Empower Spell + Summon Monster equals...

[Greenfield]

What happens when you apply the Empower Spell feat to a Summon Monster x spell?

At first glance, it looks like it wouldn't do much, if anything. Yeah, there a dice roll if you use Summon Monster X to call for a number of monsters from the Summon Monster X-1 table, but that's not usually your A game. After 1st level Summoned monsters are about half the CR of the caster, which often makes them little more than speed bumps in an actual battle.

What I'm wondering is whether it applies to the implied random numbers inherent in the monster itself? As in, ability scores and hit dice.

Monster ability scores, straight out of the book, are based on a standard array to which racial modifiers are applied. That is, 10, 11, 10, 11, 10, 11. These are the average of a 3 D6 being rolled. (The average is 10.5, hence the split).

In theory though a DM running a monster from the book could use a different base, such as the Elite array, or even (gasp) roll the actual dice.

If we read that "standard array" as the result of an average dice roll then Empower spell would add 5 to every ability score.

Most monsters use a D8 for hit points, and the hit points listed in the MM reflect an average roll of 4.5 per dice, plus CON modifier.

Applying Empower Spell to that dice roll would grant +2 per hit dice (+3 for D12 monsters). On top of the CON boost that would mean +4 per dice, more if it started with an odd number in the CON score. Considering the 4.5 base average roll, that means that we'd nearly double the hit points.

+2 AC from Dex increase, with a matching Init increase, +2 to hit and damage from the Strength boost, +2 to all Saves and Skill checks.

Suddenly summoned monsters get notably kickass.

Opinions? Should it work this way? Would you, as a DM, accept this as a valid interpretation?

How about Maximize Spell being applied?

 

[Celebrim]

What happens when you apply the Empower Spell feat to a Summon Monster x spell?

It does what it intuitively does, allowing you to summon small crowds of low powered creatures. I have a homebrew feat in my game called 'Summon the Legion' that automatically empowers (for free) any Summon Monster spell. As you note, since low CR creatures are generally only of small utility, I judge this feat useful but not broken, as it makes summoning 'Wall of Wolves' (as the players jokingly refer to it) more of a thing worth doing and you only get a couple extra low CR monsters.

In standard RAW, it's basically NEVER worth empowering a summon monster spell, as you'd always be much better off casting the version two levels higher. Summon Monster V, using the Summon Monster III list, is vastly better than getting a few extra creatures from the Summon Monster I list. And indeed, since the creatures you get are random, without something like my 'Summon the Legion' homebrew, why would you ever risk getting 1d3 creatures from a lesser list (ending up with perhaps 1 weak creature) rather than summoning 1 strong creature.

What I'm wondering is whether it applies to the implied random numbers inherent in the monster itself? As in, ability scores and hit dice.

No, I would not accept this as valid interpretation. Summon monsters are always normal stereotypical examples of their species. They never have unusual stats. You never roll for these stats. (I personally never even roll for hit points on summoned monsters.) Nothing in the rules implies that a monsters standard array is actually the average of generated by 3d6 rolled randomly. That's a demographic interpretation not born out from how D&D normally played. It is by no means certain that NPCs generate stats at all, much less that they do so randomly. And unless the random stat generation was explicitly mentioned in the spell-text, Empower Spell wouldn't apply to it anyway.