Why is spirit higher than hit?
What Jessamy said explains slight differences due to modeling. I don't remember exactly when the program was updated, but it was generating incorrect hit values, but has since been corrected. It may have been fixed before that report, I don't recall.
The difference between a .01 change in hit/spirit is insignificant to the point of not worth mentioning, but it was causing issues when only modeling positive deltas and thus was remedied.