Velocity is great...if you can control it. Calvin Elston is a smart guy and has done some amazing things with headers. But, in the end it comes down to designing a system from the top down, or from the intake. If you think about it terms of valve sizing, and start there, it's a bit easier to understand why.
If you can use a bigger intake valve, AND you can get the port volume to support it, AND you get the correct cam lobe the engine will always make more power. Usually the limiting factor of intake valve diameter is how close it is to the exhaust valve. If you can make the exhaust smaller, and make the exhaust side more efficient, you can use the bigger intake and a smaller exhaust and make more power. IF you can make everything else happen. Then of course you can use smaller primary tubes and a merge collector that is sciences out for your combination. That includes the exhaust lobe of the cam.
It also comes down to how well can you fit a header into the car. IIRC Calvin likes the primary tube to match the port exactly and then come straight off the head for at least 4 inches. Sometimes that's impossible. And he likes to keep the radii big and to a minimum. Also sometimes it's hard.
I learned from my testing that most guys use too big an exhaust valve, too long of primary tube and too short and too small a collector in CONVENTIONAL 4-1 headers. 4-2-1 headers are a horse of a different feather.
In the end, you want the exhaust to act like the expansion chamber in a 2 stroke exhaust. You want the pulse to head back up the primary to close off the exhaust port without letting exhaust gas back into the chamber. Like anything else, the smaller tubes are more responsive to changes and the merge collector acts similar to an expansion chamber in a 2 stroke. It stops exhaust pressure (and atmosphere) from heading back into the collector.
If you google Calvin Elston you will find his web site and you can read his stuff. He's a sharp cat.