Jerome Kelleher

Sounds about as good as that sort of thing gets with real data to me - if it was any closer I would doubt its value

Nice!

Just to note here that we're supposed to be working with things that **arent'** tsinfer too, so we need to be very careful about assumptions we make about metadata.

I think code snippets like this as implicitly assume to be "do whatever you want", we probably don't need to worry about assigning a formal license. The overall repo license...

I think we can just sentence saying "All code in these tutorials is free to use for whatever purpose (CC-BY)" or something. I think that's the de-facto anyway, I doubt...

Reminder to myself: I wrote a presentation about this for the PopSim workshop which could be adapted for a first pass on this. Some good examples in there which would...

Here it is: [arg-ts-presentation.pdf](https://github.com/tskit-dev/tutorials/files/6001602/arg-ts-presentation.pdf) This was the presentation I gave at the PopSim consortium meeting in Aussois, Oct 2019, so some of the information will be a little out of...

> Hello, just adding a suggestion as discussed with @hyanwong . It would be handy to see how to extract the number of recombination events from the ARG representation, and...

Hmmm. So, one way we can get at topology changes is to look at the tree [ranks](https://tskit.dev/tskit/docs/stable/combinatorics.html) in the [simplified](https://tskit.dev/tskit/docs/stable/python-api.html#tskit.TreeSequence.simplify) tree sequence ```python sts = ts.simplify() for tree in sts.trees():...

> If one of the recombination events was completely "invisible" (no branch length or topology change), then am I right in thinking we would see just one tree instead of...