Jim Nordin wrote (in part):

The idea of shock cooling has been a subject brought up many times on the
LML and others, generally as a passing statement ... "I avoid pulling the
throttle at altitude to avoid shock cooling." I still struggle with the
implied harm from the effect ... that is ... what is it? I think most
imagine a cylinder head cracking...

Jim,

Shock cooling is a serious and real problem with aircraft (and other
air-cooled internal combustion) engines. While it's theoretically possible
to crack a cylinder or other component, I've never seen a crack that could
be definitively associated with shock cooling. The basic shock cooling issue
is caused by the different amount of thermal expansion between the aluminum
and steel used in engine components. There are a number of places where this
manifests itself, but one very important and easy to visualize problem is
the cylinder to piston interface.

Since the aluminum pistons expand more for a given temperature rise than the
steel cylinders that contain them, they are actually fitted with a
significant (and carefully measured) clearance when at "room temperature."
Through years of experience and actual measurement, the manufacturers have
determined the correct "room temperature" relationship to achieve an optimal
clearance--and therefore minimal blow-by and piston "slap"--for normal
operation.

Shock cooling disrupts this fit. If you pull the power to idle while at
altitude, the internal heat source is immediately removed and the intake air
source essentially cut off, but the internal engine components initially
remain at operating temperature. The cylinder walls, however are exposed to
the cooling air flow through the engine compartment (typically below zero if
at altitude) and they cool and contract very quickly against the hot (and
expanded) piston. If the temperature difference is great enough, the
lubricating oil and hot gases are forced from between the cylinder wall and
the piston, causing the piston to "scrape" on the cylinder and track
aluminum onto the cylinder walls. I've never heard of this actually causing
an engine seizure, but at least theoretically it can happen. The more common
occurrence is for the engine to continue to rotate, scratching the pistons,
tracking aluminum on the cylinder walls, and destroying the piston rings in
the process. When you get on the ground and all engine components are again
at a common temperature, the piston/cylinder clearance is basically
restored, but the surfaces are no longer polished/smooth and the rings no
longer hold an air seal--hence no compression. A borescope will quickly
confirm these things.

This is a very real and observable occurrence, and I hope I've given you at
least a start on understanding the physical component properties that cause
it. This is one "lesson" you might want to learn from others; the cost and
risk to learn it from first-hand experience is significant...

Bob Pastusek