From a different perspective, I haven't been happy with powers that let you trick or force enemies to attack each other. They're too situational, and even worse, the designers don't seem to have fully grasped the math involved in needing to successfully attack one monster in order to give it the chance to successfully attack another monster, at which point you finally deal damage.
For example, Turathi Highborn, selected by one of my players because it really fits her character, has "Bolts of Bedevilment." It says Bolts plural, but it only attacks one enemy. If you hit, that enemy makes a basic attack as a free action against another enemy. The enemy gets a +2 on the attack roll, and deals an extra 2d6 fire damage.
That may seem nice at first, but think about it. First you've got to have an enemy eligible to attack an ally without moving, but lets ignore that. Lets say you have a 60% chance of hitting with the bolts. Then the enemy has to attack its ally, which means maybe a 60% chance of hitting again due to the bonus you've granted it. Then it does regular damage +2d6.
Well, you're already looking at a 36% chance of actually dealing damage. And then the damage you deal is about equal to a regular level 11 encounter power, maybe.
It seems like the coolness of controlling an enemy caused the designers to overvalue the power. Even in optimal situations (forcing a solo to attack another enemy, for example) you're still not looking at good damage simply because of the terrible, terrible chance of hitting.
My group's solution, by acclimation, is to let Bolts of Bedevilment attack two enemies. I've felt like a lot of powers of this type seem underpowered, but this is the only one our group uses, so its the only change I've made.
