Anomalous variance of MCN1

[msyriac]

Add an option for MCN1 with lower sample variance in :
https://github.com/simonsobs/so-lenspipe/blob/master/solenspipe/bias.py#L98
(This requires sims with new seed choices, but can implement assuming they exist)

[msyriac]

With help from Julien Carron, we have identified that a large contribution to this was due to the QE not being symmetrized. We can close this once symmetrization is implemented.