I've found this discussion/debate about the role of Perception and similar checks quite interesting.
As I understand it, [MENTION=97077]iserith[/MENTION] treats it all as an issue of GM framing - it is the GM's job to establish the scene ("describe the environment" is the term used in the Basic Rules, but I think the GM can reasonably add in other stuff too, even in 5e, eg after a particular bit of action has been resolved the GM might narrate "You've outrun the imperial guards and are back at the castle, panting and sweaty. What's next?")
If the GM wants the scene to include the PC's noticing missing gauntlest, then s/he incorporates this in his/her framing. Otherwise s/he doesn't - but if the players want to mention that they look around the castle for strange stuff that might give clues to whatever-it-is-that-matters, then they're free to do so and the GM might tell them some stuff, or call for a check, as seems appropriate depending on the details:
[sblock]
how you describe the environment is entirely up to you as DM.
<snip>
Again, you describe the environment how you want - you're the DM! You don't need dice for permission on how much detail to give. You could just say that the armor is now missing some element. Or not. If you don't and if the player does not establish that the character is examining the armor, then you needn't provide any additional detail about that armor.
As to the PCs never spotting it <snippage> If I need them to be aware of it for some reason, then I won't gate that information behind an ability check. To do so is to create a problem that one then has to solve by creating another problem - calling for an ability check for an action the player hasn't described. As I've said several times, the DM doesn't need the permission of the dice to describe the environment.
[/sblock]As I understand it, the contrary view is driven mostly by the idea that "in real life" people may or may not notice things, depending on how much sweat is dripping into their eyes after a hard run, and whether or not they're distracted by the chirping of the birds as they walk down the corridor, and the phases of the moon, and myriad other causal factors. And so the GM's narration of the scene should reflect that, which is achieved by the use of Perception checks:
[sblock]
The problem with this Iserith, is that if you play it your way (do not assume players are examining until told), the players always fail to spot the gloves.
<snip>
Unless of course, in the fiction ofthe world, they spot it by accident when moving past. What mechanic exists like that? A Perception check. Or at the very least, a dm examination of passive pereception, maybe giving a different description to a play with a passive score of over 15. Or at least, that's the way I'd do it.
Some of the description is sometimes driven by random chance: that randomness being whether you by accident happen to notice something or not.
if the history check fails the PCs just have to carry on without whatever clues might have been hidden in the Dwarven runes - if any. This is why pre-emptive checks can be useful - sometimes things just get found (or missed) by random chance en route to doing something else unrelated.
[/sblock]There is another reason being suggested for GM-called for/deterined Perception-type checks, by [MENTION=6873517]Jay Verkuilen[/MENTION], which is that they serve a metagame purpose of mixing things up and putting the players on edge:
[sblock]
that's exactly what I use an informational check for, as well as tension building. A failed check often does move the tension up. The players know there were failed checks with potential information missed, which makes them start to wonder what's going on. (Well at least I would hope so, but clearly that would depend on the player.)
I've definitely curbed my own propensity for calling for rolls where there isn't any consequence but in this case or when the player's description is just fluff, but something like the check I outlined has consequences.
I think this often depends on the table. Folks I've played with for many years will often call for checks where there's something that the player seems to be missing and it is possible the character might know something. I'll also call for checks from out of seeming left field to stimulate the player or push them in a different direction than they're looking. For example:
Nattick Nimblefingers' player: "Search for traps."
DM: "Make a History check."
[/sblock]I think iserith's reply to Jay Verkuilen on this particular point perhaps misfires a bit:
[sblock]
there is no need for the History check here. It is superfluous, and arguably any purpose it may serve relies on the kind of "metagaming" you're rolling all those extra dice trying to avoid. Not only did the player not describe an action that would call for such a check, the information can be imparted as I described: There is a discharged trap. It has dwarvish runes on it. What do you do?
[/sblock]Unlike [MENTION=29398]Lanefan[/MENTION] and others, Jay Verkuilen seems to
want the metagame effect of the players knowing that something is afoot. But that said, I tend to sympathise more with [MENTION=97077]iserith[/MENTION] in respect of this particular GMing technique: if I want to put the players on edge I'd normally try and do this via narration than by calling for checks (or rolling dice secretly "behind the screen", which was a popular technique at least back in the late 70s and early 80s).
But on the main issue, about the GM choosing the narration vs "letting the dice decide" so as to simulate the vagaries of "real life": the "let the dice decide" approach makes some sense, I think, in the context of (mega)dungeoneering play. If there is an expectation that the players will play through some bit of dungeon multiple times, gradually trying to map it fully and loot it dry, then having some stuff gated behind random chances to notice it can make sense and be part of the GM's approach to "content revelation". And in a system that measures PCs' perceptive skills, connecting that random chance to those perceptive skills also makes sense.
But as soon as we move even a little bit in the direction of "story driven" play, where the GM has some sort of affirmative responsibility to present the players with situations that are engaging in ways that go beyond simply "there's the dungeon, have at it!", then [MENTION=97077]iserith[/MENTION]'s approach seems more rational to me. And Gygax noticed this back in 1979, in his DMG (the passage is from p 110):
t is your right to control the dice at any time and to roll dice for the players. You might wish . . . to give them an edge in finding a particular clue, eg a secret door that leads to a complex of monsters and treasures that will be especially entertaining.
Consistently with how I understand iserith's posts above, I want to say: if the complex of monsters and treasures will be especially entertaining, then why gate it behind a random chance of having fun? Just tell the players that their PCs notice the secret door! (If the GM wants to achieve this by providing the information to the player whose PC has the best Perception score, as a tip-of-the-hat to that PC's build, then that seems harmless enough, but very much a secondary consideration.)
I should add that this idea of the GM wishing that the players have some information, and so giving it to them, is different from the idea of the players needing the clue so the game can progress. The latter idea is what motivates the GUMSHOE approach of the GM narrating the (basic) clues without calling for checks; but the idea that the GM willl just provide, via narration, information that s/he thinks is interesting for the players to know needn't be connected to the idea that there is something the players must do with it, or that the information is a clue of some sort.
As soon as a DM starts paying attention to "what's good for the story" she's put herself on a slippery slope towards heavy railroading and-or plot-protecting the PCs, neither of which are all that desireable.
And apropos of the immediately above, this is just nonsense. Gygax deciding that the it would be fun for the players to know how to get to the "especially entertaining" bit of the dungeon isn't railroading anyone into anything. Me starting my last Prince Valiant session by establishing a scene that was salient for both knight PCs, the squire PC and the travelling entertainer PC wasn't railroading anyone into anything.
Nor does framing scenes with an eye to them being exciting and engaging rather than not have anything to do with "plot protection". (Which may or may not be a bad thing, although it's weird for someone who plays only D&D, which has at the heart of its resolution system the most famous plot-protection mechanic of all time - hit points - to say that it's undesirable.)