Ok, here's my interest rate visualization.

The horizontal axis is linear time, so you can see how the updates have been spaced out over time. The years are marked off, because we commonly think about what happened in 2015 or 2016 etc. Each line represents one subgrade. I've made all the "1" subgrades red, all the "2" subgrades blue, and so forth. The dots on the bottom red line show the dates of the changes. The lines in between updates are diagonal, but of course the interest rates actually stayed constant until the date of the next change, at which time they jumped.
Obvously the changes are not evenly spaced in time. There appears to be no policy about when to change rates. Late 2015 and early 2016 saw a rapid burst of changes. At first I thought "oh, that's when they noticed there was a problem", but that's not quite right. Look closely at those changes in late 2015. They're small little wiggles. In retrospect I have no idea why they started these rapid changes. In 2016 it is obvious that they realized there was a problem.
As lascott mentioned, the thing that happened to B grades at Oct 2016 change is nutty. Almost looks like a mistake that they corrected in the most recent change.
Other than that change, and the little shift up in E4, the tail end of 2016 and the first part of 2017 have essentially no changes. This would seem to imply that the LC number crunchers think they're doin' ok. At the present time, investors are not reaching the same conclusion. LC doesn't share the data that drives them to these changes, so we're left scratchin' our heads. Will take a few quarters to see whether they got it right.