SMB, thanks for the info. Yeah, the car was used on track pretty much from day one. The factory told us that running in isn't really necessary on modern engines (run in on the bench etc), but to take it easy for the 1st 500 miles or so - which we did. The car is road legal, so we ran in on the road, avoiding high/constant revs etc. Some sources suggest revving the nuts of the engine from day one to avoid polishing the bores, not yet convinced myself. In hindsight, following your advice of a longer run-in period may have been wise... However, it hasn't spent its entire life on track, about 50/50 normal(ish) road driving and track driving. Track driving was only track days, not real racing, so not totally brutal.
Tempering the engine prior to adding the blower does sound plausible, but unfortunately everyone I ask (I have mates in touring car and F3 in the UK) has a different theory.
Having searched the web, there are a few stories of K20A blowing, mostly user error - lack of oil killing the big ends, or missing a cog and buzzing it. One race series in the US has had reports of 'soft' con-rods causing failures (I think that may be rubber chicken's background).
The factory line seems to be that K20A's can only be killed by over-revving, so we must have done. Unsurprizingly, I'm not convinced.
I've now had a propper spanner-man have a look at the wreckage. The bottom end of the engine is absolutely fine - no spun bearings, crank journals all fine, no evidence of heat damage. The con-rod that hit the block is only bent, not snapped. The gudgeon (wrist) pin, is still in one piece (my earlier post was incorrect on this). The piston has completely disintegrated and destroyed/split the bore wall, and mashed all 4 valves. All other pistons/valves and the valve train itself has little or no damage. There is a tiny amount of scoring in the other bores, but this was probably caused by the schrapnel after the failure.
In the view of the mechanic (experienced, but not an engine builder), the piston let go before it hit the valves. Over-revving is still a possible cause, but not the only one? He thinks other possibilities may be detonation causing piston failure, or localized over-heating in pot one causing the piston to seize, and the con-rod to tear itself out. There is so little of the piston and bore wall left that its hard to check for heat damage. May need to send bits away for analysis.
You're on the right line with the type of kit-car, I'd rather not say which at this point to avoid unfairly harming the reputation of the garage with my whinging and conjecture. Will post details once our little dissagreement is resolved.