How to beat the "20 always succeeds" rule

[Jeff Wilder]

Seriously, Glitterdust makes you blind, which is essentially the same as making everyone else invisible. That gives everyone 50% miss chance. In order for Unluck to give a 50% miss chance, the attacker has to be able to hit on a 11+ on their attack roll.

Not exactly.

If a creature hits on 11+, and has to take the worse of two d20 rolls, the chance of the creature hitting is 25 percent, not 50 percent. (Each roll has a 50 percent chance to be 11+, but both rolls must be 11+: thus 0.5 times 0.5 equals 0.25.)

50 percent would be hitting on 7+. (Each roll has a 70 percent chance to be 7+, but both rolls must be 7+: thus 0.7 times 0.7 equals 0.49.)

IMO, unluck is a pretty decent spell. (In addition to significantly underestimating the probability effect of the spell, you may have also overlooked that it affects damage rolls ... including for spells. It even affects, e.g., the "1d3" roll when summoning a lower-tier creature with a summoning spell.)

It's not a ridiculously over-powered spell like glitterdust, but in a group with multiple spell-casters it gets pretty close.

 

[Wyvernhand]

As I said, Unluck, as it functions, would be much better as either a swift action (functioning similar to Assay Spell Resistance, for saves), or if it didn't allow a save (like Curse of Impending Blades), or heck, even if it affected multiple foes like Slow does. Compared to Slow, Unluck is strictly worse in nearly every case, to the level that you'd be better prepared for 99%+ of adventuring days if you prepped 2 Slows instead of 1 Slow + 1 Unluck. If I'm a wizard and I'm casting a 3rd+ level spell, I want more bang for my buck and Unluck just doesn't deliver compared to the other tools on the market.

That is, unless you consider spells like Glitterdust to be too good. Then I'd suggest filling up all of your slots with Heightened Flares. Glitterdust is good, but its far from overpowered.