Please check pseudo stat filters ― some seem broken

Testing has uncovered that the "Adds # to # Damage to Spells" and the "Adds # to # Elemental Damage to Spells" pseudo stat filters do not exhibit expected behavior, and this leads me to believe that perhaps other pseudo filters may be similarly affected.

I have prepared screenshots of 3 carefully constructed trade searches performed in quick succession to demonstrate the existence of erroneous behavior. First, using a weighted sum, I generated a stat group that should theoretically duplicate the behavior described by the "Adds # to # Elemental Damage to Spells" filter. The results appear as expected.
Next, I AND in a second stat group with the "Adds # to # Elemental Damage to Spells" filter itself. Since the two filter groups should behave identically, the AND should prove redundant and we would expect to see the same results. This is not the case, however.
Further, swapping in "Adds # to # Damage to Spells", which should be even more general than "Adds # to # Elemental Damage to Spells", improves nothing.

The pseudo intelligence filter doesn't seem to be working either:

At least in the case of the intelligence filter, it works just fine if it's not in a weighted sum for some reason.

This screenshot also exhibits these problems.