|What are the odds, buddy?|
In an ideal world, conditional difficulty should be part of the system so the DM doesn't feel like s/he's rewording the Old Testament just so that the lock to the crypt can be a little tougher to open than the lock on your sister's diary. It should be easy for the DM to assess difficulty without taking a lot of time, rather than trying to decide whether it's a -22% penalty or a -24% penalty, it should be broad enough that the DM can make a decision quickly and we can all get to the matter at hand: slaughtering orcs and stealing their treasure.
Here's how it goes: the thief (or the Maestro del Dungionni, depending on the circumstances) rolls d20 and a pile of six-siders at the same time. If the 20 sider + thief's level are equal to or greater than the total of the pile of 6'ers: Success! How many six siders in a pile, you ask? Good question: depends on how difficult the DM feels the challenge is. Here's a handy table to use as a guideline:
Table IX.A.2.d(17): Dice Piles and Difficulty of Thieving Abilities
|Avg. Success Rate for a Thief of:|
|Dice Pile||Difficulty Rating||Description||1st Level||5th Level|
|1d6||Super easy||Fail this and you're out of the
|2d6||Easy||Only a newb would screw this up||75%||92%|
|3d6||Modestly difficult||Still, you better succeed if you want to earn your keep||58%||77%|
|4d6||Difficult||We have faith in you!||40%||60%|
|5d6||Pretty Difficult||You might as well give it a try...||24%||43%|
|6d6||Very Difficult||Better have a Plan B just in case||11%||26%|
|7d6||Fat Chance||Give us a minute to back up a ways before you try that||4%***||13%|
Pretty fancy, right?
Some guidelines on how I'd adjudicate this:****
Picking a lock on a typical dungeon door: 3d6... and so forth
Picking a lock to the King's treasure room: 5-7d6, depending on the wealth of the king and his security measures
Sneaking past a guard who's chatting with his buddy about the Misfits show he went to last night: 2d6
Sneaking past a guard who's on high alert: 4d6
Sneaking past a guard who just saw you a minute ago and knows that you're trying to sneak past him: 6d6
Climbing a craggly wall with ample hand/foot holds: 1 or 2d6
Climbing a rough hewn yet slick wall: 3 or 4d6
Climbing a sheer, polished, monolithic wall: you're a thief not Spiderman.
Some of you who are more statistically gifted than I are probably thinking "Why not just tell the dudes they've got a 75% chance to defuse an Easy bomb or 40% chance of deciphering a Difficult language or whatever?" You could do this if you have a good idea of the odds of success for every thief ability under a wide variety of circumstances. I don't have that info handy, nor do I trust myself to make up reasonable odds for such on the fly.
The beauty of the Pile o' Dice system is that as a DM I don't have to think very granularly about the odds of success, I let the dice do the dirty work. Yes, on average a 1st level thief has a ~58% chance of hiding in a 3d6 shadow, fine. But, as DM, I don't actually have to say that. Instead, I'm saying "your odds are somewhere between 20% and 95%, but are most likely somewhere in the middle, so let's just see what happens here..." This is probably just my laziness but I find it much easier to assign a range of probability than a single, precise value.
* The exception here being picking pockets of leveled characters, of course. Which clearly was a reaction to those dickhead players who insist that it is their thief's moral obligation to steal from his colleagues mid-dungeon, yet get indignant when their moronic character finds himself hogtied and naked while his erstwhile colleagues offer his carcass to a band of gnolls in exchange for free passage through their domain.
** Assumes 1 is a fumble, otherwise 100% chance of success
*** 5% if you abide by the 20 always succeeds rule. I know, I'm being inconsistent here.
**** I loathe the word adjudicate as used by gamers--though I don't deny it's usefulness. Still, I hope you appreciate the personal pain I am submitting myself to in order to deliver this post to you.