This is an old revision of the document!
Turns out that each priest has a variance in the number of holy levels they contribute when calling.
Relatively small sample size but is strongly suggestive of holy ± 1 with equal (or almost equal) weightings.
This is more complicated, but not unnecessarily so…