Changing stopping rules is perfectly fine in a Bayesian setup. It's the likelihood principle. This is one of the central differences that arises when you cond

Changing stopping rules is perfectly fine in a Bayesian setup. It's the likelihood principle. This is one of the central differences that arises when you condition on the data (Bayesian) rather than the parameter (frequentist).