Failing to record eclipses doesn't seem very remarkable. There are on average about 1.5 lunar eclipses per year. Clearly people throughout history were not recording every single one. Even the densest parts of your graph are missing most of them.
Edit: also note that this graph is "timed" eclipses only. I take that to mean that it only includes eclipses which were not only observed and recorded, but for which the record includes the local time of day of the event. This is necessary for measuring the Earth's long term change in rotation, as that slide does, but unnecessary for synchronizing calendars.
On the off chance you are familiar with this type of research, do you have any conception for how accurate dates like these are and how they were arrived at? I find it very difficult to find out what the ground truth evidence is for any historical (ie pre-printing press) data.
Pretty much. I start to doubt the importance of what I read when it becomes difficult to figure out where/what/when of the underlying data. At that point it seems ok to wildly speculate and see if any testable predictions come out of it.
I haven't put forth that much effort at this (it's more of an interesting "waiting for something else to finish" type task), but still find it curious I have had so much difficulty following, or even finding, citations to the original sources.
Imagine that applied on a large scale after a revolt or war. The enemy would be so vilified a law would be passed to wipe the time they were in power from history.
Then later generations try to piece together what happened and can't tell whether that "forgotten" dynasty lasted 100 or 200 years, but they do their best and guess 200 years (nope it was 100). Then we base our history off records of that earlier guesswork.