Walter

@rbouckaert I may be wrong, but according to Joseph's tutorial, the `factor` looks equivalent to `ploidy` in BEAST1. I will go through the code to see whether I could confirm...

OK. I also found the EBSP analyser in both, but they required to log trees. I will try these 2 changes.

I have completely re-designed the tests as Alexei's advised: 1. use HCV data set. 2. use `factor=1` and `ploidy =1`. After reading both code, I can confirm they are the...

Here is Joseph's plot https://bmcecolevol.biomedcentral.com/articles/10.1186/1471-2148-8-289/figures/7

I found the difference between beast 1 plot and beast 2 was caused by the `MixedDistributionLikelihood` only existing in beast 1 xml prior section. ```xml ``` BEAST 2 used ```xml...

Yes. You are correct about the `mixedDistributionLikelihood `. I also double-checked the Java code and XMLs for this and other priors except of `Coalescent`. They look same. So the only...

I ran the another comparison tests with new setting: stepwise, and not using mid points. BEAST 2's plot seemed wrong: [b2.pdf](https://github.com/CompEvol/beast2/files/7275379/b2.pdf) But BEAST 1 looked fine: [b1.pdf](https://github.com/CompEvol/beast2/files/7275381/b1.pdf) XMLs and logs...

@alexeid I found another potential problem as well: after I used the beast 1 analyser to plot beast 2 log as you advised, the plot shape looked same to beast...

I added two new xmls `examples/beast2vs1/beast1/testEBSP-b1.xml` `examples/beast2vs1/testEBSP-b2.xml` to replicate these problems. The previous 2 xmls were renamed to testEBSP-old.xml.

We have the method "parseSetsBlock" starting from line 1008 in NexusParser to parse TAXSET, but it does not take ranges: BEGIN Sets; TAXSET 'con' = 'con_SL_Gert2' 'con_SL_Tran6' 'con_SL_Tran7' 'con_SL_Gert6'; TAXSET...