Recreating Normal Distrbution using logp - sampling diverges

There’s no observed argument in model1, is that a typo? When I add it everything runs fine.

In general, it can be more difficult to sample from priors using MCMC than from posteriors. See this discussion for details.

1 Like