Parabola Diagnostics

[pernak18]

After reverting back to fd459233eed913074da10e6784f22fb420ec4588, we're looking at the data going into the quadratic regression a little more closely. Some questions we have:

• [ ] 1. why do some trials have identical dCosts for more than 1 weight scale? e.g., trial 81 of iteration 95 has a -0.6125413177496171 at both the 1 and 2 points (or 0.05 and 0.1, or plus and 2plus)?
• [ ] 2. Why does weight scale not always line up with winner dCost? e.g., trial 29 of iter 95 is initially at a -0.0046732679841312574 -- there is no zero point, so the weight scale is correctly chosen because the initial (unperturbed) dCost is the lower in the init/plus/2plus formulation; however, the dCost of this trial after recompute() is invoked is -0.02511183448149268, which is the dCost associated with 2plus; this could be related to what i was trying to fix in c5f6d702ea59c649624368431d7854e89d6d5a66
• [ ] 3. it looks like i'm assigning 1e6 to some trials...this should happen only for the trials that optimize a given iteration, but there are multiple trials like this -- 110 and 117 of iter 95; these trials are in the same band, so band 11 must have been the winner in iteration 94, but at this point the init dCost should have been recomputed

[pernak18]

1 is because of how i handled negative kmajor values -- i just assign a weight scale of 0 (init) instead of opting out of the calculation like Eli requested. this is not the most efficient route -- i should just not do these calculations -- but i didn't know how to quit the process without crashing the code. the easy workaround here is just to ignore trials where this is happening, since this effectively is what we were doing in the first place

with 2, i'm almost certain this is because of the k-distribution netCDF reassignment with 1plus and 2plus. 2plus is the last to be assigned, so that explains why its value is always the winner. so i need to re-implement some (but not all) of things i was trying to do in c5f6d702ea59c649624368431d7854e89d6d5a66

[pernak18]

should be all set with 93c43bd64434d1b08a95490ab77507e246118fd6. with that commit, dCost points used in the parabola fit, associated trials and bands, weight scale (or zero-point from regression), and final dCost after recomputation with regression roots are printed to CSV files for each iteration