Nate Pope

There is something that I am not grasping about the dependence between the TMRCAs for non-overlapping pairs of samples in the same genealogy, and how this relates to the binomial...

@jeromekelleher Sure, I can join you all on the 16th. If you need to reach me outside of Github with details on how to join, I'm at nspope at utexas...

Hey @petrelharp, > More generally, I'm wondering what is the interpretation of this estimator as an empirical description of the genealogies? For instance, is it related to the ECDF of...

Thanks @petrelharp, >The numbers x and k are the "raw" input here, and these are of independent interest (as a more fundamental descriptor of the lineages, and the ingredients for...

> I'm inclined to return coalescence rates instead of "effective population sizes"? If people want the latter, they can just invert them? I agree with coalescence rates as a default,...

> how much independence we get depends on how long ago we're talking about (more recent times would need bigger windows). I agree, though, the binomial confidence intervals seem promising...

Thanks for the suggestion @vsbuffalo! Yup, I think we'll ultimately want to do a single pass over nodes. The tip-by-tip approach that I have currently implemented is really just a...

@jeromekelleher @vsbuffalo In my opinion, it'd be useful to include a way to calculate counts of pair coalescence times in tskit, both across time/genomic windows and between defined groups (e.g....

I had one more thought regarding the discussion in #1315 (which prompted this PR) and the question of what fits into tskit (if anything). The way I've been viewing things...

@petrelharp ok, I'll work out a prototype and we can see if it's too redundant with the existing stats machinery. I still think it's possible to avoid multiple rows per...