"Did you try an unstable pole in the engine control loop, so the engine would run faster under load than when free running? "
Now you're talkin'. Igor Burger has done this with electrics. You can search here for his stuff, or perhaps he will elaborate.
I was 80% sure it was Igor -- but when I tried searching on his name I didn't come up with much. Hopefully he'll see this thread.
In theory, the unstable pole thing would make the system "sense" a higher load, and respond by running faster. Unfortunately, it'd also sense any glitches, bearing issues, bugs on the prop, etc., and take those as reasons to run faster or slower, too. All such problems would be exacerbated by the increased variability that one encounters with slime power vs. electric. No matter what, much excitement would ensue; only if one were lucky and well prepared would it be the sort of excitement that one enjoys at the time.
Much as I dislike the religion called Systems Engineering, you might start by determining the requirements. Stunt people will be of little help: they can't decide on the definition of "power". Constant RPM might be something to shoot for. I don't know what frequency response you need, but if might be fun to determine experimentally the response of an engine with an intake throttle.
I was thinking that a wise control engineer would start by seeing what could be done with constant RPM, as that seems to work so well with electrics. Constant, high RPM and a big, low-pitch prop ought to come close to what an electric does with regulated RPM. But experimentation would be the only way to know.
Frequency response would be an interesting problem, because it'll vary from engine to engine, and from run to run on the same engine. I can easily envision getting a system set up to work perfectly, then tweaking the needle valve and having the thing burst into Really Embarrassing oscillations once it's in the air. Measuring the response of the engine to the throttle would be the first thing to do once one had ones sensors and actuator on line.
You must be an engineer, or work closely with them, and given your cynicism about systems engineering it almost has to be in aerospace. Systems engineering can be great stuff when it's approached with the right attitude, but I've seen it turn into a magnificent vehicle for empire building by folks with little realization for the practical impacts of what they're doing, and I've seen it turn into a great excuse factory. On the other hand, I've seen systems designed without good systems oversight that cost way, way too much for the job they do, and I've seen systems designed without good systems oversight that stumble and fall because they're just not up to snuff.
Done right a good systems engineer asks "how can this goal be accomplished with the least expenditure of time, money and other resources", then comes up with an answer to the question that doesn't have some huge glaring holes (like a need to build an aircraft half out of titanium and half out of tungsten, when the competition are doing the same thing with aluminum and balsa wood).
Systems engineers -- can't live with 'em, can't live without 'em.
You can make your own stunt fuel for about $12/gallon.
Could you please find my thread on fuel sources and give me a run down on where to get the stuff to do that? It may be the best solution, unless I change over to electric now, before I'm carried off in a sea of nitro and castor.