I don't doubt your methods or results, re the P-40B, but it indicates to me you've discovered a potential FM error.
How could it accelerate at less than half the rate of its contemporaries? Look at the distribution there - the P-40B is way, way off by itself. I don't see how it possibly could have been that bad in RL. No wonder it falls out of the sky after a couple tight turns.
Look at the difference between the P-40E and B. The B is 900 lbs lighter and has about 100 less hp. How could its acceleration have dropped so far? I think your testing has uncovered something significant here.