[Someone]

I don't think the 'not enough energy' argument is particularly strong. To flip a plane over, you have to lift up the plane's center of gravity by about half its length (probably less, as the nose can go down a bit during the flip. Also, if the pivot point is below the fuselage, the engines will provide torque during flipping.

Let's take a 60m long body and assume perfect conversion of horizontal speed (=kinetic energy) into vertical speed (=potential energy). To lift the body by an average 30m, it would need a speed of sqrt(2gh) = sqrt(600) or less than 25m/s. That's less than 90 km/h.

Google gives me a landing speed of about 250 km/h for a 777. That gives us a factor of 3 in speed, or 7 to 8 in kinetic energy to cater for inelastic collisions, air resistance, energy loss digging the nose wheel or the whole nose into the ground to create a pivot point, etc.

Also, and probably more convincing, there's the example from a DC-10 crash in Sioux City in 1989. Both http://www.airdisaster.com/reports/ntsb/AAR90-06.pdf and http://aviation-safety.net/database/record.php?id=19890719-1 mention the word 'cartwheel', and that's also what I remember from the video (which I can't find now)

[momavujisic]

I'm not going to question you, because the plane did indeed have enough energy to the "flip". I'm just going to suggest things about the speed. The approach speed a 777 at 250km/h, is correct, that's 134 kts which is maybe a tad bit slower then what it should be. However, the NTSB confirmed today that the aircraft was significantly slower than that – which why the plane ended up coming up short and stalling. Furthermore there is at least a couple seconds of the plane skidding on the ground which would have slowed it yet even more before it did the "flip".

[Someone]

OK. Let me question myself, then :-) . Energy-wise, I still think it is easy to flip a plane in a crash. I would only call it cartwheeling if there were two flips in succession. That would be tight, as the impact on the ground after the first flip would be quite inelastic.

Also, for both flips, you would have to have the mass of the plane straight behind the pivot point. That is hard to achieve, as everybody who has done cartwheels will know (if not, imagine maiign a handstand on one hand with too much rotation that continues into a landing on one leg)

[mikeash]

The Sioux City DC-10 did not cartwheel. I believe a wing detached which is why people thought it did.