D&D 5E Table practices for handling skills in 5e?

[el-remmen]

If a roll is warranted, I would base the DC on the overall situation, which includes the goal and approach. I don't base the DC on how "good" or "entertaining" the roleplay is. In this case, I would call for an opposed roll rather than a set DC, something like: "Roll a Charisma(Deception) check. You can roll with advantage since you are wearing a guard outfit. The guard will be rolling a Wisdom(Insight) check. If you succeed, the guard is none the wiser lets you pass. If you fail, the guard is going to think you are full of it and react accordingly."

I am much more likely to use a fail forward approach with this kind of social check, esp. when a roll misses by less than 5 (or something). So in the example above, if the check fails by not too much, I might have the guard be suspicious and call for his boss to get confirmation (giving the PCs a chance to come up with another or follow-up option while they wait) or I might have him ask a follow-up question, "Oh you guys are from the north barracks? How's Wally holding up? That owlbear wound he suffering getting any better?"

That said, while I get why you want to tell the PCs ahead of time the results of failure, I prefer to keep them hidden until they actually happen because 1. it keeps my options open, and 2. I think the possibilities of what happens with a failure are in the players hands to figure out when assessing risk (though I would of course fill in gaps of knowledge that the character would know even if the player doesn't - so if the PCs decide that the worst that would happen is being questioned, I might let them know that they had heard the Baron had put out an order for summary execution for any outsiders found on the property - I am not trying to "get" them).

 

[payn]

I am much more likely to use a fail forward approach with this kind of social check, esp. when a roll misses by less than 5 (or something). So in the example above, if the check fails by not too much, I might have the guard be suspicious and call for his boss to get confirmation (giving the PCs a chance to come up with another or follow-up option while they wait) or I might have him ask a follow-up question, "Oh you guys are from the north barracks? How's Wally holding up? That owlbear wound he suffering getting any better?"

That said, while I get why you want to tell the PCs ahead of time the results of failure, I prefer to keep them hidden until they actually happen because 1. it keeps my options open, and 2. I think the possibilities of what happens with a failure are in the players hands to figure out when assessing risk (though I would of course fill in gaps of knowledge that the character would know even if the player doesn't - so if the PCs decide that the worst that would happen is being questioned, I might let them know that they had heard the Baron had put out an order for summary execution for any outsiders found on the property - I am not trying to "get" them).

That leads into what I mentioned as a negotiation between GM and player. Sometimes a player has an idea but is operating on not enough info. When they realize what eff around and find out really means, they might choose a different course of action. It also allows me to include a possibility that perhaps didn't come to my own mind. Though, I do like a fog of war aspect to these things too. Thats where the player trust amongst each other gets tested.

 

[the Jester]

Whereas I find that gameplay which is predicated on hiding the players' own information from them is generally an ineffective and counterproductive strategy as a GM. It encourages players to distrust the things you tell them (always bad when you are their only source of information) and sets up adversarial behavior from both ends.

I see it as more of a way to enforce "you don't know what your character don't know". It seems effective and productive in my experience.

 

[Swarmkeeper]

I am much more likely to use a fail forward approach with this kind of social check, esp. when a roll misses by less than 5 (or something). So in the example above, if the check fails by not too much, I might have the guard be suspicious and call for his boss to get confirmation (giving the PCs a chance to come up with another or follow-up option while they wait) or I might have him ask a follow-up question, "Oh you guys are from the north barracks? How's Wally holding up? That owlbear wound he suffering getting any better?"

Fail forward, aka success with a cost, is definitely a great tool to employ for many checks. In the case of deceiving the guard, a failure might be that he lets you pass but raises one eyebrow as you walk past. He then alerts others to keep an eye on the PCs. Or... something else. Getting through a locked door via force is a good one to employ success with a cost. On a failed roll, the door smashes open creating a huge racket and you end up splayed on the floor in the next room... or similar.

That said, while I get why you want to tell the PCs ahead of time the results of failure, I prefer to keep them hidden until they actually happen because 1. it keeps my options open, and 2. I think the possibilities of what happens with a failure are in the players hands to figure out when assessing risk (though I would of course fill in gaps of knowledge that the character would know even if the player doesn't - so if the PCs decide that the worst that would happen is being questioned, I might let them know that they had heard the Baron had put out an order for summary execution for any outsiders found on the property - I am not trying to "get" them).

Sometimes the stakes I declare are kept vague (such as in my example above) which keeps my options open. Sometimes I like to tell them straight up what might happen. A capable adventurer, knowing things that the player might not, can decide on a different course of action at that point.

I get the old school approach about challenging the players and telegraphing dangers. And I like it!

 

[EzekielRaiden]

I see it as more of a way to enforce "you don't know what your character don't know". It seems effective and productive in my experience.

I mean, does that not still give the players an active reason to distrust what you, their only source of information, tell them?

It is one thing for a character to lie to them. Characters are (fictional) people who may be trustworthy or untrustworthy. It is quite another thing for the narrator-adjudicator to lie to the players. Because if you believe the narrator-adjudicator will lie to you, how do you know the difference? You literally can't get information except from them. You can't preserve player trust in an environment like that.

I would much rather have players trust that every time I say something, I mean it, rather than being able to very occasionally model a situation where the players make a decision based on incomplete information. Particularly since I can just say openly some variation of "You aren't sure," and achieve essentially the same thing without any temptation to metagame and without any loss of trust.