I’m using the magic method by creating a rough mesh over the entire model. This same mesh is used to tally flux and for the weight window generator. Plotting the lower and upper bounds of the weight windows seems to indicate more importance placed in the central region where most of the neutrons come from. It is also asymmetric for some reason, as seen in the image.
I ran the simulation with 1e5 particles for 20 batches but am still getting this strange distribution in my weight windows. As I understand it, the weight windows should take the results of the flux tally over the whole simulation and give greater importance to regions where there is higher error or standard deviation in the flux (like in the magnet region circled in red). However, the weight window generator does not appear to be doing that.
Any advice on how to get weight windows that prioritize the magnet region? I suppose I could try manually assigning the bounds of each weight window instead of using the magic method.