If we're talking about the speed with which something exciting happens, then the thief pick-pocketing the king has just mainlined that process. No matter how that action resolution turns out, something exciting is about to happen:
- You're going to get something valuable from the king.
- You're going to get involved in an epic chase through the palace.
- You're going to be arrested and thrown in prison and need to concoct your escape.
- You're going to be captured and sentenced to clear out Norworld and make it habitable for civilized folks.
Or, you are going to die, which is by far the more likely outcome. Trying to steal something from the king, even so much as touching the King's person without his prior consent, isn't merely 'pickpocketing'. It's High Treason, to be punished by drawing and quartering on the morrow, and the body peices being hung in an enchanted gibbet so that no resurrection is possible. This is probably the resolution I'd steer for, because it involves far less pain to the play group as a whole.
But let's suppose that an epic chase does happen. The rest of the party must either choose to side with the pickpocket, which is quite possibly suicidal, or else sit back and watch the rest of the session be monopolized by the thief while they sit on their hands. And this is going to be especially painful if the whole reason that they were here in the first place was to win the trust of the king. That's not going to be very 'exciting'. Of course, technically they can hunt down and kill the pickpocket (I would, or would want to), but the pick pocket's player is likely to see this as 'betrayal' given the personality we are probably dealing with here.
So let's assume that the result is actually capture and not death. Well, unless this king is an actual booby, he's going to take basic precautions that will make his prisons all but escape proof. It might take 20 years to concoct and implement an escape plan from a reasonably well designed prison. In the mean time, what are the rest of the players supposed to do while the campaign has been effectively permenently forked? Is this going to be 'exciting'?
I think the pre-determination that any of this is less interesting or slower paced than "have a chat with the king where he offers you a contract to go clear out Norworld" is to make a rather large presumption that the GM's predetermined plot is the only interesting outcome.
What I see in your reply is the pre-determination that no matter what happens, you are going to treat the player with kid gloves and reward their choices. If I know that nothing can happen to me as result of trying to pick the king's pockets except that I end up with the same predetermined mission to 'clear out Norworld' that I would have ended up with anyway, I think I'd be quite free with my sleight of hand skill. I'm not saying that you can't sometimes make a silk purse out of a sow's ear, but sometimes rigging the game such that the PC doesn't get his logical reward violates suspension of disbelief and other necessary factors in gaming to the point that your game is so cheapened as to be not worth playing in. Sometimes when the player sticks his hand in the fire it not only needs to get burned but stay burned because that's the nature of serious stories. Serious stories aren't about zero sum games with no serious consequences.
I tend to play my games straight up with no dice fudging or narrative fudging. I feel that to do anything else risks disempowering and deprotagonizing the PC's. The consequence of that is that for the most part, the PC's don't get treated with kids gloves and earn their failures as fairly as their successes.
However, once in my current campaign I actually cheated. I had a player successfully use the Intimidate skill to bully someone into doing something. I ignored his play and proposition completely, because I knew the player had no idea what he was doing and that if I didn't ignore it would absolute wreck the entire campaign for everyone. You see, the player - stretching his 'first time roleplaying' muscles - decided it would be fun to bully a stranger that he met on the road into giving him his horse. This would have constituted banditry in the eyes of the law, which is punishable under the law by death by exposure. The offended NPC, once the player was out of sight, would use his newly hostile attitude to the player - and by extension his comrades on the road - to go to the local watch, who I happened to know was run by a quite compotent watch commander and report the fact that he'd been robbed by armed bandit's on the Prince's road and swear out a complaint versus the PC party. The stranger, who was in fact a highly respected and wealthy citizen, would have been readily believed and the distinctive appearances of the PC's would have made them unmistakable and unable to hide. As low level characters, they would have stood now chance of resisting arrest. All these facts where part of my preexisting 'sandbox' as it were. They were setting facts, events, and background.
Worse yet, before the arrest could take place, events that the characters hadn't yet learned about would transpire which would have resulted in the murder of the very stranger who they'd bullied. This would have resulted in the watch commander making the quite understandable assumption that the PC's murdered the man, and lacking any other suspects they would have been bundled together in these charges. The result therefore of the one players ill-advised action was almost certainly going to be a messy TPK against overwhelming odds and the pointless death of every character in the party. You better believe that this occuring in my campaign's second session was not something I would have enjoyed, so I simply by fiat treated the event as if it had not happened. I had absolutely no in game reason or justification for doing so. Now, maybe that is 'cheating', but its no bloody less cheating than allowing a player to rob a king in public and it being treated as a minor incident not impacting the dignity of the king or his ability to rule.
Sometimes the DM needs to stand back and let the player's play their characters without worrying to much about getting what he wants. That's my default stance as a GM. But every once in a while what I want really is better for the group than what one player wants.