r8s is far better than BEAST, which is the most easily mis-used program in phylogenetics -- a complete black box like none I have seen in 30 years in the field.  I have seen literally dozens of chronograms that are patently wrong!  A common problem is to include multiple alleles in a population, BEAST  treats this by giving divergence times WITHIN the population of up to 2 years!  When used by real experts, BEAST can do many things.  Unfortunately I share the opinion with many that the experts who created BEAST have persisted in using the program to promulgate the incredibly misleading tenet that substitution rates get slower since time of divergence.  Despite many reasoned responses from Linda Charlesworth, Mike Hickerson and others, proving that this phenomenon is because the "experts" are not taking into account Ancestral Polymorphism (with a single allele, you are adding the age since the MRCA of the population to the age of the lineage.  This causes recent divergences to appear VERY fast -- and VERY MUCH an artifact -- and appear to slow down as the within-population allelic divergence is dwarfed by the divergence between populations.  I attach Hickersons latest and definitive paper.  r8s is intuitively obvious, gives you tools to understand its performance, and etc.