[455-4+1]

I think a really important part is being missed here when comparisons are being made on the dyno.
Unfortunately due to pretty poor processing power, and sampling rates, the industry (Superflow? ) settled on 300 rpm / sec as the default test rate.
This is extremely slow and does not compare to actual acceleration rates in vehicle. Often these are thousands of rpm per second, and an engine will respond very differently under these conditions.
If the intake CL remain nearly the same, and volume goes up, so must CSA.
Often cylinder filling with larger CSA heads can be less effective at low rpm, but the slow rate of acceleration on the dyno allows this to "recover" somewhat.
At the track the lack of bottom end, may be made up for with increased performance in the higher ranges, resulting in similar ET. (mph may be a good judge of this?)
I would like to see the tests carried out at higher rates, ie 900 rpm / sec and 1500 rpm / sec, but alot of dyno setups probably couldn't maintain that?