The Fast Lane Car tried to sketch some answers by going to a shop in Denver, Colorado to run a 2011 Jetta TDI with a six-speed DSG transmission on an all-wheel-drive dyno. The thinking was that if you ran all four wheels the car would think it was on the road, whereas if you ran only two the car might think it's being tested. We'll get straight to the numbers: the stock sedan is quoted at 140 horsepower and 236 pound-feet of torque. When run with all four wheels turning it produced an uncorrected 114 hp and 214 lb-ft at the wheels. When run with just two wheels in motion the Jetta got 113 hp and 188 lb-ft at the wheels. Reading the graph, we're told that power differences between the two runs were as much as 15 hp and 32 lb-ft.
You'll need to take some salt with these numbers, though, because the dyno and test protocol in the video are nothing like those used by the EPA. The shop attempts to trick the Jetta into 'emissions testing mode' by using the front wheels only for the two-wheel-drive run, but we have a feeling the software code at issue is far more sophisticated than that, since the ICCT, UVA, the EPA, and CARB investigated the situation for more than a year and couldn't figure out. Also, the technician adjusts for being a mile above sea level with a correction factor of 1.2 applied to horsepower and torque, which inflates the disparity in the final power differences over the two runs.
Go to YouTube and read the lengthy comments on the video left by Andrew Price for a more thorough dissection of what could explain TFL's disparities. You can watch the video above, and feel free to try and dissect the results yourselves in the comments below.