Ok, I’ll look into this. Might be easier than I expected.
A crude solution would be to split the logfile into 2 files based on the hour of the flip, and analyse them separetely. I haven’t tested that thought.
I agree that the exclusion of AF runs and data points is very useful. But I felt I had lingered enough regarding posting this new version, so hence I posted it. Clearly I need to address this when I have time.