Thanks on the shading effect.
The usual story with series is with no shading, you can calculate that each panel will do the same amps (same watts/panel), but that was not so with mine as described in the OP.
I am thinking that makes parallel the better choice, so you can just add the individual amps and get the most for a total.
In previous set-ups, I was unable to realize any measurable advantage you read about where higher voltage does better in low light overcast so there is still some voltage above battery voltage. So IMO that would not counter the loss in series where it uses the weakest panel's amps. ( I know you would have to do a bunch of proper comparison testing to prove that, but I have done some, and that's what I found. )
