Inside Line reviews the new M5

Originally Posted by dunderhi

I will gladly take your advice to read about experimental design and statistics (to be fair, I get paid to do it anyway).

Since richschneid is the recognized authority in these fields, then he's the man and I will no longer question his mastery of Design of Experiments, test processes, data collection/reduction, statistical analysis, or the analytical rigor associated with the typical automotive journalist's testing process. Thanks for the insight.

I am being both "argumentative" and "serious". That is how scientists and physicians relate to each other. The study of physics and the study of medicine are both massively dependent on the "Design of Experiments, test processes, data collection/reduction, statistical analysis, or the analytical rigor associated with the typical automotive journalist's testing process." The actual testing procedures of cars are not designed by the journalist. I assume the procedures and testing equipment are designed by people with formal training in science and engineering.

I have an undergradute degree in physics and mathematics including advanced statistical analysis and a graduate degree, an MD. The MD degree requires a deep understanding of experimental design involving large numbers of variables, some known and some unknown. This is the same basic process as the variables in car testing. Only the human body is a lot more variable and complex than an automobile and so is the variability in testing conditions. Dunderhi does an excellent job of pointing out the same types of variations in automobile testing. (BTW, I have read every issue of Car and Driver and Road and Track since 1964.)

I am not trying to be argumentative just for arguments sake, I am just trying to join in the discussion about the relative merits of various automobiles. Just like in medicine and physics the principle in automobile testing is to maximize the probability of a meaningful conclusion by keeping as many variables the same in vehicles being tested. The fewer the differences in the variables the greater the probability of the results and conclusions being accurate.

More to the point, I am on my fifth consecutive V8 BMW since 1992, including an E39 M5 and an '06 650i sport package. The 550i xDrive is my first AWD vehicle. My "seat of the pants" impression is that the xDrive system is so sophisticated in it's torque transfer front to rear that the the handling is superior to my RWD cars. I would very much like to see if this translates into actual improvements in track times on a tight road course. A test I read of the X6 described the on track handling as extraordinary. The low speed acceleration is also enhanced in high torque vehicles equipped with AWD and might be a significant advantage in accelerating out of low speed hairpin turns. On the other hand AWD adds weight which will hamper performance in both handling and acceleration. Only a formal comparison test would determine which factors will predominate in a given vehicle.

But all this is conjecture on my part. That's why I would love to see a direct comparison test done by a professional driver and a reputable magazine. People on Bimmerfest spend a lot of time discussing and argueing about these questions. So, it makes sense to me that such a test would be welcomed by all, even if it not "conclusive" but just "highly suggestive". In mathematical parlance this is what we call a "trend" and is used if the test shows a p value of, say, 0.07 or 0.08, instead of 0.05, but even that is an arbitrary assignation.

Thanks, all, for a highly intelligent and informative discussion.

Originally Posted by richschneid

I have an undergradute degree in physics and mathematics including advanced statistical analysis and a graduate degree, an MD.

We should expect nothing less from a 550 driver

Originally Posted by richschneid

That's why I would love to see a direct comparison test done by a professional driver and a reputable magazine.

While the results of proper test done by professionals would be welcomed, I'd be much more intrigued by the benefits the torque vectoring/ awd give to the average/above average driver.

Below is the 2009 article (which I'm sure you've read) from 2009 about the system.
http://www.caranddriver.com/features...ling-tech_dept

Cars behave not unlike canoes, which turn more sharply if you drag an oar on one side or pull harder on the other. With a car, the stability control applies drag, or brake force, to individual wheels to veer the car away from disaster. With torque vectoring, however, the computer "paddles" harder on one side to help steer. How?

BMW's rear differential, supplied by German maker GKN Driveline, has a planetary gearset on either side, engaged to the wheels through motor-operated clutches. Cruising straight, the clutches remain open and the rear wheels receive equal torque. The computer watches steering angle, throttle position, yaw sensors, and so on. Steer into a corner, and the clutch controlling the outside gearset closes.

The 1.1:1-ratio planetary gearsets mean that on slick ice, the outside wheel would actually spin 10 percent faster. On dry pavement, the oil-bathed clutch plates-which never fully lock-can't spin the wheel any faster. They only transfer a percentage of the torque, "just a push," says GKN's Theodor Gassmann, manager of advanced engineering. That push on the outside wheel is enough to help the big SUVs carve a fiercer line.

Acura's "Super Handling All-Wheel Drive" also has planetary gears, but it's a part-time system and only pitches in on acceleration. BMW's full-time system works under deceleration as well.