If you have a stock-ish V8 with steep axle you could could generate actual, hard test data by doing long-distance, constant-speed testing. I've actually done a bit of this in my Rambler to determine optimum highway cruise speed on long trips. (It's 60 mph -- so that's what I drive on trips.)
I drove at least two full-tank fillups at one speed (easy to do on Route I40 east or west through AZ/NM). There's some margin of error but it equals out. I did a few 500+ mile one-way runs at different speeds.
Given a known axle/cam combo, you can drive at various engine RPMs and work out which engine rpm is best. Yes, there's many other variables (friction and all that) but in a narrow rpm range I think the biggest variable will be the cam/carb combo. (For example, below, I don't believe there's much variation in carb behavior from 2300rpm/126cfm to 2500rpm/143cfm; I believe the difference is the engine hydrodynamic junk. I know my spark is all in by 2000.)
Then look at curves for that cam grind; (in this case, what's it doing at 2300 rpm) and then get a cam grinder to tweak one for say 2000 rpm, assuming you have enough grunt at that speed to be drivable -- my little LP six doesnt!) and pick an axle to match.
These are propane MPGs; 14 mpg/ LP --> 18.93 gasoline; 15 mpg --> 20.28 gasoline. I've got every single mile and fillup recorded on this car back to 1994 (OK so I'm fanatical).