My thought was the fail probability was more like 10 (-9). Oh well, at some point, it was 1/1.
barbiesboyfriend re: sequence/failure. My feeling is that this was not blade fail. The trail of lunch is more likely to be a start/failed disc, with its segments most likely retaining blades (if but for a millisecond). Then again if the disc migrated aft, the blades could have severed clean to start the cascade: debris/fail/debris/FAIL . Sioux City (UAL) was such a failure. The disc (why do I remember it was N1?) parted at an existing crack, and the conspiracists claimed the crack had been identified at a C check, but determined to be "within limits" (sic).
Since the entire midsection of this large powerplant is reported "missing", the shaft failed completely, and it is a wonder and a testimonial to RR that nothing more than a one hundred million dollar event needs to be expensed.
I don't think billet/forging/annealing or existing metallurgical defect is at play here. It seems more likely at this point, given the AD's that addressed the specific location of most likely failure, that it is a maintenance/inspection issue. The lubrication flow problem and required inspections, mitigations fall into an easy line of thought. Who wants to borescope this monster? Who else wants to STRIP it? It doesn't get any more remote than this junction to expose a potential million dollar repair. 400 cycles? that is a very pressurized demand on maintenance. For a New Engine? The implications are falling into place as we post. No one wants to be the bloke with the stick at the Hornet's nest. Given the AD(s), and I'm no expert, the (possible) dire consequences for the fleet are incalculable imo.
Let's capitalize the "Q" as well, it is a
proper noun. 
bear