<<Caveat Emptor on those fuel burn vs. power ratings.  Unless we mate a real
time torque sensor to the engine in flight at altitude, power cannot be
measured accurately.  Eric Ahlstrom>>

The snip above is just to remind the reader of the subject (as if they had
to be.  I think I'll stop with this one!)

George, as usual, said it better than I could, but here are a few other
points:
My estimation of BSFC for SI engines (George has the real numbers) was
assuming a compression ratio of 8.5 to 10, not the 6 that Eric mentions - I
don't know of an engine with a compression ratio that low.  Engines
basically have 3 "limits" that determine the most advanced the spark should
go.  One is, as previously mentioned, detonation.  Second is MBT - there is
no reason to advance the spark more than is required for best torque.  Third
is pre-ignition, not usually a factor in these engines, but common in high
performance marine engines; and it is usually aggravated by retarding the
spark not advancing.  Because the aircraft application is by necessity
conservative, most engines are not "octane limited," meaning that they will
get to MBT before they reach detonation and is generally a good thing.  The
PRISM system holds the promise of being able to increase the compression
ratio so that in most conditions the detonation limit and MBT could
coincide, as the system will protect the engine from detonation.  And
finally, as George says, running LOP REDUCES the octane requirement, not
increases it as Eric says.

Gary Casey