I've read a couple of books on this topic and there's two main types of induction charging; inertial charging and pressure wave charging.
-Inertial charging you keep making the pipe longer until there's no gain, this type will have a smooth torque curve.
-Pressure wave charging is where pressure waves travelling up and down the intake runner as a result of valve opening and closing are used to increase the pressure at the intake valve just before closing. If this effect is strong the torque curve will wave a wave super imposed on it (assuming it isn't filtered out by curve fitting)
There's a couple of places you can get a rough approximation for pressure wave charging lengths. For peak torque boost due to the intake at ~6000rpm you need a runner length of ~600mm (inlet of the runner to the valve) for the 2nd harmonic. You'll also get a bump at 3000 and 4500 rpm as well due to the other harmonics. (
http://www.bgsoflex.com/intakeln.html) There's a quoted range because it assumes values for the other influencing factors.
Inertial charging, as far as I know, appears to be done more empirically unless you have access to some pretty sweet CFD based on real test results or some packages that have simplified the flow equations and tuned them to match actual results. Generally longer is better.
The plennum is the other factor that influences the intake charging. Small plennum reduces the peak in size and rpm value but spreads it over a larger rpm range, the bigger you go the closer you get to individual runners which have the largest but narrowest peak at the highest rpm (assuming all else is equal).
Take all these factors and look at factory/ oem designs and their intended use and you start to get a good feel for what the manifold you want will look like. Optimising it is probably beyond the means of most enthusiasts. Mine will have long runners a smallish plennum to boost low rpm torque before the turbo kicks in.