@flinty That OP wanted multiple samples but used x[[i]] for a single observation when something like x[[i,j]] was necessary. But your suggestion to use LogLikelihood allowed the avoidance of dealing with any of that.
Someone needs to tell SW to try and save a notebook with a couple of high-res images in it. Takes forever and just hangs there. Really needs a progress bar if saving is going to take more than a second