Granted, I don't mind thieves simply being really good at ability checks, but, yeah, do something special with them.
They shouldn't be just really good
they should be significantly better
. Anyone can "hide". Anything
can be hidden for that matter. I'm hiding behind a keyboard right now! I didn't spend months or days studying camouflage, the nuances of shadow, movement and sound like a ninja would have. Significant difference that can be quantified as Skill
vs my natural or circumstantial Ability
. A Thief would spend days upon days practicing lock picking, the same amount of hours that a Wizard would spend studying ancient text, collecting and organizing material components and memorizing the profane language of ire-some gods to cast a spell, and the time that a Warrior would spend honing his sword, practicing his footwork and taking steroids and "working out". The only way I like quantifying these differences in game terms (without getting into playable vs realistic, fun vs simulation, etc.) is totally and arbitrarily letting only
those experts do certain things, not just do them better. Certain
things being the keyword. Using hiding as an example again, a wizard on his tip-toes is one thing, but a Thief having special shoes, laced a certain way, knowing which fabrics to wear, and rolling the edge of his feet slowly, brushing any objects out the way before committing his body weight to the balls of his feet, and breathing matching the breeze as he's practiced over and over is something only a Thief can do.
Unless you want to go *really* Old School and argue that the thief class doesn't even need to exist...
Need is a subjective term!
Ok, here's my argument for keeping %. In AD&D an uber high level Thief has a 99% chance of success for most of his special abilities. More or less. And without taking any other modifiers into consideration.
Now, should uber high level Thieves be so awesome at what they do, that they only fail 1 time in 100? Works for me! I don't care if percentile dice are used or not, but it's that level of efficacy, which I think should be retained. Which is why I don't like a lot of the old school versions of the Thief, which have been created over the past couple of years. If a Thief is gonna fail one time out of every six, or even one time out of every twenty, he's just really not all that impressive, is he?