The extinction probability, E(t), are estimated incorretly

[hoehna]

The extinction probability, E(t), are estimated incorretly

[drabosky]

We have addressed these and related concerns on the project website.
http://bamm-project.org/likelihoodmodel.html
For the first question (rate shifts on extinct branches), the short answer is that this doesn't matter in practice, and there is no way to address this without paleo data. Moreover, all other methods that have been proposed for studying rate shifts make the same assumption, so it's a more general issue.

For the second question (arbitrary swiveling of nodes), we have incorporated a solution into the new version of BAMM (detailed on http://bamm-project.org/likelihoodmodel.html).