I'm convinced compression readings vary with everything including
day-of-the-week and the phase of the moon etc.

My 172 M has (eventually) met compression for 31 years (I guess that
would be about 120 readings) but for about 10% of the readings it
required a retest 10 hours later to get a cylinder above 60, at which
point it would be back up. Through the years numbers though were
completely inconsistent except that the inconsistency was inconsistant
too.

I always used the same compression measuring rig (I called it the heart
failure model) so that was not the source of the variation.

John Thorpe preached that removing the upper spark plug before
compression testing could drop debris on any open valve faces reducing
the reading.