The first problem with the recorded barometric altitudes post upset is that they bear no relation to the recorded vertical speeds. That "zoom climb" for instance takes place while VS is (very) negative.
Secondly, the altitude changes over time for the first "climb" works out to a VS of 95280 fpm or 488 m/s, followed by a descent at -187800 fpm or -955 m/s (M3.15). If the transition was linear, the plane pulled -42G there. And something similar happened another nine times over the next twenty seconds.
That's just not believable, is it? I'm just some random dude on the internet though, so if you don't think my calculations can be trusted either, that's fine.
While the last reported instrumentation readings may not be reliable for directly showing the flight path of the plane, they may still be indirect evidence of it. Assuming (yes, big 'if') that the sensors were still functioning properly, can anyone with knowledge of the plane's systems suggest what gyrations could produce those readings?