So a brief digression to consider the impact of perception and perceptual concerns at the gaming table. I might be running a little far afield before I come back to my point. I've been watching Lie to Me on Netflix recently. I like crime procedurals generally-- unfortunately my particular viewing habits mean that I tend to burn out on shows pretty fast-- I mean I'll watch them all, but I won't like them as much as it goes along. I tend to watch things in a bundle-- as I've done with CSI, Numbers, Law and Order(s), and Bones. Good and interesting characters will keep me hooked for a time, but eventually that will wear off. More and more as I watch I'll be conscious of the magic and super-powers inherent in the TV presentation of these things, most of which revolves around either Mentat-level deductive powers and knowledge or technology decades beyond what's presently possible. In the case of Lie to Me, we have a little less of that tech voodoo, but we do have super-powers-- however they don't feel quite as out of left field as the others. Tim Roth does a pretty good job of selling the premise, one which I was initially skeptical of. Essentially he can read people's reactions, scary well.
Which is interesting in rpg terms in that one of the skills/abilities I normally limit is lie detection-- even fallible lie detection as a skill. That's one of those things that has a pretty high chance of eliminating a whole class of stories and tensions. I'd put danger sense, fast regeneration and the ability to track people perfectly as some others that players have to convince me pretty seriously about if I'm going to allow them in the game. Obviously solid and clear lie detection can harm a game. But Lie Detection based on a roll creates an additional set of problems. There are a couple of ways to handle it-- the first being that the player rolls, in which case they know the results. Or there's a resistance roll for the NPC which can be an irritating proposition for the player. Nothing like rolling a success and then having it negated. Especially in a social situation where the player generally makes fewer rolls and hence any negation reduces a larger percentage of their action/play time.
Or you have the classic “GM rolls hidden” version. That's perhaps a little more palatable, but ultimately unsatisfying to the player. It is a blind game for them. Players can't actually act on their ability with any certainty. Mind you, they could play based on whatever the GM tells them, but at that point it becomes more of a narrative device. I don't like either of these choices-- and my approach is instead slightly differently. Generally I like to give players access to social skills which allow me to add information or inflection to their interactions. Usually I've borrowed the skill “Human Perception” from the Cyberpunk rpg. I like the idea of being able to make small reads that can give a player a clue about the situation. I think social interactions can be hard to pull off. I try to give players mechanical fall-backs: the ability to ask for further information, to gain a reinterpretation of the materials, to gain a sense of agendas or even get to replay a scene, depending on their skills. Some players make use of these options, but others avoid them. They seem to react to taking up those options as giving in or admitting defeat. I've found that a little odd in the past, but stubbornness of various kinds can be an especially damning problem at the table-- depending on the flavor of it. Some I can play with and off of while others produce a stalemate or shut-down.
But I'm kind of coming back to the point I started with. Lie to Me, especially if you watch a number of episodes in a row, reveals a basic formula for about half of the episodes. The problem arises, the investigators interview the subject(s) and ask them a question. The subject lie/tells the truth-- and the investigators are assumed to have perfect ability in this (mostly). The determined truth/untruth reveals that the interviewee is to blame/not to blame. Then other evidence arises which contradicts this. Then they suddenly realize that the question they've asked is ambiguous or has other answers. And the cycle begins anew. You find yourself simply sighing when they ask one of these open ended questions-- knowing that's going to be the loophole they will drag the plot through.
Why wouldn't this work in an rpg? Because players tend to ask very closed questions-- or else immediately realize that their question and answer has actually given them nothing. I ascribe some of that to gamers fondness for logical puzzles and years of trying to work out what would be the most ironclad wording for a wish without repercussions.
But there are some things to be gathered from the series-- little details which can be thrown into a game to enhance verisimilitude. And that's honestly what the show is selling-- a kind of story about how behaviors and speech can be read on the human body. They sell it like it is really true-- but in a couple of cases I've seen them use research (like a point about 'gaydar') as if it were solid, certain and with concrete applications which is less than honest. But at the table were selling a version of the truth as well. I make stuff up all the time about history, science and physical phenomena and run the game as if it is true. I'm glad most of my players don't have iPhones and immediate access to Wikipedia.
I want to continue on this track a little further-- tangential, but still connecting, about the question of at what level players and GMs do their hand-waving about elements. We'll be back to that.