Rob J Goedman

Of interest on LOO etc might be http://www.jmlr.org/papers/volume11/watanabe10a/watanabe10a.pdf .

Hi Josh, Thanks for the feedback! Having a `digits=...` argument is a good suggestion. On your 2nd question, the approach taken in MCMCChains is slightly different in that you can...

For `psisloo()` and `waic()` wouldn't you need the `pointwise_loglikelihoods(model, c)` matrix? Not sure for wdic. I really like the name `ModelSelection.jl`, maybe in JuliaStats?

David, is your thinking to have something like an MCMCModel

The current version of StatsModelComparisons.jl uses StanSample.jl for testing purposes. Wouldn't that be an ongoing source of issues if StatsModelComparisons.jl is made part of testing MCMCChains.jl. In the past there...

Just to make sure I understand this right. For the import statement in MCMCChains.jl, StatsModelsComparisons.jl needs to be in the dependencies of MCMCChains right? In StatsModelsComparisons.jl I can turn all...

Very nice changes David! Thanks a lot. I'll remove the mc_path. At some point I ran in troubles with making it a constant but that was a few years ago....

Yes, this recent change to only auto-merge packages with OSI-approved licenses was a surprise to me as well. A problem is that TuringLang is a product while StatisticalRethinkingJulia is just...

To the best of my knowledge this has not been done. From my point of view, I would prefer this, at least initially, to be done in a separate package...

Hi Chris (@ChrisRackauckas), There have been a few attempts over the years but none successful. Two of the more recent threads are [Stan forum](https://discourse.mc-stan.org/t/installing-cmdstan-using-conda/25536/10?u=rob_j_goedman) and [Julia forum](https://discourse.julialang.org/t/adding-a-build-step-for-cmdstan-inside-a-julia-package/51844?u=goedman). The Conda route...