I dunno why this 'calculations' thing keeps popping out. And it always does in the same context, if something is desired to be dismissed, it's a good excuse, and rarely proved.
I think 'too good for my taste' is the real problem with these data, and it doesn't matter if it's flight test either, you can always find excuses : the Russian and Finnish 109G datasets are coming from flight tests curves were also 'dismissed as abberant', simply because the data they are good results. If the 'calculations' excuse doesn't work, dismiss them because 'they are abberant'. It's very simple to show this way whatever you want with a few tests, if you are showing only the ones that you like and dismiss all the rest. Curiously, wheter it's a calculation or not is never an issue with poor/low results.
There are no such reservations when it comes to british calculations either, see the calculated MkXIV specs, but you can't see a 3-paragraph whining how these figures were unlikely to be reached by serial production planes, and with how much reserve they must be treated.
So when I see someone talk about 'calculations', I take it as an excuse. Companies wouldn't do calculations if they would not spare them money with similiar results as prototypes and tests. Companies do not forward overly optimistic specs to Air Ministries , only to make their poorer performing planes not reaching those specs of their own turned down on delivery, and loose profit. The company is interested to be conservative in it's specifications, it can't have trouble from that.