I've been away for a couple of days and yes I seem to have started something off
As the owner of a very expensive IO540-engined plane I have read everything I can find on engine management generally and shock cooling specifically, and have not found anything other than the same old stories being regurgitated by everyone, with no supporting data whatsoever. Nobody has any data to support a CORRELATION between shock cooling and cracked cylinders - never mind an explanation of what exact mechanism would cause it which accounts for anomalies like the supposed lack of problems caused by shock HEATING on takeoff for example, and the fact that most schools don't get cracked heads despite doing PFLs all the time.
Except one: a year or so ago I read an article by a commercial glider towing pilot, who over 20 years or so has worn out or damaged many engines. He found that they used to get cracked heads regularly, until they modified their operating procedures so that instead of cutting power abruptly after releasing the glider they allowed the CHT to stabilise at its CRUISE value (which doesn't take all that long, a minute or less), and after that they could cut power all the way back (and descend quickly) without any problems.
So, based on this bit of data, it would appear that shock cooling does not happen unless the CHT is above a certain point to start with.
For someone who wants a bit of reading, see
http://www.avweb.com/news/columns/182146-1.html
for some relatively respected stuff on engine management.