In an IR setup the air/fuel mix does not flow continuously into the cylinder, each time the inlet valve closes flow stops, however the air/fuel mix is still moving and as it stops, it compress's a bit in the inlet runner before bouncing back out again. How far it bounces back out is a function of the speed and size of the venturi's plus all the other stuff at the time-rpm, runner length, valve size, throttle opening, engine load etc... So we effectively have two types of reversion, the low rpm type with low airflow rates where the overlap literally blows the fuel mix back out,........Then we have the higher RPM type that gets 'bounced' out!
I need to clarify and add a couple of extra thoughts here... Overlap occurs while the piston is at or either side of tdc, with 'le mans' cam exh valve is still open when the inlet opens @ 52° btdc , exh valve does not close until 42° atdc.... This is where having a longer rod helps... On your 3.00" stroke with the longer rod the amount of piston movement during the overlap phase is quite a bit less than what it would have with a shorter rod, since it hangs around TDC for a bit longer & does not move as far or as quickly it will not tend to pump as much out into the inlet port at low rpm when there is little or no exhaust scavenging taking place.......... That's about as simple as I can make it sound & I am sure some boffin somewhere would kick my butt for that terminology.
Its also why a 4bbl has much more going for it since there are 2 or 3 inlet valves open at any given time and a continuous flow demand as a result, the stop/start scenario is still taking place in the individual runners of the manifold, but the plenum in the manifold helps to kill the effect of these. Tom from germany was in the process of fitting one of the 'new' ida intakes with what i would call a divorced plenum prior to his accident, but i don't think it ever got fitted to the car, was not on it at time of accident.