3,524 lbs is NOT heavy....it's best-in-class.
#141
Burning Brakes
Merit? Don't like being corrected friend?
Sorry you said 996TT to the current 991TT, you just missed an entire generation of the 997TT(best one at that), only roughly 7 years lol. Judging by your other post you should let him explain it again. Yes, technically the TT is a year or so behind the new release of a gen just like our beloved Z cars but damn I was trying to keep it simple for you. Sorry for your mistake.
Sorry you said 996TT to the current 991TT, you just missed an entire generation of the 997TT(best one at that), only roughly 7 years lol. Judging by your other post you should let him explain it again. Yes, technically the TT is a year or so behind the new release of a gen just like our beloved Z cars but damn I was trying to keep it simple for you. Sorry for your mistake.
I definitely know my cars and facts as well as Porsches.
The OP had no problem with what I said either guy. He understood, whats your beef?
I guess no one on the forums here can know or have anything other than a Corvette, or they must have read it in Road and Track? I happen to have and have had quite an array of cars and driving experiences.
Beyond amusing.
#142
Drifting
The mistake is yours funny man. I said "A" previous generation turbo, not "THE" previous turbo. You obviously can't read.
I definitely know my cars and facts as well as Porsches.
The OP had no problem with what I said either guy. He understood, whats your beef?
I guess no one on the forums here can know or have anything other than a Corvette, or they must have read it in Road and Track? I happen to have and have had quite an array of cars and driving experiences.
Beyond amusing.
I definitely know my cars and facts as well as Porsches.
The OP had no problem with what I said either guy. He understood, whats your beef?
I guess no one on the forums here can know or have anything other than a Corvette, or they must have read it in Road and Track? I happen to have and have had quite an array of cars and driving experiences.
Beyond amusing.
#143
Burning Brakes
Very nice sports car and it should be. Base price $280,000, as equipped tested by Forbes $325,000.
If absolute lowest weight is the only point of comparison then the Z06 loses. If having a prancing stallions emblem is a requirement your choice has been made. Every product is a large bundle of benefits and drawbacks and how important each is to an individual determines what is the best overall product for that individual.
GM is not a boutique auto company and it isn't going to be producing exotics. If that is what you are looking for there are a number of companies that can provide what you want.
If absolute lowest weight is the only point of comparison then the Z06 loses. If having a prancing stallions emblem is a requirement your choice has been made. Every product is a large bundle of benefits and drawbacks and how important each is to an individual determines what is the best overall product for that individual.
GM is not a boutique auto company and it isn't going to be producing exotics. If that is what you are looking for there are a number of companies that can provide what you want.
Except for cost, the McLaren wins hands down. Its the real deal.
#144
Burning Brakes
#145
NO.
AFM, DI, AND VVT TOGETHER add 35.2 pounds. How did you decide that AFM is 30 plus pounds alone? So, the hardware for VVT and high output (read heavier) components like fuel injection pump and fuel pump (for Direct Injection) only add 5 pounds???
And how much do YOU think the battery weighs in a Z06???? The battery AMG uses SAVES 45 pounds???? That's literally IMPOSSSIBLE (unless you find a conventional car battery that weigh's 90 pounds).
Jimmy
AFM, DI, AND VVT TOGETHER add 35.2 pounds. How did you decide that AFM is 30 plus pounds alone? So, the hardware for VVT and high output (read heavier) components like fuel injection pump and fuel pump (for Direct Injection) only add 5 pounds???
And how much do YOU think the battery weighs in a Z06???? The battery AMG uses SAVES 45 pounds???? That's literally IMPOSSSIBLE (unless you find a conventional car battery that weigh's 90 pounds).
Jimmy
Seems like you have to see it to believe it here is the S63 weight reduction: http://www.caranddriver.com/news/201...easures-page-2
#146
Burning Brakes
Member Since: Mar 2010
Location: Manassas VA
Posts: 839
Likes: 0
Received 0 Likes
on
0 Posts
#148
indicating curb weight vs torque without considering RPM is a huge mistake, some of those cars hit 9000rpms hence make shitloads of power there regardless of the low torque, HP vs curb weight is the correct way of doing it and changes everything.
#149
Burning Brakes
I've been trying to say that you guys are thinking emotionally...you cannot dispute math. Below you'll find a Pearson coefficient correlation test which scientifically proves, without doubt, that torque-weight correlates 49% better than horsepower-weight in setting Nurburgring lap times. Other metrics don't even compare. Please do not dispute cold, hard facts.
Let me state again: the weight to torque ratio is, by a large margin, the single most relevant performance metric because it's the one that correlates closest to real lap times. When it comes purely to racing one car against another around a track to set the best time, objectively & removing all human emotion out of the argument, torque versus weight matters most. Above all else...no other metric even comes close.
But then, car enthusiasts generally don't run mathematical correlation & regression analysis to prove their point so I can understand how all this forum banter can lead you to believe the wrong thing.
Correlation Summary
Weight to torque = 0.55 (55% moderately high correlation)
Weight to horsepower = 0.37 (37% weak correlation)
Weight = 0.19 (very weak correlation on weight alone)
Horsepower alone = -0.33 (high HP alone lowers ring time)
Torque alone = -0.59 (high torque alone lowers ring time)
PART I - Baseline measurements
Tested vehicle (Ring time in seconds, LB/HP, LB/Torque)
2012 Corvette ZR1 (439/5.22/5.52)
2009 Viper ACR (442/5.59/5.99)
2012 Z06 (443/6.29/6.76)
2011 Aventador (445/5.49/7.49)
2008 GTR (446/ 7.86/ 8.79)
2009 458 Italia (448/5.83/8.23)
2010 MP4-12C (448/5.33/7.12)
2012 Ferrari F12 (453/4.54/6.61)
2009 911 Carrera S (454/9.12/11.32)
2013 Camaro Z28 (457/7.72/8.21)
2009 911 Turbo (457/ 7.10/ 7.19)
2009 Gallardo LP 560-4 (458/5.88/8.31)
20109 Audi R8 V10 (459/6.55/8.80)
PART II - Weight to torque correlation = 0.55
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 100.34
Mean = 7.718
∑(Y - My)2 = SSy = 26.8
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 67.038
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 67.038 / √((550.923)(26.8)) = 0.5517
PART III - Weight to horsepower correlation = 0.37
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 82.52
Mean = 6.348
∑(Y - My)2 = SSy = 19.838
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 38.798
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 38.798 / √((550.923)(19.838)) = 0.3711
PART IV - Weight-torque versus weight-HP
Comparing the difference in variance,
(0.55 - 0.37)/0.37 = 49%
Let me state again: the weight to torque ratio is, by a large margin, the single most relevant performance metric because it's the one that correlates closest to real lap times. When it comes purely to racing one car against another around a track to set the best time, objectively & removing all human emotion out of the argument, torque versus weight matters most. Above all else...no other metric even comes close.
But then, car enthusiasts generally don't run mathematical correlation & regression analysis to prove their point so I can understand how all this forum banter can lead you to believe the wrong thing.
Correlation Summary
Weight to torque = 0.55 (55% moderately high correlation)
Weight to horsepower = 0.37 (37% weak correlation)
Weight = 0.19 (very weak correlation on weight alone)
Horsepower alone = -0.33 (high HP alone lowers ring time)
Torque alone = -0.59 (high torque alone lowers ring time)
PART I - Baseline measurements
Tested vehicle (Ring time in seconds, LB/HP, LB/Torque)
2012 Corvette ZR1 (439/5.22/5.52)
2009 Viper ACR (442/5.59/5.99)
2012 Z06 (443/6.29/6.76)
2011 Aventador (445/5.49/7.49)
2008 GTR (446/ 7.86/ 8.79)
2009 458 Italia (448/5.83/8.23)
2010 MP4-12C (448/5.33/7.12)
2012 Ferrari F12 (453/4.54/6.61)
2009 911 Carrera S (454/9.12/11.32)
2013 Camaro Z28 (457/7.72/8.21)
2009 911 Turbo (457/ 7.10/ 7.19)
2009 Gallardo LP 560-4 (458/5.88/8.31)
20109 Audi R8 V10 (459/6.55/8.80)
PART II - Weight to torque correlation = 0.55
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 100.34
Mean = 7.718
∑(Y - My)2 = SSy = 26.8
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 67.038
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 67.038 / √((550.923)(26.8)) = 0.5517
PART III - Weight to horsepower correlation = 0.37
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 82.52
Mean = 6.348
∑(Y - My)2 = SSy = 19.838
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 38.798
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 38.798 / √((550.923)(19.838)) = 0.3711
PART IV - Weight-torque versus weight-HP
Comparing the difference in variance,
(0.55 - 0.37)/0.37 = 49%
A NISMO GTR does the Ring in 429 seconds, 10 seconds faster than the next fastest on your list. It has 595 HP and only 481 ft/lb of TQ, weighs ~3800 lbs. Pretty much invalidates your post.
#150
I just saw Ching Ho's other explanation and it makes even less sense now. Dear Ching ho, using nurburgring times to validate your point is a VERY bad example on how to approach this, you have to take into consideration that ASSUMING you are on the throttle for 50% of the time, you are on the brakes the other 50% of the time on a road course (those %'s are not accurate by any chance, just trying to make a point for the sake of the argument), then your calculations are falsely taking into consideration how good the brakes are, and the conditions of the track, the weather, driver, tires, transmission shift speeds, gear ratios, aerodynamics and so on.
The second problem I see here if your using advertised HP and curb weight figures for those vehicles, both numbers which are usually underrated or measured with different vehicle options, which could change the weight a good bit.
Third, those vehicles aren't being driven by the same driver (guess I mentioned this already), this itself is big factory in the equation.
Finally, the sample size is far from enough to even conclude such a thing, the three things that make a car fast is the amount of torque it can generate at the rate the engine is spinning at (rpm) vs the weight it has to act upon which inevitably equates to HP/lbs regardless of how we try to look at it.
The second problem I see here if your using advertised HP and curb weight figures for those vehicles, both numbers which are usually underrated or measured with different vehicle options, which could change the weight a good bit.
Third, those vehicles aren't being driven by the same driver (guess I mentioned this already), this itself is big factory in the equation.
Finally, the sample size is far from enough to even conclude such a thing, the three things that make a car fast is the amount of torque it can generate at the rate the engine is spinning at (rpm) vs the weight it has to act upon which inevitably equates to HP/lbs regardless of how we try to look at it.
Last edited by zeshawn; 08-17-2014 at 09:04 PM.
#151
The "Fat Vette" will curbstomp the ZR1 let alone the out going fabled C6 ZO6 ten ways from sunday.
It's already beat the pants off of the ZR1 at the lutz ring by well over a second a lap, on it's very first time touching the track without further development.
Lap Times Matter. And the C7 ZO6, has a full interior (unlike the C6 which frankly might as well have never had one to start with, it's fit and finish was so atrocious that even -I- who cares two **** about interior noticed).
Not only that, it has an Electronic limited slip diff, which adds weight. It has cylinder deactivation. again Weight.
I bet everyone whining about the weight of the C7 ZO6 could easily shed 50lbs themselves and not be hurting for it....
let alone that there are C7s on this forum with shots taken of race scales that weigh well less than 3400. The GM figures are for fully loaded, (in fuel and in trim options)
God you people are acting like it's a 1976 Stingray with a 200 horse 454 tipping the scales at 3800 lbs
It's already beat the pants off of the ZR1 at the lutz ring by well over a second a lap, on it's very first time touching the track without further development.
Lap Times Matter. And the C7 ZO6, has a full interior (unlike the C6 which frankly might as well have never had one to start with, it's fit and finish was so atrocious that even -I- who cares two **** about interior noticed).
Not only that, it has an Electronic limited slip diff, which adds weight. It has cylinder deactivation. again Weight.
I bet everyone whining about the weight of the C7 ZO6 could easily shed 50lbs themselves and not be hurting for it....
let alone that there are C7s on this forum with shots taken of race scales that weigh well less than 3400. The GM figures are for fully loaded, (in fuel and in trim options)
God you people are acting like it's a 1976 Stingray with a 200 horse 454 tipping the scales at 3800 lbs
An economy chassis car with a somewhat powerful engine, aka pony car is supposed to weigh a certain amount. Once a beloved sports car creeps up there, people will be disappointed. Whether the expectations were realistic or not, or pure fantasy, it's really nothing to be upset about. We are a mix of disciplines, we like cars but approach from different viewpoints. Some are track time folks, some 0-60, some style, some GT, some a balanced mix of everything.
Loosing 50 lbs off my midsection would be impossible; 50 lbs would have to come from other parts of my body as well and I'd be dead.
Last edited by User24; 08-17-2014 at 09:15 PM.
#152
Le Mans Master
I don't find the weight to be a bother. If it crept past 3800 pounds, I'd start to feel differently. I have no problem trusting that, for a street car, all the materials and parts used on the C7 Z were necessary and not fluff. Maybe in time they will offer delete options for people who are really serious to shed weight and get rid of everything short of a steering wheel and a seat.
#154
Burning Brakes
I don't find the weight to be a bother. If it crept past 3800 pounds, I'd start to feel differently. I have no problem trusting that, for a street car, all the materials and parts used on the C7 Z were necessary and not fluff. Maybe in time they will offer delete options for people who are really serious to shed weight and get rid of everything short of a steering wheel and a seat.
#155
Drifting
Hard to tell from the maker's specs. http://www.braillebattery.com/index....batteries/i65s says "19.5lb/5.22kg" but 19.5lb = 8.9kg and 5.22kg = 11.5lb so let's say it weighs 19.5lb and we know we can buy one for $2,400. Here's the most expensive battery for the C7 at Batteries Plus http://www.batteriesplus.com/product...2L-615CCA.aspx which is $170 for 45.5lb. Wow, $2,230 extra saves an entire 26lb. But if that forces all sports car snobs to admit that the C7 Z06 is now a sports car because it's 26lb lighter, it's money very well spent.
#156
Race Director
Whatever is the weight of the AFM 20,25 or 30 lbs, i don't see any true sports cars like Ferrari 458 or Porsche GT3 with that stuff you want those cars to be as light as possible.
Seems like you have to see it to believe it here is the S63 weight reduction: http://www.caranddriver.com/news/201...easures-page-2
Seems like you have to see it to believe it here is the S63 weight reduction: http://www.caranddriver.com/news/201...easures-page-2
You better get used to mileage stretchers on higher volume, high power sports cars as mileage standards tighten. All 911 Carreras have stop/start now, which adds weight through heavier, higher capacity batteries, more stout starters, and alternators. For very limited cars (like the GT3/Speciale) there's NO need to do these type things. In fairness, while you keep comparing those two cars (GT3/Speciale) to the Z06, you should also keep in mind that there will probably be 3 times as many Z06's built as GT3/Speciales combined. In theory, you're comparing "track" cars, in reality the Z06, because of it's production level, HAS to, at least give the impression of being somewhat "green".
Jimmy
PS. While I'm not sure why the M-B S63 AMG got into this conversation (the car was GROSSLY overweight, now the new one is just really overweight), I looked up the battery required for the 2013 model. It is a HUGE battery (Motor Trend called it "Freakishly huge") and it weighs nearly 60 pounds! That said, I seriously doubt that the new lithium battery weighs 15 pounds so I think Car and Driver got this wrong.
Last edited by jimmyb; 08-18-2014 at 12:03 AM.
#157
You should definitely do some research..
Last edited by Lavender; 08-17-2014 at 11:21 PM.
#158
#159
Safety Car
I've been trying to say that you guys are thinking emotionally...you cannot dispute math. Below you'll find a Pearson coefficient correlation test which scientifically proves, without doubt, that torque-weight correlates 49% better than horsepower-weight in setting Nurburgring lap times. Other metrics don't even compare. Please do not dispute cold, hard facts.
Let me state again: the weight to torque ratio is, by a large margin, the single most relevant performance metric because it's the one that correlates closest to real lap times. When it comes purely to racing one car against another around a track to set the best time, objectively & removing all human emotion out of the argument, torque versus weight matters most. Above all else...no other metric even comes close.
But then, car enthusiasts generally don't run mathematical correlation & regression analysis to prove their point so I can understand how all this forum banter can lead you to believe the wrong thing.
Correlation Summary
Weight to torque = 0.55 (55% moderately high correlation)
Weight to horsepower = 0.37 (37% weak correlation)
Weight = 0.19 (very weak correlation on weight alone)
Horsepower alone = -0.33 (high HP alone lowers ring time)
Torque alone = -0.59 (high torque alone lowers ring time)
PART I - Baseline measurements
Tested vehicle (Ring time in seconds, LB/HP, LB/Torque)
2012 Corvette ZR1 (439/5.22/5.52)
2009 Viper ACR (442/5.59/5.99)
2012 Z06 (443/6.29/6.76)
2011 Aventador (445/5.49/7.49)
2008 GTR (446/ 7.86/ 8.79)
2009 458 Italia (448/5.83/8.23)
2010 MP4-12C (448/5.33/7.12)
2012 Ferrari F12 (453/4.54/6.61)
2009 911 Carrera S (454/9.12/11.32)
2013 Camaro Z28 (457/7.72/8.21)
2009 911 Turbo (457/ 7.10/ 7.19)
2009 Gallardo LP 560-4 (458/5.88/8.31)
20109 Audi R8 V10 (459/6.55/8.80)
PART II - Weight to torque correlation = 0.55
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 100.34
Mean = 7.718
∑(Y - My)2 = SSy = 26.8
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 67.038
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 67.038 / √((550.923)(26.8)) = 0.5517
PART III - Weight to horsepower correlation = 0.37
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 82.52
Mean = 6.348
∑(Y - My)2 = SSy = 19.838
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 38.798
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 38.798 / √((550.923)(19.838)) = 0.3711
PART IV - Weight-torque versus weight-HP
Comparing the difference in variance,
(0.55 - 0.37)/0.37 = 49%
Let me state again: the weight to torque ratio is, by a large margin, the single most relevant performance metric because it's the one that correlates closest to real lap times. When it comes purely to racing one car against another around a track to set the best time, objectively & removing all human emotion out of the argument, torque versus weight matters most. Above all else...no other metric even comes close.
But then, car enthusiasts generally don't run mathematical correlation & regression analysis to prove their point so I can understand how all this forum banter can lead you to believe the wrong thing.
Correlation Summary
Weight to torque = 0.55 (55% moderately high correlation)
Weight to horsepower = 0.37 (37% weak correlation)
Weight = 0.19 (very weak correlation on weight alone)
Horsepower alone = -0.33 (high HP alone lowers ring time)
Torque alone = -0.59 (high torque alone lowers ring time)
PART I - Baseline measurements
Tested vehicle (Ring time in seconds, LB/HP, LB/Torque)
2012 Corvette ZR1 (439/5.22/5.52)
2009 Viper ACR (442/5.59/5.99)
2012 Z06 (443/6.29/6.76)
2011 Aventador (445/5.49/7.49)
2008 GTR (446/ 7.86/ 8.79)
2009 458 Italia (448/5.83/8.23)
2010 MP4-12C (448/5.33/7.12)
2012 Ferrari F12 (453/4.54/6.61)
2009 911 Carrera S (454/9.12/11.32)
2013 Camaro Z28 (457/7.72/8.21)
2009 911 Turbo (457/ 7.10/ 7.19)
2009 Gallardo LP 560-4 (458/5.88/8.31)
20109 Audi R8 V10 (459/6.55/8.80)
PART II - Weight to torque correlation = 0.55
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 100.34
Mean = 7.718
∑(Y - My)2 = SSy = 26.8
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 67.038
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 67.038 / √((550.923)(26.8)) = 0.5517
PART III - Weight to horsepower correlation = 0.37
X Values
∑ = 5849
Mean = 449.923
∑(X - Mx)2 = SSx = 550.923
Y Values
∑ = 82.52
Mean = 6.348
∑(Y - My)2 = SSy = 19.838
X and Y Combined
N = 13
∑(X - Mx)(Y - My) = 38.798
R Calculation
r = ∑((X - My)(Y - Mx)) / √((SSx)(SSy))
r = 38.798 / √((550.923)(19.838)) = 0.3711
PART IV - Weight-torque versus weight-HP
Comparing the difference in variance,
(0.55 - 0.37)/0.37 = 49%
Hypothesis: Correlation analysis is insufficient to capture the relationship between torque and car performance.
Theoretical approach:
1. Your model cannot distinguish between two engines that are equal in every way except one has more torque. It cannot distinguish between the Hell Cal engine and the LT4. This is in fact, a most important point. The manufacturers are hell bent on using forced induction and DCT transmissions nowadays. DCTs are generally torque capacity constrained. That's why we see turbo cars like the GT-R and 911 increasing power from year to year while keeping torque relatively unchanged. Your model cannot account for increases in torque at the top of the rev range unless peak torque increases as well. In other words, your model cannot even capture the torque curve, but rather one single point in a continuous distribution.
2. Your model fails to take into account power and other publicly available information about the cars in question. Why limit yourself to kerb weight and peak torque when you can improve the predictive ability of your model significantly using information readily available just about anywhere? The coefficient of correlation may be the best predictor for car performance if we only observed weight, torque, power, but we were only able to use a ratio of two of them. In reality, these restrictions don't exist, and therefore your correlation analysis can be easily beaten by the smallest modifications.
3. Correlation does not imply causation. To see why this is true consider that I could use correlation to predict the chance of rain. I observe the number of umbrellas out of my window and run a correlation analysis by observing whether indeed it is raining outside. Even if this correlation is .9 and I can easily use this model to predict the chance of rain, the number of umbrellas does not CAUSE rainfall. Therefore, if you and me both opened our umbrellas we would not make it more likely for rainfall to occur. In your model, it is unclear whether the weight to peak torque ratio predicts performance due to torque characteristics or other features of high torque vehicles.
Empirical proof:
1. Regress Nurburgring time on your ratio variable while controlling for dry weight, and a polynomial fit to the power curve. Given that no other confounding factors exist you should get a coefficient that is statistically insignificant for the ratio. Anything else is a spurious relationship.
2. Tire patch, power, and many other 3rd factors are correlated to both Nurburgring times and torque/weight. A simple predictor like weight/torque ratio is therefore a biased estimator of Nurburgring times.
QED, next.
#160
Instructor
Thread Starter
Let me give it a shot.
Hypothesis: Correlation analysis is insufficient to capture the relationship between torque and car performance.
Theoretical approach:
1. Your model cannot distinguish between two engines that are equal in every way except one has more torque. It cannot distinguish between the Hell Cal engine and the LT4. This is in fact, a most important point. The manufacturers are hell bent on using forced induction and DCT transmissions nowadays. DCTs are generally torque capacity constrained. That's why we see turbo cars like the GT-R and 911 increasing power from year to year while keeping torque relatively unchanged. Your model cannot account for increases in torque at the top of the rev range unless peak torque increases as well. In other words, your model cannot even capture the torque curve, but rather one single point in a continuous distribution.
2. Your model fails to take into account power and other publicly available information about the cars in question. Why limit yourself to kerb weight and peak torque when you can improve the predictive ability of your model significantly using information readily available just about anywhere? The coefficient of correlation may be the best predictor for car performance if we only observed weight, torque, power, but we were only able to use a ratio of two of them. In reality, these restrictions don't exist, and therefore your correlation analysis can be easily beaten by the smallest modifications.
3. Correlation does not imply causation. To see why this is true consider that I could use correlation to predict the chance of rain. I observe the number of umbrellas out of my window and run a correlation analysis by observing whether indeed it is raining outside. Even if this correlation is .9 and I can easily use this model to predict the chance of rain, the number of umbrellas does not CAUSE rainfall. Therefore, if you and me both opened our umbrellas we would not make it more likely for rainfall to occur. In your model, it is unclear whether the weight to peak torque ratio predicts performance due to torque characteristics or other features of high torque vehicles.
Empirical proof:
1. Regress Nurburgring time on your ratio variable while controlling for dry weight, and a polynomial fit to the power curve. Given that no other confounding factors exist you should get a coefficient that is statistically insignificant for the ratio. Anything else is a spurious relationship.
2. Tire patch, power, and many other 3rd factors are correlated to both Nurburgring times and torque/weight. A simple predictor like weight/torque ratio is therefore a biased estimator of Nurburgring times.
QED, next.
Hypothesis: Correlation analysis is insufficient to capture the relationship between torque and car performance.
Theoretical approach:
1. Your model cannot distinguish between two engines that are equal in every way except one has more torque. It cannot distinguish between the Hell Cal engine and the LT4. This is in fact, a most important point. The manufacturers are hell bent on using forced induction and DCT transmissions nowadays. DCTs are generally torque capacity constrained. That's why we see turbo cars like the GT-R and 911 increasing power from year to year while keeping torque relatively unchanged. Your model cannot account for increases in torque at the top of the rev range unless peak torque increases as well. In other words, your model cannot even capture the torque curve, but rather one single point in a continuous distribution.
2. Your model fails to take into account power and other publicly available information about the cars in question. Why limit yourself to kerb weight and peak torque when you can improve the predictive ability of your model significantly using information readily available just about anywhere? The coefficient of correlation may be the best predictor for car performance if we only observed weight, torque, power, but we were only able to use a ratio of two of them. In reality, these restrictions don't exist, and therefore your correlation analysis can be easily beaten by the smallest modifications.
3. Correlation does not imply causation. To see why this is true consider that I could use correlation to predict the chance of rain. I observe the number of umbrellas out of my window and run a correlation analysis by observing whether indeed it is raining outside. Even if this correlation is .9 and I can easily use this model to predict the chance of rain, the number of umbrellas does not CAUSE rainfall. Therefore, if you and me both opened our umbrellas we would not make it more likely for rainfall to occur. In your model, it is unclear whether the weight to peak torque ratio predicts performance due to torque characteristics or other features of high torque vehicles.
Empirical proof:
1. Regress Nurburgring time on your ratio variable while controlling for dry weight, and a polynomial fit to the power curve. Given that no other confounding factors exist you should get a coefficient that is statistically insignificant for the ratio. Anything else is a spurious relationship.
2. Tire patch, power, and many other 3rd factors are correlated to both Nurburgring times and torque/weight. A simple predictor like weight/torque ratio is therefore a biased estimator of Nurburgring times.
QED, next.
Of course correlation doesn't equal causation - we're not going for FDA approval here. Winning in professional basketball correlates highly to being tall (and black). Does it prove causation? No. But everyone knows it's true. 55% correlation on torque-weight alone is pretty damn strong.
Once again, I'm not building a multi-variable regression model... I'm just providing mathematical proof on what the single most important variable is. It remains 100% true that no other individual metric exceeds torque to weight. Test any other individual metric...I challenge you to run the numbers yourself & prove me wrong.
Last edited by Ching Ho; 08-18-2014 at 03:02 AM.