power to weight ratio than the heavier car, so it should expect to have better acceleration, but a lower top speed. Wether or not the two cars will be competitive with each other depends on the characteristics of the track.
Update 3: PP values of some FF cars @ 223 BHP
Decided to look into the PP variations of front wheel drive cars. I set all the cars to 223 BHP to see how well they line up with each other. The only variation here is the weight, I took the data on stock weight, +100 kg, +200 kg and finally stage 3 weight reduction.
Volvo C30, Alfa Romeo MiTo and Mini Cooper S are all nicely aligned along the same line. The Ford Focus is slightly overweight while the Hondas are much lighter than the others. For instance: At 410 PP, the Honda Civic is more than 200 kg lighter than the Ford Focus, yet they both have 223 BHP.
The Focus has 50% more torque than the Hondas though...
I tested two different ways of reducing the PP of a car, to see what effect it had on the power to weight ratio (kg/BHP).
In the chart below, the blue line is when the reduction is done by adding ballast, the dotted line is when reduction is done by limiting the engine and the red line is the values of the dotted line, flattened to a linear (because BHP is a decimal number, and in the settings menu it's only displayed in integers, so when a change is made from, say 50,9 BHP to 50,2 BHP they both are displayed as 50).
The horizontal axis is the decrease in performance points and the vertical axis is the increase in power to weight ratio.
What it shows is that for this car (Daihatsu Copen), when the PP is reduced by limiting the engine (power decrease), the changes to the power-to-weight ratio is pretty much the same, for every 1% the kg/BHP increases, one PP is dropped.
When the PP is reduced by adding more ballast, the power to weight change is at first not as high as for the power decrease, but once it reaches a certain point, the required change is bigger than for the power decrease.
This chart shows the absolute Power to weight ratios:
At -12 PP, the weight increase method has got a power to weight ratio that's nearly 1 kg/BHP better than that of the power decrease method. After that point, the weight increase curve climbs and gets closer to the power decrease curve.
Why is this happening?
My guess is that the power decrease curve only have an effect on one variable: power to weight. The weight increase curve has an effect on two variables: power to weight and handling. Up to a certain point, the PP system thinks that the weight increase impacts on handling so much that the car needs a better power to weight ratio than the car that has only dropped in power for both cars to perform equal. After a certain point it figures that the handling can't get much worse if you add more weight, so if you want to drop the PP more, you need to add so much weight that it impacts on the power to weight ratio, and for those reasons the kg/BHP needs to increase so much that it catches up with the car that has reduced power.
What does this mean?
Well, the car with added weight would probably have better acceleration than the car with reduced power. The car with reduced power should have better braking and cornering, because it's lighter. What method is the best depends on the track and your driving style, although I suspect that for most people the car with the better power to weight ratio is the better one.
At -12 PP, the car with reduced power did 0-100 km/h in 16.7 seconds, while the car with added weight did 0-100 km/h in 16.1 seconds. In a race situation that's a difference of half a second only in the start.
Implications for other cars are a bit sketchy. The Copen is a lightweight car to begin with, and possibly the effect of added weight is greater for light cars than for heavy cars. Next car I'm going to test is the Audi A3, which is 500 kg heavier than the Copen.
Audi A3
Now this result seemed a bit odd at first. They are the opposite of the Copen: Adding weight gives a worse power to weight ratio than reducing power.
I would guess that this is because now we're dealing with aerodynamic drag to a much higher degree than what the Copens had to deal with. Because a high powered, heavy car would have an advantage over a low powered, lightweight car when it comes to high speed performance, it seems to me like the PP system in this range values power higher than weight. While it's true that the lighter car may have better acceleration, power is what's needed to reach those high speeds, and thus reducing power has a bigger impact on PP than adding weight. The car with reduced power has slightly better power to weight ratio than the heavier car, so it should expect to have better acceleration, but a lower top speed. Wether or not the two cars will be competitive with each other depends on the characteristics of the track.
Update 3: PP values of some FF cars @ 223 BHP
Decided to look into the PP variations of front wheel drive cars. I set all the cars to 223 BHP to see how well they line up with each other. The only variation here is the weight, I took the data on stock weight, +100 kg, +200 kg and finally stage 3 weight reduction.
Volvo C30, Alfa Romeo MiTo and Mini Cooper S are all nicely aligned along the same line. The Ford Focus is slightly overweight while the Hondas are much lighter than the others. For instance: At 410 PP, the Honda Civic is more than 200 kg lighter than the Ford Focus, yet they both have 223 BHP.
The Focus has 50% more torque than the Hondas though...
Last edited:
- 3,955
- SoCal
- GTP_MooCow
I find it strange that you have power to weight ratio listed as "kg/BPH"... shouldnt it be BHP/kg ?
Any particular reason why you inverted them?
Is it more practical for your calculations?
It shouldnt change the conclusion much, but i had to twist my brain to read your chart... i am was thinking about doing something similar myself ... until i saw this thread...
Any particular reason why you inverted them?
Is it more practical for your calculations?
It shouldnt change the conclusion much, but i had to twist my brain to read your chart... i am was thinking about doing something similar myself ... until i saw this thread...
- 845
- California
- VELOCI_2NR
That's just the way the game displays the power-to-weight ratio, so I'm guessing that's why he did it that way.
- 2,931
- Austria
- GTP_SnowCrash
Some additional interesting facts:
I'm thinking it might be a sum of several summands, where...
...a certain power (area under the curves) equals, say, 100PP. A 5% increase yields +7.5PP.
...a certain weight equals 100PP. A 5% decrease yields +7.5%.
...additional constants for traction, centre of gravity, etc.
- The oil change boosts the car's power by 5%. That increase in power is done across the whole rpm-range. However, the effect in terms of PP is surprisingly even across all cars.
<290PP: +5PP
290-360PP: +6PP
360-450PP: +7PP
>450PP: +7-8PP
Above 450PP the PP-increase does not get any higher anymore. This might come from the fact that another part of the PP-formula cancels part of the PP-increase out. This makes some sense as the more power you have in a car, the less effective the same power increase in relative terms will be.
Note: Exception for this rule are hybrid cars as only part of their power-source gets the boost.
- If you increase a car's weight by 5%, you get the same 7-8PP drop in the PP-figure.
I'm thinking it might be a sum of several summands, where...
...a certain power (area under the curves) equals, say, 100PP. A 5% increase yields +7.5PP.
...a certain weight equals 100PP. A 5% decrease yields +7.5%.
...additional constants for traction, centre of gravity, etc.
- 627
- São Paulo
- Boxout7x
- Boxxout
Pretty much this. The PP should take in consideration other upgrades other than Power and Weight.
Let's imagine that you got both of exact same cars, Both tuned to the same PP/HP/Weight
Except Car 1 has Brakes upgraded, Transmission upgrades, Clutch upgraded, Suspension upgrades. While Car 2 runs with stock parts.
It's obvious that the Car 1 will have a huge advantage than Car 2
PD should re work this sort of things and make these upgrades count on PP as well, all of them, from clutch to tires.
- 28,471
- Windsor, Ontario, Canada
- Johnnypenso
The system is already wildly inaccurate, introducing more variables into the situation won't make it better it'll make it worse. I think it's pretty much assumed that in any given race, everyone has easy access to the same "standard" upgrades for their cars, including tires, drivetrain etc.
