Bayesian inference. What i expected and what i got after 1 800 000 generations.
The slide command plots split posterior probabilities based on non-overlapping samples of trees taken from an MCMC run. This is equivalent to a non-overlapping "sliding window" analysis where the MCMC run is divided into equal size segments.