Quote Originally Posted by Kurald Galain View Post
That's what I want them to do, yes.

WOTC has either not done the math yet, or they think that it would be a positive trait for their system if the dumb barbarian can routinely make knowledge checks that the smart wizard fails at, or said wizard to punch down doors that the barbarian can't. There's only a +7 difference between their skill mods, after all.
-10 take half doesn't actually have anything to do with the disparity of a wizard and a barbarian. If they wanted to make a high-strength barbarian that much stronger than a low-strength wizard, then they would simply allow the barbarian to get a higher strength score. Giving the barbarian a strength score of 36 and the wizard a score of 6 would be the same as making every point count for 1.

The real issue here is the amount of randomness being applied to the numbers. If you wanted there to be less of a range of how good or bad a character can do due to random chance, then you need to use a less random replacement of the d20 roll. Use 2d10, 3d6, 1d10+5, 1d6+7, etc etc.

However, if you start to narrow the range of possible results like that, then you run into our old friend of stat inflation again. If it's really impossible for the wizard to out-armwrestle a wizard, then you're likely also going to make it impossible for a trained fighter to miss a goblin, or for a dumb barbarian to resist a clever wizard's magic. Everything becomes certain, and thus dull.

A wizard being able to out arm-wrestle the barbarian is a feature, not a flaw. It's essentially the whole basis of DnD. Does it cause things you don't like sometimes? Yes, but that's the whole point of it being random. Sometimes it works in your favor, sometimes it doesn't.