The BMW M3 is a (virtually) universally appreciated car. Pretty, practical, fun to drive—there’s not much to dislike. If you want one today, the E36 body style of the 1990s tends to be the cheapest. But, the question becomes, how much “M3” is still left in a decent example after all that time?

has taken it upon themselves to find out, as host Alex takes his own M3 to a power-measuring dyno (or “rolling road” in British English) then conducts some informal zero-to-60 testing on a real road.

This particular M3 actually looks really nice from where I’m sitting. I think I see about 130,000 miles on the odometer in a shot of the gauges. Apparently the car has , but the video seems to indicate its heart is healthy.

Alex and Surrey Rolling Road’s dyno operator Charlie establish that the car made “317 brake horsepower” when it was new.

“A really, really good one of these will still do the 317,” Charlie explains. “But it’s very rare. An exceedingly bad one of these will do about 250. Usually when the cats are blocked and the air-flow meter has failed. An average one will do 300.”

A dyno measures a car’s power by running on what’s basically a giant treadmill. After taking the test, the M3 in the video puts down 279 horsepower.

The car seems to have lost the most juice at high RPMs. Charlie explains that the engine’s valve clearances could be the culprit; which “tightens everything up” and leads to restrictions in the top end.

Of course, there are different ways to measure horsepower, most significantly at the brake and at the wheel. Measuring at the wheel, like a dyno does, nets a lower number since it puts more parasitic loss between the engine and the measurer. This video doesn’t make it totally clear that they’re comparing stock-wheel horsepower to wheel horsepower or stock brake horsepower to brake horsepower, but I’ve got to assume the guy owning the dyno shop knew to take that into account.

The second test this old M3 goes to is empirically useless—four zero-to-60 runs averaged and compared to the stock figure which Alex calls “between 5.2 to 5.7 seconds.” That’s already way too big a range to compare anything to, but apparently the best he’s able to do anyway is a 5.96-second stopped-to-60 pull on his fourth try.

The interesting thing is seeing a real-world measurement of a “realistic” car. We’ve learned to take any number reported by an automaker with a pinch of salt, and it’s cool to see what really comes out of an engine—especially after a few years of use.