One other thing to add based on your second post...
Measuring engine output at anything less than WOT is difficult; partly because of that velocity change we talked about. If you opened that 70mm throttle body enough to simulate the same flow as the original 65mm throttle body, you wouldn't get the same dyno results as if you never switched - partly because you enlarged the plenum behind it and the airflow is now going from fast to slow.
There are other factors involved, but I don't know if you want to speak theory or practice... The engine's ECM is a big factor. In practice, if you hold your 70mm TB open enough to simulate the flow of the 65mm, the ECM is not at WOT fueling maps, but in theory you can get close to your old dyno numbers by simulating the flow of 65mm through a 70mm TB.
You have also made a very subtle change... We're discussing theory of going to larger intake passages, but in practice you probably haven't made any massive changes that have caused any "mismatches" like I discussed earlier.
In my hotrodding background, we typically discuss theory while hypothetically comparing a single 750 cfm carb with dual 600s. Then you're talking about massive differences and the velocity/flow potentials are vastly different.
... which brings up another point (and then I promise I'll shut up
) MPFI versus TBI/Carb are a bit different. The TBI and carb mix fuel and air before the TB, so the mass of the airflow is much greater as it goes through the TB, plenum, and intake runners, while the port injection doesn't inject fuel until after all of that. An airflow change on a port injection engine has less impact since the inertia of the "dry" air is much less. A "wet" intake has several design factors to consider. Not only do massive changes in velocity tend to have larger dyno impact given the greater mass/inertia of the incoming charge, but they also tend to make fuel fall out of suspension and puddle.