My experience with with the small Cox TeeDee's and SPI (and a few reed versions but no experiment with them)
Bottom line, on these engines, is the fuel load is heavy with castor and moves slowly..I suspect that adding high speed but very brief pure air into the rapidly spinning crank case volume--- offsets, or enhances if you will, the inefficient fuel/air charge from the venturi NVA induction
Important variables:
Atmospheric air density and pressure
Sub and super sonic air moving
Controlled duration of the clean air charge
Points of diminishing returns
Starting and unloaded NVA mixture and leaning
Muffler, vs open exhaust...SPI engines should not run with a muffler...too much dirty air in the SPI charge
Very brief timing of un-shrouding of bottom of piston before closing of fuel and boost ports... there IS a point where there is too much SPI
I have fussed about with SPI on Cox TeeDee .049 and .051 and there is a distinct minimum and max that yield increase RPM and torque
I have fussed with opening TeeDee Venturies, all on bladders, and the effects of SPI are more pronounced than the venturi drilling
One of the trials included increasing the SPI and Venturi...engine fell flat...went with smaller spi larger Venturi...xx rpm
then with larger spi and smaller venturi...xx more rpm
It would take a much more controlled environment than I was willing to deal with to determine bench vs in the air and prop load differences but my modest experiment proved to me there is a cross over relationship and a LOT of variables
In the end, I morphed 4 of my 7 TeeDee's to be able to swing a 5.5x 6 prop and haul a 9.7oz Lite Hawk variant out on 42" lines with authority---- similar to same birds with Norvel Aero .049s on the nose