"whatever diameter your header's collector ends in is the minimum diameter your exhaust needs to be. anything smaller hinders the header from scavenging the exhaust gases effectively."

I'm not trying to be a pain, just thought I'd add some.
One thing that needs to be said is that the reason why header collectors are bigger is because they are "collecting" the exhaust gasses from each primary, hence the name. Obviously where two different streams of air, traveling at diff. speads (because not all cylinders are firing at the same time), collide you are going to need a little extra room where the gasses can "simmer down" a little. Once the gasses have been combined into that one pipe the name of the game is keeping that gas moving. To scavenge the gasses you absolutely need velocity. Let's say you have a golf ball representing the actual amount of exhaust gas produced from one cylinder in an engine(I know solid and gas but humor me here). Let's say that that golf ball comes out of a header collector into a 1 in. exhaust pipe. Obviously the ball will not fit but in the case of a gas it will, it will just be a major restriction. This setup is going to have very good velocity but bad flow capability. Because the "ball" of gas is going to have to be pushed thru the too small pipe(in other words back pressure), you can throw out the scavenging effect idea. Now, let's say that that same ball comes out of the collector into a 4 in. pipe. Well that fixed the flow capability but now you are going to lose the scavenging effect because that ball is not filling or even coming close to the pipe's inside walls to create a sort of suction created as the momentum of the ball pulls the air behind it. The last setup is the pipe that "fits just right." The ball comes out of the collector and because the pipe diameter is big enough for the ball to travel through it, velocity is okay. But also, as the ball travels through the pipe, it fits snuggly enough that it pulls, or scavenges, the air behind it creating a sort of momentum that goes clear back to pulling air out of the cylinders which is where the power from an exhaust is actually coming from. That said, I did leave out a factor -engine speed. As engine speed increases, the size of the "ball" is going to increase proportionally, obviously. So the goal is to create a pipe size that is effective at the most used rpm's. I think most would agree that choices for a stock to mild KA would be 2.25 in. and 2.5 in. Because a KA could probly work with both sizes it is just personal preference. In my opinion, since I probly spend 95% of the time in the lower 3/4 of the rpm's I would go with a 2.25 although I think I could live with a 2.5. You also have to factor in that 2.5 is going to be louder, too. If you are worried about top end I really don't think you will lose any with a 2.25 if you have a good flowing muffler, high flow cat, and mandrel bends. The idea that isn't true though is that you need to use whatever piping your header collector is because you need to base the size of your piping on the amount of exhaust gasses your particular engine will produce. Consider that stock piping is 2 in. with a restrictive muffler, resonator, cat, and cast exhaust manifold. 2.25 piping has approxiamately 21% more cross section area than 2 inch. I will run 2.25 in., when I get the money:) and time. One more example I thought of is the hour-glass shaped header collectors on those high output Winston Cup engines. At the smallest part you would think it would barely be big enough for a honda but, it builds power by using the velocity-scavanging-and other factors like wave pulses and stuff. Granted, those things are precisely tuned for each specific engine but still. Okay, I have definately put in my 2 bucks here.:oface
If I am wrong anywhere please explain where :thinker
-Scotty