I posit this as the theory. We are playing a game where the play is discussion, problem solving and negotiation. (I'll get to the picture in just a second).
Player skill is not something to be avoided. It is not something I am personally interested in bypassing. My thesis is this.
If something can be handled at the table with player skill, then it should be.
A certain talented someone once said something buried in some comments. I dub it the Zak Sabbath Method of Determining When You Should Roll Dice For Something at The Table! (ZSMoDWYSRDfSaTT!, Pronounced smode-wised-sat!)
In his own words:
"Combat is an especially privileged situation in RPGs largely because it _always_ involves: (1) trying to do something quickly, (2) in competition with someone else, (3) with death (or "failing-to get to play with the PC you patiently levelled up") as the consequence, and (4) successful performance isn't remotely model-able at the table (unlike, say, talking, which can totally be modelled right there at the table via acting).'P.S. Thank you for writing my post for me Zak.
These are the 4 conditions which suggest that an action (combat or otherwise) could not be performed easily by an imagined character--that is, conditions that suggest that -elements (people, monsters, forces) within the fiction- would want to contest the character's success.
They also suggest that the thing occurring has a great deal of narrative tension--character death means you stop playing the game you enjoy playing and have to start again playing a different way.
I mean, the narrative tension in the game may or may not necessarily "care" whether your character climbs a wall on a sunday morning or not, but if you're in combat, someone inside the story being told wants you to fail--always. Otherwise the fighting part of the fiction wouldn't have happened. (yes, marginal examples exist--test fights, whatever)
Nearly every other skill sometimes is and sometimes isn't performed under these 4 conditions.
When they do they're diced.
The more of these 4 qualities -any- action usually has (fast, competitive, life at stake, not modellable at the table), the earlier in the history of RPG evolution you see subsystems describing them being developed."
So, let's just take a second and look at why we are rolling dice.
The first thing we notice is this has less to do with the method of skill resolution or the type of skill system in place and much much more to do with the quality and skill of the players in play. I believe we are universally agreed that the terrible example of hours of a back and forth with the party thief interacting with the dungeon by doing nothing but making endless search checks. You can hear an actual live play example mp3 live play podcast. This is an audio proof example of actual players acting out my white room example of search for hours and endless hours. (Player: I search, DM: Roll, Player:(number), DM: You don't find anything, ad infinitum)
We are rolling dice for the 4 reasons outlined above.
1) Does the event need to be accomplished under time constraints?
For all resolutions that are opposed this will be the case naturally. For ones that are not opposed it will have to be handled on a case by case basis.
2) Are you in conflict with another entity in the game?
The basic stance I have in running a game is that the players can accomplish the tasks they set out to do. I say "yes" or "yes, but" to practically everything they suggest. As soon as someone is in competition with them my ability to do this is suddenly in conflict, since I can't say yes to both players or to the player and NPC (hence there is a conflict that needs to be resolved which is the whole point of conflict resolution systems, i.e. skills.)
3) Is there a serious consequence for failure?
Trying to open a lock with an hour to spare. Trying to tumble past the coffee table. Trying to recall information resting on the open page in the book in front of you. Using diplomacy on your mother to get her to do your laundry. Using search/perception on an empty corridor.
This is a common problem in play, and there are even rules in the game to allow you to bypass these types of pointless checks (take 20).
If there are no stakes, there is no reason to resolve the conflict.
4) Can I model this at the table?
If the answer is yes, then why eliminate game play by rolling dice? Isn't that infinitely less interesting than actually exerting your creative problem solving ability? Are some people bad at thinking on their feet? Are some people not good at coming up with creative solutions outside of the box? Then those people are bad at Dungeons and Dragons. That is fine! There's nothing wrong with playing a game badly, as long as you are enjoying yourself.
[edit: added this after comments]
5) Does the success model a partial result?
This is one of the key things that make a success exciting. You've either partially achieved the result creating interesting situations and choices, or you've chipped away towards your eventual victory. A partial result is more exciting.
What does this all mean?
Diced resolution of actions is only necessary to resolve dramatic conflict with significant consequence for failure that cannot be reasonably (not realistically) resolved by the play of the game (discussion, problem solving, and negotiation).
Dramatic conflict is defined as exterior resistance to the players accomplishing their goals. Significant is defined as consequences that matter to the players at the table. Modelling realism is not a priority, it is more important that the result of the conflict resolution be reasonable and fun (i.e. not destroy our suspension of disbelief and be an enjoyable game mechanic) then to statistically model any specific naturalistic probabilities. The more uncertainty, time constraints and chaotic factors involved, the more likely the event cannot be resolved by the play of the game.
We'll be delving into individual skills as this series continues.