Only roll at the precise time when the outcome of the roll matters- for everything!

[Mastikator]

Before I share the idea I've been mulling over I'd like to share a scenario that has happened many times at my tables (in different groups, in different systems, on different planets even!). Those who are not interested can skip to the TL;DR.
This scenario happened in a game I was DMing, 5th edition D&D.
A player- Bob, wants to hide a specific item, they are in a fortress in the barracks and don't want to bring said item on their next adventure. So the player asks: "What do I roll to hide my stuff?"
I, the DM responds: "Make an wisdom (stealth) roll" (don't question my choice of ability score and proficiency, if you disagree with it just pretend I picked whatever you would've preferred. Just pretend!)
The player rolls the dice: "I got a 3 on the D20, my wisdom is +2 and I don't have stealth proficiency, so uh... 5?"
Then another player- Bill, pops in, tempted by the lure of metagaming and says: "Can I check that he hid it well? My passive perception is 12, do I find it?"
Me: "Yeah you find it easily"
Bill: "Hey Bob mind if I help you hide the stash of items? BTW I rolled a 17 for a total of 19"
Me: "Yeah it's much better hidden now"
Then they fly off on their hippogriffs and continue their adventure.

FAQ:
What did I do wrong here? I made them roll immediately even when there was no consequence for rolling poorly. Which allowed the players to invalidate the dice until they got a result they liked.
Couldn't the players have just paid to have the stash locked away by the quartermaster or whatever? Yes, absolutely. But the players were swimming in money and not willing to let go of a single coin.
Couldn't I have just told the players not to metagame? Yes. And I should have. I chickened out. And I am prepared to receive my 10 lashes on the public square for my dishonorable weakness.

TL;DR
The idea that I've been mulling over is to not let players roll the dice until the very last second.
If a player wants to sneak around I don't let them roll for stealth when they begin, but immediately when they can be detected by an enemy.
If a player wants to search for traps on a door, I don't have them roll for investigation when they begin, I have them roll for investigation when someone opens the door and potentially triggers the trap (if they succeed then the player who found the trap stops them at the last second so as to not trigger the trap).

This is already how most games already work, you want to attack someone, roll for attack and immediately hit or miss. Want to dodge a grenade, roll immediately to dodge. So it's not a weird really, but it's much much harder to invalidate a bad dice roll by metagaming.
Now I'm not talking about characters who have special class, race, perks, feats features that let them reroll or add bonuses after the fact, those typically cost resources and are fine.

How would the above scenario worked out if I had applied this rule?
Bob asks: "What do I roll to hide my stuff?"
I, the DM respond: "Nothing, not yet. I'll tell you when to roll"
Bill- still not satisfied: "Can I help him?"
Me: "Yes, you and Bob spend the morning hiding your stash. Bob will have advantage when he rolls"
[... the game continues ...]
As the players depart, I say: "Make an wisdom (stealth) roll with advantage because Bob helped you"
Bob: "Ok I rolled a... 1 and a 1. For a total of 3, LMAO" (or maybe he rolled a 1 and a 17 for a total of 19, exactly the same outcome, who knows what these tricky dice get up to. But this way they didn't metagame.)

-
Edited to make it slightly easier to read

[Lord Torath]

Makes sense to me. Heck, for that particular situation, I wouldn't have the player roll until someone was actually looking for the hidden item. If no one looks for it, no roll at all.

[JNAProductions]

I don't think your idea is bad, but I also don't think it's needed.

For instance-Bob rolls poorly to hide his stuff. Bill is reasonably Perceptive, and easily spots it-the party, therefore, knows immediately that Bob did a bad job. And, in fact, they SHOULD be able to repeat it (if time allows, of course) until the result is at least equal to or better than everyone's Passive Perception.

Implementing that wouldn't be a deal breaker for me or anything, but I don't think it should be universal.

[stoutstien]

It's called delayed or deferring rolls and it perfect for things like stealth or anything else that is as yet determined.

[gbaji]

I tend to agree that rolls should not be made until they matter. However, the exact example you gave is a bit of a tricky case, since it is reasonable to assume that if someone hides something really poorly, and someone else notices that it's really poorly hidden, this is something that could/should come up, allowing for a re-roll (or at least giving the PC another option than relying on his poor hiding job). Similarly, if someone is trying to sneak off to do something, and fails badly, it might just be obvious to others right around that character, well before they actually get into a dangerous situation. Really depends on the skill and the situation IMO.

Bill, overhearing the player conversation, is a metagame problem. But if Bill's character could reasonably have been in the area, noticed Bob hiding his stash, he might have instantly pointed out "Hey Bob. That's a terrible hiding spot". By handling this openly, you allowed for the metagaming. And even just not making the rolls at the moment doesn't solve anything, because Bill could still say "I'm going to try to find Bob's stash, to test how well he hid it", at which point the roll is needed, so you call for it, and we're left in the same position and with the same problem.

IMO, the better way to handle this is to ask Bob at the moment he declares that he's hiding his stuff, whether he's letting anyone else know he's hiding his stuff. If he says that he's not, and the room/area where he's chosen to hide his stuff isn't somewhere other people would have access to, then when Bill tries to interject, your ruling is "you can't do that because you don't know Bob hid anything, and you aren't in the area he hid it, so your passive search ability doesn't do any good". And then we move on.

I find that a good way to avoid metagaming like this is to ask your players very specific questions about what they are doing, and who else is involved in what they are doing. That way, the other players can't act on that metagame info, and if they do, you can rule against any actions they take based on that. It's not going to prevent all of it, but it does cut down on it quite a bit. It's always a tricky balance between player agency and metagaming. Players have an amazing ability to just happen to decide to do some unusual action, but only when doing that action will just happen to reveal some information, or provide some benefit/defense, that their character should know nothing about needing or wanting in the first place.

Sometimes, it's blatant, and you just firmly say "No. You have no way to know to do that". Sometimes, it's wobbly enough, you just let it go. It's always a judgement call though.

[Reversefigure4]

I do this routinely. It's primary purpose is to avoid the metagaming - after all, the character is doing the best job they can, and the 3 on the dice represents them not doing it well enough (and not realising it, because if they realise they weren't doing it well enough, they just do it again). Spending extra time, doing a routine job, or getting aid are all covered by various Take 20/Take 10/ Aid actions already.

As a secondary benefit, it also helps in scenarios where you'd have to remember the roll result for multiple sessions. The PCs set up a firewall on their computer, expecting a hacking attempt. Now, you could get them to roll, write down the 13, and remember it... Or, 3 sessions later when the hacker attempts it, roll Computers vs Computers then.

[icefractal]

I tend to agree that rolls should not be made until they matter. However, the exact example you gave is a bit of a tricky case, since it is reasonable to assume that if someone hides something really poorly, and someone else notices that it's really poorly hidden, this is something that could/should come up, allowing for a re-roll (or at least giving the PC another option than relying on his poor hiding job). This, exactly. "Having someone else check your work" is a very common thing in most industries, and saying "PCs shouldn't be doing that, they should be relying on their first attempt being perfect" makes them sound foolhardy and/or just plain dumb.

Although in this particular case, I'd go a step farther - why is Bob not taking 20? Doing a single roll is for when you hear the guards approaching and need to quickly stash the stolen goods out of sight. If you've got plenty of time to hide something, you take 20, and then there's no metagaming about the roll because there's no roll.

But all that said, I'd agree that when there isn't a way to judge something in advance (sneaking in to someplace alone, for example) then waiting to roll until it matters makes sense.

[gbaji]

It can be very game system dependent too. In the game system I play most often, you generally know if you made or missed your skill attempt, and how well you did, if you did. It's not really metagaming at all. You have a sense of how well you hide something, for example. So you know you didn't do a great job hiding the stash, but it's because you just couldn't think of a better hiding spot maybe. And we actually do allow for folks to realize they suck at something and failed, and just ask someone else to help them out (there's no "take 20" mechanism, so this kinda fills in if there's sufficient time). So in that system, it's just not an issue, since it's basically allowed/known behavior.

On the flip side, however, this same system allows for fumbled skill rolls, in which case you totally failed, but think you did really really well. Which can have some hilarious effects. Sometimes, it's quite obvious. The rest of the group looks askance at Fumbles McFumbletoes, as he trips over everything and keeps dropping the pots and pans that he's carrying with him on this stealthy mission for some reason (ok. Now I'm picturing Joxer the Mighty).

Sometimes? Not so much. Like when the party tracker fumbles his roll: Well gee guys... I could have sworn that mountain range was in the other direction... No. I'm certain that's exactly where we're supposed to go. (rest of the party facepalms)

[gatorized]

Some actions are easier to accomplish with help. In such cases, the person helping you makes their own challenge roll against a Hard (2) threshold and you gain a +1d bonus to your challenge roll for every 2 net successes they roll (rounding up as always). When working with multiple helpers, use the bonus provided by the one who rolls best.

Rather than bother with the assist rules, I would just give the hidden object the same rank as the player's covert or streetwise rank, whichever is better, with +1 for each person assisting. Then, if someone tries to find the object, I just roll that rank against their perception or investigation. This way, player is satisfied they've been given the best possible result, and there's no time wasted on multiple rolls. Although, if my player actually had the foresight to hide an important item, I probably would just not have it be found.

[Vahnavoi]

The principle is fine, but as others noted, the example is suspect. A character ought to be able to check another's work and then improve it. No skill checks are being invalidated by this, since the second character is trading time for their effort, and it is not a metagaming issue at all since the entire process makes perfect sense in character.

Personally, if I wanted to do something about your example, I would simply remove the (pseudo)random component from the hider's side of the equation. The target number to find an object is entirely decided by where it is placed and the player does have to make that decision, no rolls involved. As consequence, nobody can just try to roll a die again to do better, a player genuinely has to come up with a better hiding spot.

[King of Nowhere]

Besides agreeing with all the excellent posts on how it is absolutely reasonable for bill to check on bob's work, i also want to add that there are instances where it's better to roll early.
The classic "roll a spot check - rolls 1 - ok guys, there's something i failed to notice". The fact that the party is asked for a spot means there's something to see. Instead, if you ask at the beginning of the adventure to roll a spot for when it will be relevant, the party may be aware that they are going to miss something, but they don't know when. Similar with sense motive or other insightful checks

[Crake]

Besides agreeing with all the excellent posts on how it is absolutely reasonable for bill to check on bob's work, i also want to add that there are instances where it's better to roll early.
The classic "roll a spot check - rolls 1 - ok guys, there's something i failed to notice". The fact that the party is asked for a spot means there's something to see. Instead, if you ask at the beginning of the adventure to roll a spot for when it will be relevant, the party may be aware that they are going to miss something, but they don't know when. Similar with sense motive or other insightful checks

Or, y'know, the tried and true roll behind the screen? That way the players don't know what you rolled, why you rolled it, what the outcome was, or what modifiers were applied, or if the rolls even actually matter and you're just faking them out. You could be playing yahtzee while you wait for them to decide on a plan for all they know.

[stoutstien]

Id figure you would assume the PC to be competent unless otherwise stated. That means they wouldn't hide the object in a way that would immediately be obvious nor should there be any type of "roll" to determine how well it's hidden unless they are doing it in haste.

What you need to determine is what DC/DT would be needed for someone finding it in passing and the DC in case someone is actively seeking out hidden objects. Those would need more input from the player than "I hide the things" because the method of hidding would change based on who is doing the searching.

[gbaji]

Or, y'know, the tried and true roll behind the screen? That way the players don't know what you rolled, why you rolled it, what the outcome was, or what modifiers were applied, or if the rolls even actually matter and you're just faking them out. You could be playing yahtzee while you wait for them to decide on a plan for all they know.

Yup. This is my usual go-to methodology. If I don't want the players to know what the result of a roll was, or sometimes even what the roll was for in the first place, I just roll it secretly myself. And yes, this absolutely can/should involve some subterfuge. Randomly ask players for the values of different skills/stats on their sheets, then roll the die behind the screen. Maybe grin evily, check some notes, flip through the monster manual, nod to yourself, scribble some notes, then say "you notice nothing unusual going on".

Another idea, which an old GM of mine used to do, but I haven't quite picked up as dogmatically, is to have the players write specific information on a 3x5 card for each character. Put some indentifying info, and that character's stats and specific relevant skills (perception and any "special" skills mostly). I think he also had them write down obvious visual stuff, like what they were wearing (relevant if a group of NPCs are pickng targets and maybe will be deciding between "guy in plate mail" and "guy wearing robes and a floppy hat"). The GM keeps these as a reference for when secret rolls are required. This avoids the GM having to ask a player "Um... What's your characters sense amubush skill again?" (which may clue them into the fact that there's an ambush nearby). Basically, the GM can manage "passive perception" type stuff, without ever cluing the playrs in that anything is going on at all.

I don't go that far myself, but I do put some notes down for the adventure, relating to the characters on the adventure, and part of that is listing off any special skills characters have. I find that's usually sufficient.

[King of Nowhere]

Or, y'know, the tried and true roll behind the screen? That way the players don't know what you rolled, why you rolled it, what the outcome was, or what modifiers were applied, or if the rolls even actually matter and you're just faking them out. You could be playing yahtzee while you wait for them to decide on a plan for all they know.

yes, we should do that, but at our table we got used to just rolling. and when a player announces a high roll, you feel bad telling him that it's not valid because you must roll it in secret. and if you have accepted the good roll, you should now also accept the bad roll, yes?
utlimately, it's not important enough for us to want to stamp out this habit. plus, natural 1 - 20 often lead to hilarious moments that we would not have if we rolled in secret.

so, prerolling your spot checks is not an ideal solution, but it has merits when a group isn't really into rolling secretly

[NichG]

My preference is to make any sort of reactive/defensive/passive uses of skills either just be fixed at 10+modifier, or have specific interactions based on the value of the skill modifier or skill rank. This combines with a sort of hitpoint-based stealth system rather than opposed checks. So you can't 'roll to spot the ambush', but if your party has a lot of perceptive characters then they will deplete the enemy's stealth pips more before the ambush happens than a party without perceptive characters - that may lead to spotting the ambush entirely, or more likely to meaning that only some portion of the enemies in the ambush stay hidden well enough to get a surprise round, that attackers cannot afford to both attack and get their sneak attack / death attack / stay hidden / re-hide / whatever features. Generally the pip system is designed to make the stealthy characters proactive, so you pay a certain number of pips to 'get past this person' or 'cross this space' or 'attack from stealth without being seen' or things like that.

I like the 'only roll when the outcome matters' thing on the basis of a metagame consideration where, if there are no stakes, there should be no rolling and you just take average or use a different flat system. Basically, avoid rule design that leads to rolls as being treated as part of the physics, where a character could think in terms of 'proccing a roll'. You can work with your perceptive friend to hide something even better, but it should be at best be treated as an aid another kind of modifier, not this sort of played out dynamic of 'roll to hide, roll to perceive, infer you must have rolled low so roll again, ...'

The way I tend to modify stealth systems, I guess the way I'd have it work is that when you hide an object you grant it a certain number of stealth pips (and your perceptive friend might effectively push you past a threshold where it gets an extra pip or something via aid another rules or not, but it is what it is and the thing gets a fixed score that isn't based on rolling). Then figure out how frequently the area will be passively or actively searched for the object per day lets say, and the object is guaranteed to remain hidden for that many days modulo variations in the Perception or Search or whatever scores of the people who would be looking. Or if you're just smuggling it past a guard post, its just treated as a normal stealth interaction where you have a certain number of pips for the entirety of 'stealthy actions' you need to take in getting through the post. A post with 10 separate checkpoints would mean you'd have to pay for stealth separately 10 times, but one guard suspecting you and 'searching you over and over' can't subtract pips more than once - you pay to fool that guard, and if you paid the pips the guard is fooled.

The only time you'd roll would be if you had class abilities or feats or whatever that let you e.g. roll Stealth to regenerate a certain number of stealth pips based on the roll or things like that, which some of my systems do have...

[Pauly]

Speaking as someone with a 2 year old in the house, my wife and I regularly have to hide things from little fingers.

Using that as a real life parallel to the example given.
Sometimes I try to hide something, realize I’ve rolled a 1 and try again until I roll a success.
Sometimes I hide something and think it’s a good hiding spot then my wife makes her passive perception check and tells me I failed my roll.
Sometimes my wife and I think we’ve hidden something successfully and my son never finds it.
Sometimes my wife and I think we’ve hidden something successfully then my son makes and passive perception check and finds it.

Basically irl all possible gaming options happen.

[Crake]

yes, we should do that, but at our table we got used to just rolling. and when a player announces a high roll, you feel bad telling him that it's not valid because you must roll it in secret. and if you have accepted the good roll, you should now also accept the bad roll, yes?
utlimately, it's not important enough for us to want to stamp out this habit. plus, natural 1 - 20 often lead to hilarious moments that we would not have if we rolled in secret.

so, prerolling your spot checks is not an ideal solution, but it has merits when a group isn't really into rolling secretly

This is solved by simply not telegraphing that a roll is required at all. If you want to be truly stealthy about it, just pretend you're looking up something on your phone while you roll a virtual dice for it.

[Lord Torath]

One more thing (borrowed from Jay R):

Only let the player roll if both failure and success are acceptable results.

[KorvinStarmast]

How would the above scenario worked out if I had applied this rule?
Bob asks: "What do I roll to hide my stuff?"
I, the DM respond: "Nothing, not yet. I'll tell you when to roll"
Bill- still not satisfied: "Can I help him?"
Me: "Yes, you and Bob spend the morning hiding your stash. Bob will have advantage when he rolls"
[... the game continues ...]
As the players depart, I say: "Make an wisdom (stealth) roll with advantage because Bob helped you"
Bob: "Ok I rolled a... 1 and a 1. For a total of 3, LMAO" (or maybe he rolled a 1 and a 17 for a total of 19, exactly the same outcome, who knows what these tricky dice get up to. But this way they didn't metagame.)f
Works well enough.
Only roll when success or failure matters. Or when success or failure is interesting.
DM calls for the roll.

You can cut down on a lit of dice rolling if you stick with those those two guiding principles.

[King of Nowhere]

This is solved by simply not telegraphing that a roll is required at all. If you want to be truly stealthy about it, just pretend you're looking up something on your phone while you roll a virtual dice for it.

But i like rolling my dice. I didn't spend 30 € on shiny sparkly pretty dice to have the dm roll a virtual dice on his phone for me.
I also like to flaunt my +30 to spot. There's a simple pleasure in the exchange "make a spot check" "i got a 45". Again, rolling in secret undermines that.
Finally, it's fun to roll natural 1s and 20s and get special effects on that.
You lose a bit on immersion and realism, but you gain many great anecdotes

[Crake]

But i like rolling my dice. I didn't spend 30 € on shiny sparkly pretty dice to have the dm roll a virtual dice on his phone for me.
I also like to flaunt my +30 to spot. There's a simple pleasure in the exchange "make a spot check" "i got a 45". Again, rolling in secret undermines that.
Finally, it's fun to roll natural 1s and 20s and get special effects on that.
You lose a bit on immersion and realism, but you gain many great anecdotes

shrug there's plenty of other rolls that you can use your 30$ dice on. I also personally don't use fumbles and crits on skill checks, I think they're silly. And especially with a +30 to spot, 99% of the time I wouldn't even bother, cause a 31 would be an auto succeed.

[gbaji]

But i like rolling my dice. I didn't spend 30 € on shiny sparkly pretty dice to have the dm roll a virtual dice on his phone for me.
I also like to flaunt my +30 to spot. There's a simple pleasure in the exchange "make a spot check" "i got a 45". Again, rolling in secret undermines that.
Finally, it's fun to roll natural 1s and 20s and get special effects on that.
You lose a bit on immersion and realism, but you gain many great anecdotes

This is why I usually have players roll their skills, and then tell me the roll itself. Then I apply whatever modifiers were present for the thing I had them roll for. This allows the players to have fun rolling their dice, but also doesn't necessarily tell them anything.

And yes, I do call for a lot of rolls, for a lot of different things. Some of them are super important, some may be very minor. The players don't know this though. In another thread, some folks were shocked when I mentioned that as many as a couple hundred perception rolls may occur during one of my game sessions. Yeah. Because I may just be having them roll to determine how much detail I'm going to give them about the local fauna maybe. Or whether they notice that they've moved from an area with red clay dliff sided hills to rolling green hills instead. Or whether they hear the fact that there's a river with a large waterfall nearby, maybe.

By just doing this routinely for all sorts of things going on around the players, at all times, then the players don't realize when this perception roll is actually to determine if they noticed the group of bandits hiding in the bushes along the path in front of them. And hey. Maybe it does become relevant later that the characters noticed some minor seeming change in the environment around them. Some of this is just setting information too. So when the players later are told to go talk to the local mill owner, this "fits" with the fact they previously knew about a good sized river a while back, and fills in their mental map of the world. And if they later learn that someone went off to hide in some caves up river, they may think "we heard a waterfall a day back, maybe the caves are near that?".

But all of it acts to get the players used to using their skills to explore and interact with the world around them. Which I think is a good thing. Far too many players ignore like 80% of the skills their characters actually have (even ones they don't have specific points allocated to). And yeah, in level and skill point game systems, this encourages players to maybe spend points on skills like this rather than always in a build focused way. If they've realized that these skills can provide them with useful information (ie: it's actually valuable, and the GM isn't always going to just spoon feed them scenario useful information in their absense), they will do this. And the result is more balanced characters and parties than otherwise.

[Composer99]

I haven't read through all the replies, so my apologies in advance for duplicate suggestions.

As for my two cents, I think the anecdote as presented is almost entirely fine, with no lashing required. Maybe Bob should have rerolled, to represent having advantage from Bill's aid, instead of letting Bill roll. Alternately, if Bob and Bill had enough time to find the best hiding spot, just assume they successfully hide the item as per the guidance in the DMG (pg. 237, "Multiple Ability Checks").

As for the suggested method to reduce dogpiling of die rolls, I'm inclined to say that the rules for working together, for group checks, and the DMG guidelines cited above regarding automatic success are probably just about sufficient to deal with the kind of issue you're mulling over.

What I would suggest is to keep closer track of time when it actually matters to do so and to make sure the player characters have announced their activities over that span of time. How much time do Bob and Bill have to try to stash stuff away? Ten minutes? An hour? All day? To my mind, you probably only need a roll if it's ten minutes (the old-school "dungeon turn") or less - and you only need to restrict Bill's ability to aid Bob if whatever Bill is doing reasonably precludes it over that time period.

All that said, yes, if you're calling for a roll, it makes sense to call for a roll at the point when it actually matters.

[Mastikator]

But i like rolling my dice. I didn't spend 30 € on shiny sparkly pretty dice to have the dm roll a virtual dice on his phone for me.
I also like to flaunt my +30 to spot. There's a simple pleasure in the exchange "make a spot check" "i got a 45". Again, rolling in secret undermines that.
Finally, it's fun to roll natural 1s and 20s and get special effects on that.
You lose a bit on immersion and realism, but you gain many great anecdotes

You'd still be rolling for stealth, rather the timing of when changes.

[WilliamJoel333]

When there is not a ticking clock and a dice roll won't significantly change the outcome of a situation in a meaningful way, I wouldn't have your players roll dice at all. I'd simply tell them what they find or how they accomplish the thing they are trying to do.

When a dice roll becomes necessary, I prefer to tell them up front what the DC is. That way they know exactly the number they must roll. The tension is in the dice roll and all eyes are glued.

Some exceptions are when players must come and make their roll behind the DM screen so that the party doesn't know if they succeeded of failed.