I wish it were a bit more verbose on the terms and their use. What does he mean by phase "term" and frequency "term"? "Term" -> "error"? And if so why the frequency would set to a relative error of ~25ppm? (Is it measuring the local clock frequency skew, while showing a near zero corrected phase error?)...
I can't tell if "Of-by-one errors" is a clever joke or a typo.
It's both :) ! It's a typo that is built into the clever joke (to help tell the joke)