In Delve, players and GM participate in a "what would happen?" simulation.  They choose the right moments to simulate in the right amount of detail to satisfy their taste for excitement.  There are specific techniques (mostly using a Pacing Dial) for skipping over boring parts, fast-forwarding through parts where we just care about the gist, and zooming in on the most captivating moments to savor them.

So, that's where I'm coming from with the following observation:

Character Skills for specific tasks speed up or slow down the fiction according to the in-fiction duration of the task.  If your game is moving quickly, and someone decides to try a Sleight of Hand trick, the fiction effectively goes into bullet time as dice are rolled and descriptions offered for an action that takes milliseconds in the fiction.  Conversely, if you are playing through an investigation in methodical detail and someone makes a Swimming check to cross a river, the group has effectively shifted into a fast-forward where a few seconds of real time resolves many minutes in the fiction.

Some Skills demand plenty of interpretation (is this Stealth check for darting from tree A to tree B, or for crossing the entire grove undetected?), but most contain a feature of built-in pacing dictated by their identity as units of "what we are interested in".  Do you pull off the Sleight of Hand trick or don't you?  Do you succeed or fail in Swimming the river?

This issue cropped up in a Delve playtest last Friday.  There were a few moments where the characters were investigating and their best tracker decided to look for footprints.  As the GM, I deemed that this wouldn't be too hard, but would not be a given either, so I said, "Roll for it!"  Ralph, playing the tracker, looked at his character sheet, announced his Tracking rank as "Pro", and roled the dice.  He rolled extremely well.  So I knew that the outcome would be as good as it could be.

But how good could it be?  I hadn't thought about that before requesting the roll.  Ralph's character wanted to find evidence of someone's presence and follow that person as far as possible to gather as much info as possible.  Presumably.  We hadn't discussed that either.

I wound up starting with the immediate result "You find tracks!" and then just kind of letting the "success" result ride ("Yes, you can follow them out of town.  Yes, you can follow them into the swamp.") until it seemed like a new challenge arose (the swamp is messy and might be hard to find tracks in).

I'm now considering a rule that says, "Before you roll the dice, indicate how much of the fiction you intend to resolve via the Pacing Dial."

So what does this proposed rule give you?  I looks like resolving afterward worked fine in play.  If anything, what stood out to me is you might want a stronger rules set for margins of success, but that kicks in after the dice roll.

Setting the stakes before the roll creates a different kind of flow.  For instance, this approach can downplay the margin of success.  If the stakes are already determined, how does a really good roll help anymore than one that's barely over the target number?  There are work arounds for that, but it does become an issue.

If skipping is an inherent part of your system, then declaring how much territory is covered before a roll does make sense.  However, you do need to decide what you want good rolls to do, if anything, as they are no longer extending how territory the check covers.

Dave,

I'm with you. Good rule!

Although sometimes/some players might be fine with just rolling and letting the GM handle it. Like, sometimes, I can roll for Tracking and not be terribly concerned about how it plays out: if the GM has more detail to give me, cool, let's slow down, but if not, I'm happy to play out the whole two-day search in summary mode.

So maybe it shouldn't be a *hard* rule...