@ no hoper I was,of necessity, somewhat simplifying. The vast majority of SEP's are Carburetted and the starting-procedure is , in technical terms, the epitome of crudity....Basically, you squirt a few dollops of fuel into the inlet tract, then crank, thus your induction-air evaporates this puddle of fto oo rich to fire first time, but continued cranking will lean-out as the puddle reduces, to the point where the ratio is right...the engine fires, and typically a sooty cloud of exhaust is ejected as the engine "catches" and the operator resets Mixture and throttle. if, indeed, during this brief period of starting (accelerating from cranking-speed to , say 900 rpm idling......the mixture-ratio is in the "danger of detonation" range, it is unlikely to do so, simply because the engine-loading is so light......no doubt, with the correct laboratory setup and the engine on a test-stand, you could produce this and analyse the real-time combustion events.
For all practical purposes the firing-to-idling phase is so brief, and lightly-loaded, that detonation is not an issue.
WRT to permanent-magnet," high-speed" (a relative term!) starters. They are approved!
Again , you raise a red herring....IF the starter was sufficiently fast to throw-out the impulse coupling, that would indicate that the engine manufacturer thought retardation and impulse were not required at that rotational speed....QED.
Have hand-cranked stationary-engines with impulse couplings and twirled it fast enough to start without impulse....no broken wrist or thumb!....likewise many motor vehicles and the odd few marine engines.
Normal Lycosaurus piston engine timing is fixed and works extremely well, simply because !- the average pilot is not an engineer and would be incapable of operating an Advance-retard mechanism correctly,
2... the engines are, effectively, operated as "constant-speed" devices therefore timing can be fixed and optimised for that speed.