There's no harm in the type of test drive you suggest. However, it is not a conclusive test of an engine's health. It won't tell you if an over-rev has stretched a timing chain, putting it at (or beyond) the specified tolerance (do they measure this as percentage slip?) and giving it a greater propensity to snap unexpectedly (potentially wrecking your engine). A chap on here who'd owned a 996T from new, with OPC only servicing & warranty, had his engine let go with a snapped chain and a handful of range 2's were sufficient for them to reject the claim and leave him with a big five-figure (Â£ not $) bill. He put it down to the only time he'd ever let anyone eles drive the car (which was round Silverstone).
Anyway, the point is, it's dangerous to suggest that a test drive will tell you everything about an engine's health as it most definitely will not.
The chains do not stretch. They may break but it is do to metal fatigue, failure of a chain tensioner (material fatigue), or something else comes loose and gets between the chain and sprocket.
The primary risk to the engine from overreving comes from several areas. One is some part lets go. A rod, or rod bolt or the crank lets go.
In fact these are at risk of letting go at any time the engine is running. Yet, with all the crying about the IMSB no one ever suggests refreshing the rods/bolts and crankshaft when buying a used car and spending money to freshen it up.
Anyhow, the other is the valves float and the heads kiss the pistons. The best one can hope from this is the valve heads are just bent and none break loose. The engine will almost certainly have a misfire afterwards.
However, it is a rare engine overrev event that the valves just kiss the pistons.
Generally the valves impact the piston hard enough one or more valve heads break free of the valve stem. In this case the engine immediately makes it known by sounds and behavior that the overev had a very negative impact on the engine's health.
Now the techs tell me that overrevs do not matter if (big if) the car/engine is not in for any engine warranty work.
If the car/engine is in for a possible warranty claim, then if any serious overrevs are present the warranty claim can and almost certainly will be rejected.
If the car/engine is being considered and a warranty is contemplated or desired, one must get it in writing the current overrev counter values -- if there are any in the higher ranges which can be sued to deny a warranty claim -- that these existing overrev counts will not be cause to void the warranty.
If the car/engine is not capable of being warrantied, then the buyer can do as he thinks best.
If the engine has some run time on it after the overrevs -- the techs I talked to told me an hour is a good number which given the average speed of a car over its life is 30mph this works out to a 30 mile test drive, hence the 15 mile test ride and 15 mile test drive or of course 30 miles total.
More is better. But how much more they didn't say, and there is a law of diminishing returns. If the overevs are in the distant past that's one thing, but the engine is still at risk of something failing that has nothing to do with the overrevs just the fact the engine and its parts have a finite service life. The engine is only as good as its weakest part. Crank. Rod. Rod bolt (I know of a guy who lost his high miles 2.5l Boxster engine due to a failed rod bolt. The engine had IIRC 150K miles on it and its was tracked quite a bit.) Cam chain. Valve spring and so on.
Now the overrevs counts may suggest the car and the engine have been thrashed on the track (or even on the street) and this can be sufficient cause to reject the car simply because it was thrashed.
A pressure test (compression/leak down test) is not a reliable means to determine an engine's health. It is possible the engine won't even run yet pass a compression test/leak down test just fine.
It won't tell you if a rod or main bearing is knocking, if one or more lifters is noisy, if the VarioCam (Plus) system is not working right and a number of other things that could be wrong.
Also, even if the engine is healthy enough to run it may deliver better numbers than its true condition warrants. The engine as it cranks gets plenty of oil sprayed into the cylinders. This oil at cranking speeds acts to improve the sealing of the piston/rings and cylinder.
Only when the engine is started from cold and idled/run/driven from warm to hot and run long enough can the DME flag a weak (or even a too strong) cylinder by measuring the amount of acceleration each cylinder's firing pulse imparts to the flywheel. If this acceleration is outside of acceptable bounds the DME will turn on the CEL and record one or more misfire error codes.
Running the engine driving the car a good 30 miles in a variety of driving scenarios is the best way.
Not only can you confirm the engine runs right, sounds right, the DME can go through its readiness monitor check out (the Turbo DME I'm told does this twice per trip if the trip long enough) to confirm all critical sensor systems are working properly so when they are then used to verify the engine's overall health is good there's high confidence the engine is indeed healthy.
If one wants to spring for a compression test after the test ride/drive, if the seller is ok with having his car wrenched on in this manner, then go for it. But do not fool yourself into thinking they are infallible or in my opinion even worth the money. They are better used to identify where a sick engine may be sick, not that an engine is not sick.