[Ryephile]

I can't give you a good answer because I don't know the dyno, the car, or the testing procedure....other than it appears to maybe be a 6MT with the pulls in 5th gear for some reason. The dyno might have an increasing measurement attenuation at those really high roller speeds, but that's just a wild guess.

This is why it's so important to do back-to-back dynos on the *same* dyno. That way whatever error you're introducing is the same error for all measurements. It's the delta percentage that matters, not the shape, tilt, and definitely not the absolute number.

If Elite did the "before" measurement on one dyno, and then the "after" on a different, there would be zero compatibility with the data sets.