Santiago Claramunt

Thanks for the clarification. It seems that I am specifying the stratification I wanted. However the posterior looks a bit strange. Not only for s but also for d and...

Agree. It would be nice to use the TreePL machinery to estimate under a strict clock model using full likelihood (no penalty term or smoothing parameter). Those likelihood values would...

So, setting a smoothing value very high would be equivalent to estimate under a strict clock but is there a way of obtaining the log-likelihood value (i.e. without the penalty...

The problem persists, even with other datasets and other versions of OS X such as MacOS Ventura with Apple M1 chips.