Or it might just be the difference between finding the fault on a complete and serviceable airplane and not finding or unable to duplicate the fault on an incomplete and twisted one?
For instance, if something was in error in the last few minutes of the flight, a sequence of system selections unique to the approach phase only, which generated a rogue electrical spike that did not make it to the recorders somehow? Maybe the temperature in the fuel tanks had nothing to do with the cause at all, perhaps humidity, pressure and temperature inside the hull and miles of wiring and electronic units did?
Green-dot