Originally Posted by kellykelpie
Apparently if the AOA senser is just .3 of a degree out it can create 3000kg difference.
A decent rule-of-thumb for the lift-curve slope is 0.1 CL per degree, so a 0.3 degree error is worth 0.03 CL. Your clean CLmax is probably about 1.50, so if you're flying at, say, 1.6Vs (a ballpark cruise number) your CL would be about 0.60, at which point the alpha-induced error is about 5% (0.03/0.60) so I'd expect a 5% error in an estimated weight. That would put your weight at about 60 tonnes, which is possibly plausible, so that error seems plausible too.
And to achieve accuracy of more than about 0.2-0.3 degrees takes a lot of effort and probably isn't worth it for various reasons...