‘batting_scores’: -inf
I had a similar issue recently and turned out it was because I had to downsample my observed data which caused some values to go out of the model’s boundaries. Any unexpected values in your observations?
1 Like