Variation in elapsed time of parallel chains and best practices for computing expectations from multiple chains

[maxbiostat]

@caesoma

alpha ~ uniform(0, 1e9);

Terrible, terrible prior. Maybe something like this?