exhaust systems are complex circuits. i do not pretend to be able to design one, but i know the physics behind analyzing them.
exhaust energy is a series of timed pulses; it's very much like chirping/blowing into a musical instrument.
the "resistance" to gas flowing through the exhaust varies with the pulse rate -- the frequency, eg. rpm -- of the gas pulses. it is directly analogous to electrical impedance. "impedance" is the same as "resistance" as in Ohm's Law, but dependent on frequency (as in our engines). at D.C. calculating with resistance in electricity is easy. calculating with impedance, using varying frequency electricity, is complex.
if you get it right, the exhaust has a low impedance -- low resistance -- at a range of rpm (frequency), that can even be *negative*, eg. draw (suction) in a narrow frequency range. really good, smart, tuners can calculate and build that, but it's very complex math with added complexity of real-world metal.
that's why what sometimes seems intuitively like a restrictive pipe -- a long straight "skinny" pipe might be series-resonant at a useful RPM range, and actually be optimum for exhaust flow, where increasing pipe diameter will make it worse.
in sloppy terms -- what i am able to actually build for -- overall low loss, low resistance, eg. short, large, unrestrictive exhaust is a hell of a lot better than most old/early factory systems which were designed to make the car quiet and HP wasn't really considered. (my Rambler came with 1.25" pipe with, seriously, 360+ degrees of bends, but wow was it quiet).
there's some wikipedia pages on series-resonance vs. parallel resonance and musical tuning and all that, but the practical physics of bending pipe for automotive purposes seems to remain a black art.
(lol, now if you made a "trombone" that varied in length via software and servomechanism you could to active exhaust scavenging... all you need is a huge shop and a lot of instrumentation to work it up :-)
