Big Pistons, you may read elsewhere in one of these EFATO threads that partial engine failures lead to more deaths that complete failures. Having had one, I concur.
The PA38 I referred to did indeed have a faulty fuel selector, but there is nothing in the AAIB report that says that should have been noticed by the pilot. My understanding is that he moved the selector handle to the correct position, but wear in the mechanism meant the tank was not correctly selected.
One guy died in hospital after the crash - the pilot. His passenger received terrible burns but survived.
Where I would question that pilot's actions prior to T/O is that he changed tanks then took off, so any problem with that tank - fuel, selector, or whatever, wouldn't manifest itself until he was in the air. I used to fly that very aeroplane, and I used to change tanks before the power checks, so plenty of time for subsequent problems to happen while we were still on the ground (they never did, but that was when that aeroplane was much newer).
Whatever, in my book one should be EXPECTING an engine fault at any time in flight, but especially at T/O. This doesn't mean 'sitting on the edge of the seat', merely not being surprised (OK, not being overwhelmingly surprised) when the power dies away at 300 feet.