Why still no dive acceleration difference?

[JtD]

BlackBerry, I think what you're saying about Tempest vs. P-51C is about right, even if your sources are vague and not always the right ones (you're quoting a comparison between Tempest and Typhoon, for instance). A direct test between the P-51C and the Tempest V revealed that the "Tempest tends to pull away" - which is a marginal advantage for the Tempest. You've tested a marginal advantage for the P-51.
Given the average accuracy of the flight models, which was aiming at a 5%, this is something that simply may happen between individual planes, it is no indication that the general algorithm is wrong.
Mach effects are modelled, not extensive enough for accurate high speed performance imho, but they are there.
The differences in initial acceleration between individual planes is there, if you fly them properly. Some of the acceleration differences you see as simple statements "this one is better" has a lot to do with engine management. If say an Fw 190 and a P-47 cruise side by side and then go into a full power dive, the guy in the 190 slams the throttle forward and off he goes, while the guy in the P-47 adjusts mixture, then rpm and then the throttle and then starts to accelerate. Gives the 190 a two second head start. But even without considering this, if you compare a 190A-4FR with a P-47D-22 at medium altitude in 4.11, you'll be getting something similar to the test you quoted.