If the magneto or the plug wire was intermittent, don't you think there'd be some backfiring, if not some sputtering?
Not at all..
misfire is a timing issue,
for that you need a spark while a valve is open.
and while a defective coil could cause a spark delay (rather a prematurely collapsing field) and as such cause backfire theoretically..
it has to fail in a very specific way to do so.
What I have in mind is the actual coil being just barely cracked, so when cold it's acting normal and working fine,
and as soon as it reached the "critical temperature" the crack opens up, creates a short with an adjacent winding (making it a much smaller coil effectively)
the coil no longer fully charges and the weaker field collapses into a much weaker spark
(maybe even skipping every other rotation or so);
causing the engine to loose power.
basically the temperature controlled mechanical version of a rev limiting coil
And yes, that scenario is nearly impossible to reproduce if you'd try;
the most likely coil failure is it just no longer works at all.
But since I've seen a lot of temperature failures in electrical connections lately (generally they belong to electronics rather)
I wouldn't want to rule them out any longer.
Especially since I had a somehow comparable failing coil in my 914 once..
it failed to spark on one cylinder when getting hot and one cylinder only.
changed plugs, wires, boots, distributor cap and finger first.. no difference. got a new coil.. and it's been fixed.
(haven't seen such thing before or since)
Anyways, long story short...
it's a thermal issue of sorts, either fuel or compression or spark,
you named one, I added the second
and now we need someone kicking in asking to check the cylinder head gasket and bolts to complete the list
'sid