I don't think there's a simple explanation for this one using arrows and relative airflow vectors - I'd tend to fall back on the idea that the rotor disc as a whole is, as you indicate, made up of a number of flexing, flapping, twisting objects whirling at great speed through a fluid, and is furthermore asymmetric in its travel through that fluid (ie advancing and retreating blades), so 'it just does' would be a fair enough comment!
Maybe a better way to look at it is that, because of this asymmetry, it would probably be strange to find that the magnitude of the torque spike or drop was equal for rolls either way. It's not a see-saw that goes up and down the same in both directions, so it doesn't make sense to assume things should be equal.
Last edited by Arm out the window; 21st October 2012 at 22:46.