Bad estimation of shift and location in a complex likelihood

[bnicenboim]

It also worries me that the model that doesn’t recover the true values has biased posteriors with CrI that are too precise and wrong; see my other general question here.