Hmmmm.... I see what you are trying to get at now.

The problem is not so much that the initial velocities are different, but that we don't know the velocity curve. This means we don't know how to compare the averages. (It will not, unfortunately, be as simple as a linear transformation like you suggest.) What I am not able to say is if you procedure is "good enough" for a decent approximation.

If we assume a linear resistance term, I can calculate it using Calculus:

where

(This is first the equation of motion and second the solution to the equation.)

So

, where T is the time of flight.

For the quadratic resistance case:

for

So

As you can see in both of these equations, simply scaling v0 is not going to be enough because of the T dependence. You'll just have to get the numbers and see how they turn out.

If you wish to use these equations, the velocity solutions are from "Analytical Mechanics" 4th ed. by Fowles, pg. 49. You can say I assisted you in finding the equations for the average acceleration.

-Dan