In this crash, the entire empenage was separated from the airplane near the initial impact site. The vertical stabiler and both horizontal stabs can be seen near runway centerline less than 1000' down the runway. The photos of the airplane at rest provide a clear view of the pressure bulkhead as the remaining structure at the rear of the plane. From the way the tail was ripped off, it seemed to be at a high body angle at initial contact. I'd guess at least a late go-around was initiated, posibly a late go-around without sufficient power applied with very high AOA and subsequent tail strike. The only other thing that could produce such a high body angle would be less-than landing flaps selected, but that should produce considerable automatic aural warning from the airplane. Less-than landing flaps would also produce a different pilots visual reference picture.
This accident appears at first glance to be a near repeat of the Korean Airlines crash at Guam. An inop ground station ILS which was not accounted for by the pilots. In Guam, the pilots flew a perfectly good airplane into the ground as a result.
Korean Airlines (not Asiana) had a notorious history of co-pilots refusing to countermand the Captain even when he was wrong. This was causal in the Guam crash, in a 747 freighter crash in London, and several others. I suspect those cultural influences exist at Asiana airlines also. The investigation will surely tell.
P.S. What’s the decision height at SFO and why didn’t the pilot initiate a missed approach?
Seems unlikely, even unthinkable -- I'm sure it was also on the ATIS. But that's what happened at Agana though it was night IMC).
Another possibility -- the zero fuel weight in the FMS was wrong (too light) and the FCC-generated approach speeds were off, biased toward the low side. When they were close to the flare the airplane developed a sink with no stick shaker (also based on FCC-generated speeds). But the long flight with step-climbs seems to negate that possibility.
I think these engines were Pratts, not RR's, so the previous 777 crash-landing is probably not related.
The water's edge is 600 feet short of the threshold. If they were supposed to be at 50 feet over the threshold, then fifty plus at least 31 feet (sine of 3 degrees times 600) equals at least 81 feet low. In day VMC that would be very obvious.
For us non-pilots, when you say "An inop ground station ILS which was not accounted for by the pilots.", I understand that you mean that the instrument landing system was inoperative for the two runways. What I don't understand is why that would not be glaringly obvious to the pilot as he made his approach, and how it might contribute to the crash.
Initial reports said the pilots were reassigned to the left runway while on final. If so, then my guess is they moved over but that lost them 50-75 feet they failed to correct for by adding additional power; turns, horizontal lift, you understand that point perfectly I am sure. The sight picture changed but the pilot took the risk instead of adding power or executing a go-around. As you said, the fuselage looked like a high AOA impact. The passengers in the back no doubt took a massive G-force hit.