[jluetjen]

A couple of other things to consider...

What cam are you using: Cams with a lot of overlap I suspect will be happier with higher exhaust velocities (slightly smaller headers) since the extra inertia will help scavenge the combustion chambers. But cams with a fair amount of overlap also tend to be higher rev'ing cams, so now you're moving more air through the headers, and the pulses are at higher frequencies, then with a comparabily sized engine with a milder cam and the same sized headers. So in abstract terms its hard to describe which header really is "smaller" when compared to the engine that it is attached to.

How does the header match the exhaust port? I've also heard that cams with overlap like to have a mis-match between the port and the manifold where the manifold is slightly larger then the port (especially on the outside of any bend in the exhaust shape). This mis-match tends to echo back positive pressure pulses and allow negative pressure pulses to continue into the port. This is the concept that allows Anti-Reversion headers work.

This is one of those situations where a carefully crafted set of experiments combined with some solid DOE analysis would go a long way.