Fudge Stable State Production Statistics To Avoid Thrashing

[Recon419A]

This isn't a bug, but I'd like it as a feature.

When you have a stable science setup for a long time, like in my current 1000x science run, production statistics thrash on the one and ten hour views due to phase differences. That is, at n science per minute, they thrash between n and n +/- 1, due to inherent offsets in the phase (see picture). This behavior is visually annoying, even while being technically correct, because it causes same-volume items to rapidly re-sort dozens of times per second as cyclic data scrolls off the measurement window.

Images: https://1drv.ms/f/c/d11d88eb11ba0bb7/Ig ... s?e=OBxIiX