Guys, it is not my goal to stir controversy just for the sake of it, or to declare that I hold keys to the best 4-stroke setup while everyone else is wrong. In fact, my setup was largely borrowed from those used by the Australians and French, who use 4-strokes at the WCh level.
After re-reading a bunch of older posts on 4-strokes (Saito 56 specifically), I've concluded that the main difference between mine and Bob's setups (and those preached by the Dallas group) is in the fuel system. We seem to agree on pretty much everything else: relatively large-diameter, higher pitch props (I use APC 13x6 and Bolly 12x6.5 3-blade on my 56); modest RPM - mid-8000's up to 9000; prefer a fixed venturi over RC carburator; set needle a few clicks rich of the peak RPM.
The only thing that's different is that I use uniflow tank. I've tried to make non-uniflow systems work (hard tank, clunks, etc), but the result was a significant speed-up from beginning to the end of flight (about 0.3 sec). In making my custom uniflow tank, I kept in mind the need to GUARANTEE good fuel flow to the engine. Due to much smaller fuel consumption, I was able to keep the tank short, resulting in only ~ 4" of fuel line from NVA to the pick-up. With this 3 oz tank, I get about 7+ minute runs, with lap time change less than 0.1 sec from beginning to end.
My guess is that the reason why "Dallas non-uniflow setup" setup consumes to much more fuel is because you need to compensate for fuel pressure differences simply by "flooding" the engine with so much extra fuel, it does not "feel" fuel head pressure differences. This also explains the "burnout" effect - i.e lack of power in overheads if set lean. At the top of the circle, the centrifugal force is offset by the gravity, causing less fuel pressure. Being very sensitive to fuel pressure, 4-strokes start to lose power... unless set rich. Uniflow systems seem to be less affected by that.
That's all I have to say.