Lies, damned lies, and statistics
 

Recent speculation on future Porsche pricing got me thinking, and the winding road I went down lead to a few interesting conclusions. Among them:

• The 991 GT3 is among the best "performance bargains" in Porsche's history.
• Porsche carefully manages "price/ performance", and has for decades.
• Porsche has not engaged in the recent "price/ performance" arms race between other manufactures.

These conclusions may seem obvious, but I didn't expect they'd be "provable". Through the magic of statistics, however, perhaps they are? Or you can try anyway, assuming you're bored and willing to spend a little time geeking out. Which I clearly am...

How does one define "performance"? Porsche largely seems to define it by power to weight ratio, and within their sports cars they generally suggest higher power to weight is worth more money. Surprisingly so.

Below is Porsche's current base hard top sports car lineup, with power to weight on the vertical (hp/ton), and list price (in log scale) on the horizontal. And a trend line...

https://cimg4.ibsrv.net/gimg/rennlis...2dfed69f49.jpg

The higher the power to weight ratio, the more expensive the car (in general). Cars above the line seem particularly "good value", assuming you believe better power to weight is indeed good, as Porsche seems to suggest. And the 991 GT3 is a long way, relatively, above the trend in the current lineup. Thus it seems a particular bargain by Porsche's standards...

The apparent trend got me thinking. How has it evolved over the past few decades? Surely improving technology means more performance is available for your dollar today than it was a decade or two ago? Yes and no, it seems. Today's cars clearly have higher power to weight ratios, but largely due to inflation they are also more expensive. So again plotting new MSRP, despite fluctuations in conversion rates and inflation, the trend seems remarkably consistent:

https://cimg4.ibsrv.net/gimg/rennlis...7ed533cff7.jpg

Is it chance that the unprofitable 90s are also a low point above?

Interesting... So how do other manufactures compare? Well they tend not to have clear a trend. BMW is more of a scatter plot, but some manufactures are relatively behaved:

https://cimg1.ibsrv.net/gimg/rennlis...073785ebcd.jpg
Clearly Chevy is known for low priced power. Ferrari was more of a surprise...

Finally- is the relatively large advantage Ferrari and Corvette seem to have a recent trend, or has it been this way for a while? Answer- it's recent, because the competition is moving fast:

https://cimg7.ibsrv.net/gimg/rennlis...c3c950f290.jpg

Ferrari and Chevy have clearly chosen to exploit advances in tires and electronics to push performance and hence price/ performance over the last decade. And in comparison Porsche has chosen not to.

Porsche remains hugely profitable, so I'm not sure this means they are doing anything wrong. In fact I'm not sure I like where the chase for more and more power is leading. However I'm also not entirely comfortable with where the trends above lead.

Exactly how long will they keep it up, I wonder?

Food for car geek thought...

Note the log scale on the cost axis! :o

Pretty cool stats. Well done.

gotta love it.... this is brilliant !

another interesting benchmark might be Nurburgring times vs price.

More useless statistics, just drive and enjoy! I don't need a graph to tell me how brilliant my car is!

really interesting... throw an ariel atom in there and lets see an outlier

Awesome work Pete! Really interesting data there. I almost forgot there whas a thread rating system, since I haven't used it but once or twice before. 5 stars man, 5 stars.:thumbup:

Wait, 515hp for the 991 GT3RS? As the graph from the GT4 thread states. Mathematically, that makes more sense! Yeah, that's it. Starting to see that number thrown around now instead of the previous 510.

Very interesting. Great stuff!

Just note that if power/weight is linearly related to log prices, this means that it become exponentially more expensive to buy an extra unit of power/ weight. Also, one factor that might affect the relationship above is the level of interest rates, disposable income and perhaps some other factors, so if you're really serious about it I would suggest controlling for interest rates in this experiment. Nonetheless, really cool. It would be interesting to put several brands together, and see where P's line up (above or below the regression line) relative to other brands.

Nice job and thanks for sharing

Great reading. Thanks for sharing interesting material.

We need guys like you at my firm. If you ever consider a move, PM me, I am serious :-)

Awesome stuff Pete, Nate Silver will be proud. I think I am going to geek out on it. I have a bunch of similar calculations lying around from before GT3 was released, trying to predict it's power/price. :cheers:

That's really cool Pete this is awesome. I took the liberty in taking your model and plotting an extra point for the RS. What was interesting, let's say Porsche goes by the model you presented then and if:

GT3: 475 @1430 kg - 332 hp/mt $131,000 Turbo: 520 @ 1594 kg - 326 hp/mt $151,000 Turbo S: 560 @ 1627 kg - 344 hp/mt $182,000

After plugging in those numbers and if this is the model then my hypothesis for the RS should look something like this.

GT3 RS: 500 @1415 kg - 353 hp/mt $162,000

Following the linear curve of the Turbo and Turbo S with the GT3 being an outlier, then the RS following the same outlier would should fall at a price of $162,000 MSRP.

The 458 Speciale may represent an interesting data point as well...