Because mass peak drifts up as a function of run number for (viz.) small cells, I've decided to implement a runset-depenent mass cut for 6 energy bins. I am currently using a skewed-gaussian to fit the mass distribution, and cutting where the fit falls to 1/5 the peak height to the left and right of the peak.

ENERGY BINS
============
bin 0 -- 30-40
bin 1 -- 40-50
bin 2 -- 50-60
bin 3 -- 60-70
bin 4 -- 70-85
bin 5 -- 85-100

For each 2x2 group of plots, the top plots show the fit peak position (MassM) vs set number,
the bottom two plots show the mass cut width (MassH-MassL) vs. set number.
Left 2 plots are for large cells, right two plots are for small cells.

Each set of 4 plots is for one energy bin, starting from the lowest energy

Set number is defined as 10*(day_of_year)+(set), where a single set contains 5-10 runs

This is run 12 pp510 longitudinal