Dealing with missing data and custom distribution

When all values are present it runs fine. If I remove the “charlie” row from the dataframe, everything is good, no errors.

With missing data:

User        -6.05
Task        -4.04
d_missing    0.00
ll          -4.16
Name: Log-probability of test_point, dtype: float64

Without missing data:

User   -4.04
Task   -4.04
ll     -2.77
Name: Log-probability of test_point, dtype: float64