I'm sure every DM has had this happen: one PC is trying to find something cool on a dead monster's body. The player rolls poorly, and instead of accepting the paltry 2 copper pieces, the PC calls all the party members over to also try looting the corpse. Invariably, a different player rolls well, and everyone walks away satisfied that everything's been taken down to the poor dead monster's last holey sock.

Another example: One PC is trying to find some useful information in a library. There're no guards, no reason anyone would be bothered, and no time restraints to stress about. The player rolls poorly, so all the other PCs try, too.

I'd describe this as well-intentioned metagaming. The players are afraid of losing out on content, this despite multiple sessions of me making it clear that I don't structure my games to penalize gameplay that way.

These are situations that require a roll to determine the degree of success. (At least I think so; perhaps I'm wrong here?) Failure could set back the PCs a cool item or a bit of information that would have helped a little, but these minor failures aren't game-enders or anything. As a DM I can't see a reason to say "No" and disallow the practice, but it feels vaguely like cheating to let the players roll multiple times to get better results.

How do I prevent players ganging up on skill checks after they've seen someone roll poorly?


15 Answers 15


The best advice I have seen on this issue is from Angry GM’s 5 Simple Rules for Dating My Teenaged Skill System (warning: mild, censored swearing of the $^#% variety). That’s written for D&D 4e, but like you say, this is an issue that has perplexed GMs for ages, across a variety of systems.

His answer, which is his rule #2, is very simple:

Only roll if there is a chance of success, a chance of failure, and a risk or cost to failure

You need all three to have a roll. If you have only one or two of the three, or none of the three, there is no reason to roll at all. If there is no chance of success or no chance of failure, the pointlessness of the roll is self-evident,1 but the third point is key here: if there is no reason not to, PCs should and will keep retrying until they succeed (as much as they can succeed). There is no reason to bother wasting time having them actually do so; just assume they do it and move on. Save game time for something more important.

And “it takes longer” is not a risk or cost to failure unless there is a clear and present time crunch being applied to the PCs. If they are under attack, sure, taking more time to search a corpse is risky and/or costly—they risk losing their hp, and possibly lives, by doing it. Likewise if the room is filling with water, the big bad evil guy is chanting in the corner to finish his summoning of a bigger fish, or whatever. But it has to be clear and known to the PCs to give them a reason to rush.

But if they’re just exploring an apparently-abandoned tomb, at their own leisure, there is absolutely no reason in the world for them to not take their time being thorough. It does not improve the game to constantly ask them if they’re going to be. It really does not improve the game to constantly stop and roll and check results and maybe try again when it doesn’t matter. And it also doesn’t improve the game to arbitrarily limit retries; in addition to being unrealistic (rolls represent one attempt, and the whole point of it being randomized is that not every attempt at something will be your best), it also runs into severe goblin dice problems.

Note, however, that this answer assumes implicitly that this kind of thing is a low-value use of limited play time. A whole lot of the trade-offs involved here are made in order to minimize play time spent on this issue. That only makes sense if we agree that this activity is not a major, important, or interesting part of the game. This answer presumes such a playstyle because that is the style that 5e itself seems to espouse—5e continues a progression that largely started with the acquisition of D&D by Wizards of the Coast that focuses more on the epic narrative, the quest, and the characters, than it does on careful dungeon delving, handling preparation and logistics, or on player skills. But even in 5e, that’s not all playstyles—many people play with different styles with different emphases. Particularly since in older editions of D&D, it was presumed that this sort of thing was a very important, interesting part of the game. Said editions wanted to spend more time on this in part because they also emphasized a fairly heavy use of player skill as opposed to character skillplayers had to be thorough, players had to think of places to search, and so on. And many people continue to play newer editions that way. If that’s your playstyle, the choices made in this answer would make no sense, since they are emphasizing different things than you are.

  1. Some will object that choosing not to roll at all gives away information to the players about something’s difficulty. That is, they will argue that not allowing players to roll for something they cannot possibly do tells them explicitly that they cannot do it, when otherwise they could not conclusively know that unless they rolled enough to achieve a natural-20 and still see a failure. Suffice to say that the article acknowledges and addresses this concern—some may not be convinced, but personally I think the discussion on that subject is very well-considered and convinces me, anyway, that this is for the best. The short, short version is that rolling enough to get that natural 20 is a huge waste of time that doesn’t add nearly enough to the game to be worth it.
  • 3
    \$\begingroup\$ @SevenSidedDie I have made an attempt to address the implicit assumption of playstyle in my answer; I think it has improved it, but only if it accurately captures its own playstyle assumptions. I would be interested in your opinion of how well I have done so. Like you, I have little desire to try to offer suggestions for other playstyles (I don’t have enough experience with them to judge any particular approach), but I do want to be sure my answer is clear about its own biases. \$\endgroup\$ – KRyan Mar 26 '18 at 19:51
  • 15
    \$\begingroup\$ While I have "8 simple rules" bookmarked, I do think it's worth also pointing OP to DMGp.237, which concisely lays out the same "don't roll for everything, guys!" philosophy. But with much less explication than Angry, so I would never fault someone for pointing primarily to Angry over the DMG. I just think it's worth pointing out there is core support for your solution. \$\endgroup\$ – nitsua60 Mar 26 '18 at 23:04

Option 1

One roll with Helpers

The party gets 1 try based on who is primarily initiating the action. Party members assisting can confer advantage but not keep trying. The party is not allowed to retry the same check unless something changes.

Option 2

Hidden result

The person doing the investigation does not roll, they tell you (the DM) what their modifier is and you roll for them. This makes sense for skill checks where a person doesn't actually often know how effective they were; Stealth, Insight, Investigation etc.

You only tell them if they succeed or fail, not what they rolled.

Option 3

Add a time cost to these actions (or some other cost)

In theory, even with an investigation roll of 1 an adventurer can find goodies on a corpse given 10 hours. If they are willing to search for longer, consider lowering the DC. In a dungeon, adventurers may not have 5 minutes to loot every corpse, so they need to roll to do it fast.

If the first player rolls poorly that constitutes using the entire window they determine (10 mins, 1 hour etc). They can ask allies help or retry, but that will cost more time.

  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$ – mxyzplk Apr 17 '18 at 3:10
  • \$\begingroup\$ +1 and In option 2 I think it's more effective to tell the pc the result of their action based on the roll, not necesarily if they succeeded or failed the roll itself. \$\endgroup\$ – lightcat Jan 12 at 10:59

I always make information-based rolls in secret. There's no magic to who rolls a die, so mathematically it's the same, and it 100% solves the metagaming problem.

There's a small downside, which is that we like to roll our own dice because it feels like then we're in control of our fate, and rolling for a PC takes that illusion away.

But that is a small downside that's worth it, as a trade-off, for a smoother-running game that doesn't strain suspension of disbelief, doesn't encourage metagaming, and (importantly!) doesn't discourage the DM from bothering to add cool hidden things to the game.

That last is important. If everything will always be found, why hide anything? And if the DM gives up putting cool secrets into the game, the players will eventually notice. And then why should the players keep exploring on their own initiative — the DM will tell them everything without trying, right?

At least, that is my experience: allowing the metagame to control finding hidden things leads to an apathetic DM and less-engaged players. (It also especially robs the players who specifically enjoy the feeling of accomplishment from discovering something that would have stayed hidden without their clever observation and investigation.)

So yes, I have strong reasons for rolling in secret for rolls that would give away a secret just by a player comparing the die result to what happens. The illusion of control is just infinitesimal compared to the benefits I experience in our games (not just as a DM, but as a player when my DM uses secret rolls too).

If your players take some convincing, you can explain it like this: This isn't about the player character's skill, it's about revealing parts of the world. You're just using their PC's skill to help decide how much of the world to reveal, and that's firmly DMing activity.

  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$ – mxyzplk Mar 28 '18 at 4:43
  • \$\begingroup\$ Do you include picking a lock roll as "information based" or is that not in the class of checks that you treat this way? (I already up-voted). I ask for clarification due to a norm among my early edition DM's who rolled all trap find checks behind a screen, and some did the same for lock pick checks. Your style is one I am very familiar with and I agree (experience based) that it provides fun/mysterly/tension on both DM and player side. \$\endgroup\$ – KorvinStarmast Jan 11 at 16:26
  • 1
    \$\begingroup\$ @KorvinStarmast I usually treat lockpicking as not secret. Unlike spotting an ambush, a PC is never picking a lock they don’t know about, and unlike discerning whether an NPC is lying, they’ll know their success or failure immediately. (I can imagine rare exceptions to the last point though—maybe it’s a “lock” with many settings in some large byzantine device, and whether they accomplished their picking goal is something they learn elsewhere in the device.) \$\endgroup\$ – SevenSidedDie Jan 11 at 16:44
  • \$\begingroup\$ Makes perfect sense, thanks. \$\endgroup\$ – KorvinStarmast Jan 11 at 16:44

Turn Time into a Mechanic

Let your players know, that if they roll poorly, or even before they roll, they may choose to take time to succeed. Anyone can do anything given enough time, assuming trying doesn't injure them.

The character takes an amount of time in order to acquire what is the equivalent of a passing roll. This can apply to alot of situations. If they need to break down a door, but they're not in a rush, they can take 10 minutes to simply acquire a roll of 10, plus their STR modifier. In your case, if they're searching a body, they can take 10 minutes to thoroughly search the body acquiring a roll of 10. If they want a higher roll, they take more time.

This has the added benefit of not just letting players succeed on a roll in the middle of a tense situation. If they're in a stand off with an enemy, they could try to take 10 minutes to talk them down and acquire a roll of 10, but during that time it's just as likely the combat starts and their roll is interrupted. Or if they're running their way out of a collapsing dungeon, this mechanic doesn't remove all tension, because they can't stop and sit around for 10 minutes remembering how they got in, they have to roll and keep going.

I've used this to great effect in my group. In my experience to Take 10, they need 10 minutes, and to Take 20, they need 30 minutes.


What is a skill check and what does the d20 represent?

If the d20 represents "how well I did it this time", then what your players are doing makes some sense, especially if players have an idea of how well they are doing. Restricting players to "1 roll per player" doesn't really help here.

This, however, leads to boredom; pointless rolls.

Roll for difficulty

An alternative is that the d20 represents "how hard the task actually is". When you roll d20+STR vs DC 15, and you get a 1, it means that while most doors in this dungeon are medium-difficult to open, this particular door is exceedingly hard. In this case, a reroll only occurs if the situation is "reset" in a fundamental way.

If someone offers help?

Aid just modifies that existing roll. So if your STR is +5, you need to find +9 points of modifiers to open that door. Aid another? That is worth +2. A +5 luck bonus from the Bard? Almost there. Potion of giant strength granting +3 more strength? Finally, the door opens!

Repeated rolls under this system, where the roll reveals how hard something really is, don't make sense. DCs in this system are just rules of thumb the DM is following; a "typical" door is DC 15, to find out how hard this particular door is you roll your d20.

What more, the information revealed by the door being hard to open can influence later checks. If the door was unopenable because it was swelled from water, trying to pick it after won't do much good; but a spell that dries wood could. If it was unopenable because it had metal-reinforced bars, picking the lock might help.

Roll for skill

If the original approach holds -- the d20 represents how well the player tries -- then you have to decide when to ask the player to roll.

A player should only roll when there are consequences to the roll.

  • If there is no time pressure and failure means you waste time, then there are no consequences to the roll.

  • If the task is impossible, but a bad failure will cause damage, then there are consequences to the roll.

In general, the state of the game after a roll shouldn't be one such that "I try again" would always make sense. If it is, consider failing forward.

Failing Forward

Imagine an impossible task. You decide that on an DC20 check you fail, but you learn it is impossible; if you roll 19 or under, you take 3d6 damage and learn that it is impossible. This is a check with consequences and "I try again" doesn't make sense.

Now, imagine a dangerous task. On a DC20 check you succeed, on failure you take 2d6 damage and can try again. In some circumstances this is reasonable, with the failure damage generating tension.

But it might be better to fail forward. DC20 you succeed. DC 15 you succeed after taking 2d6 damage. Under DC15 you take 2d6 damage and can choose to either take another 2d6 and succeed, or break it.

This is "fail forward" -- you fail, pay a price, and you proceed forward in the plot anyhow. It even has a gradient of failure (or a gradient of success depending on how you define it).

Notice I offered a bargain above. They could choose to give up on the obstacle, or pay a price and pass it. This could be overly metagamey for you.

Example: "I loot the bodies"

Going back to the critters. You kill a bunch of Orcs. You then do a scavenge check to see what they have of value.

Under "roll determines difficulty", your roll actually determines what is worth scavenging. On a low roll, stuff was destroyed or lost prior to finding it. On a high roll, more stuff was there to be found. Your scavenging skill factors in, but your efforts are assumed to be maximized; you find everything you can find.

Under "fail forward" with a bargain, maybe you say "you find 28 silver: If you want to risk getting a disease I'll give you another scavenge check." Without the bargain, the poor scavenge check might just give you lice (a minor disease), but regardless you find the decent quality sword owned by the orcs.

Under "only roll when it matters", have a fixed "scavenge skill to loot" result. Unless the players are pressed for time (say are being chased), in which case you can roll to determine how fast they can get the loot ("It's taking you a while to search; do you really want to when the worgs are after you?").

  • 1
    \$\begingroup\$ I like this answer, but it's a really unattractive wall of text. I'm reluctant to try editing it for fear of leaving out some of your intent. \$\endgroup\$ – goodguy5 Mar 30 '18 at 14:16

Others have already mentioned about only rolling if there is a chance of failing.

I will add that, in your actual example, rolling to determine what is on a dead body is, in my DM opinion, certainly not something that ordinarily requires a roll. The DM decides what a body has on it if anything, and no amount of rolling skill checks should change that!

On the other hand, if the PCs know the body may have something important but they are on a time pressure (to escape the room before the poison gas overcomes them!), then a skill check of some kind may be in order (with a failure just meaning they haven't managed to find it yet, so need to spend another action in an increasingly dangerous situation).

It isn't that dissimilar to searching for information in a library. Given enough time any character could find it, so if there is no time pressure I (as a DM) would be fine with saying they find what they are looking for, assuming the information is actually there.

But if it is important to the plot, then a skill check means they find it this day and every check failure means another day of searching while the army draws ever closer to the city...(with a consequence you could maybe fit into the plot at a later stage).


Pathfinder resolves this issue beautifully. You can take 10 or 20 (minutes) to get a roll of 10 or a 20 for your roll. So, if you have a lock and you have a lockpick, instead of just rolling until you get a 20 or a 1 (break the lockpick?) you just take your time, and pick the lock after 20 minutes.

Similarly, Neverwinter Nights (or similar DnD video games) also take this into account and if you have no risk of failure and the chance to roll the dice for as long as it takes, they just give you the highest amount you could roll.

You can just integrate a similar system to your game. You don't have to have it actually take time, but yeah, you could just assume they roll the highest amount they could.

  • \$\begingroup\$ I like this answer. Fifth edition is really missing "take 10" and "take 20", since the alternative is to keep rolling. The only case this doesn't work is something with no rerolls like knowledge checks, where 5e allows e.g. untrained Arcana checks, so if the wizard rolls poorly the fighter might roll, and succeed. \$\endgroup\$ – Quadratic Wizard Mar 27 '18 at 14:09
  • \$\begingroup\$ @QuadraticWizard in Pathfinder afaik, you can't take 20 on knowledge checks (if there are negative consequences, you can't take 20, and not being able to know a thing is a negative consequence imo). You can take 10 though, since that's the whole point of taking 10 (rolling average instead of risking a low roll). \$\endgroup\$ – John Hamilton Mar 28 '18 at 5:12
  • \$\begingroup\$ @QuadraticWizard I don't see why that's a problem. The wizard rolls really badly, and fails to know something. The fighter rolls really well, and does know it. The fighter presumably picked up that scrap of information someplace. The GM might want to raise the difficulty of knowing something a bit, to make up for the increased chance that someone in the party knows it. \$\endgroup\$ – David Thornley Jan 11 at 21:45

When there's no consequence to failure, I don't have my players roll at all. I just have them designate the "team lead", decide on the difficulty (and therefore a target number - not necessarily "rollable" if dice were being used) of finding the Special Thingies, and then pick a relevant skill that the team lead uses at full value, and anyone else uses at half-value as an assist, and a second skill that can be used instead of the relevant skill, but at less than full value. I then sum up the skill levels being used, and compare it to my difficulty/target. If they have enough levels, they find the Special Thingies; if not, not.

(N.B. "full value" and "half value" refer to the actual skill level as shown on the character sheet - that is, if the character has Investigation-3, then in counting toward the target number, "full value" is 3, half value is 1, 1.5, or 2, depending on circumstances [round down, don't round, round up])


Once possibility not yet mentioned is to metagame around the players metagaming attempts.

So a player searches a corpse but only finds 2cp. Let the other players roll, but regardless of their result, do not let them find anything else. That first bad roll then becomes "what there is to find" rather than "how well the character finds it".

Over time, the players will stop trying to reroll, because it does not change the original result.

Of course if you WANT them to find something - a clue or whatever, you can let them find that no matter how bad their roll.


I think the "take 20" mentality could be of value here. If the characters are going to (and can safely) search every nook and cranny, they should find everything (or everything that's not blatantly hidden).

I'm not sure if 5e has an equivalent, but 3.5e has the option for users to "take 20". It basically meant they have the conditions are such that they can take their sweet time doing something perfectly. So instead of rolling the dice, you just treated it as a dice roll of 20.

I think if your players are adamant about finding loot, just hand them the loot. No need to make it difficult. Anything hidden you can perform a hidden search check (unless they announce they're searching for something, in which case they should roll).


You said this in a comment:

In an effort to "fix" the problem I ruled only one roll per check per person unless you've somehow changed the situation. Unfortunately that's what's wound me up exactly where I am now...

It seems to me that only one or two of the other answers actually address , and none of them on why it can be — but is not always — a problem.

First thing you need to consider is whether or not they are actually metagaming. Metagaming only occurs when one has lost confidence in a game system to provide them with whatever it was that the game did — whether that be a story, escapism, a challenge, or even education and testing of their problem–solving skills.

Of course, that fourth one is not very pertinent here — and, indeed, seeking out loopholes in your mandates and thinking outside the box could be not only acceptable but encouraged. Anyway …

If they are indeed metagaming, it means that they are using your game to achieve or satisfy some need which your game does not proffer.
Maybe they don't care, and are simply using it as an excuse to socialize. That doesn't seem to be your problem here.
Maybe they like your settings, but they don't like one of your rules: so, they want to find a way around the one thing they don't like while keeping everything else.

Have you discussed with them the reason for your limitation on “only one role per check per person”? Have you explained to them what it means in the mechanics of your dungeons? (Other answers talk about what possible mechanics to use in that regard if you hadn't devised them already.)
Is it possible that they are unaware that you consider them to be metagaming? Maybe they think that you wanted them to discover and exploit your loophole. Maybe it is possible to work out a compromise between yourself and your players. Of course, pardon me if I seemed to be patronizing, but there it is.

The gist of my recommendation is this:

  • you devise mechanics not to prevent something but to make it unpleasant and hopefully to deter any interest in it;
  • you design mechanics to give yourself and the players an enjoyable experience.

If your players are attempting to metagame, then that means something went wrong either with designing a set of rules for your group or with recruiting a group for your rules.

  • 1
    \$\begingroup\$ "Metagaming only occurs when one has lost confidence in a game system" I wouldn't think so, if you consider the range of reasons people play RPGs. Some power players if not restrained would metagame, for sure. \$\endgroup\$ – person27 Mar 28 '18 at 5:43
  • \$\begingroup\$ That's exactly what I mean, @person27. A player has certain desires or wishes which aren't quite satisfied by a certain system — so, they metagame. Not all powerplayers do so, of course. By powerplayer, I take it you mean one who strives to take full advantage of every rule and function available to them. Non-immersional, to be sure, but that's not what I would classify as metagaming. \$\endgroup\$ – can-ned_food Mar 29 '18 at 1:49

For the library situation: An idea from Rogue Trader: we have the mechanics of exploration checks. The DM decides how many DoS are required to succesfully tackle a task. Then every character can contribute with an according skill (needs to explain how this skill helps), then the DoS get added up and that way we determine if we succeeded. This can AFAIK not be taken directly (never played 5e) but can give you an idea on how to tackle the task. It also distributes the success among the players and it's less like always the same player doing some tasks.

Concerning Metagaming: Punish them for it. If a player uses Metagaming / OOC knowledge and so on, give him half the XP. If he repeats, only a quarter. In case he continues: 0 XP. See if he does it again.

  • \$\begingroup\$ Maybe the group isn't big on roleplay, though. I didn't see the question complain about OOC or anything like that … maybe you have more experience and can read it between lines. \$\endgroup\$ – can-ned_food Mar 29 '18 at 1:52
  • \$\begingroup\$ Maybe I interpreted too much of my taste into his writing. I hate OOC / Metagaming / illogical behaviour cause it kills roleplaying. However, I have to admit, I've never played 5e but around 20 years of p&p rpg experience. \$\endgroup\$ – Shade Mar 29 '18 at 7:51

You seem to be not clear what you are actually rolling for.

Is it a perception check? If so, how do you justify that the PC misses a +2 Greatsword if he is not in a hurry? As mentioned in other comments, if a PC is calmly searching a corpse, you should assume that he doesn't miss anything that is not intentionally (and well) hidden.

Is it a random loot check? In such case, the roll determines what was on the body, so a re-roll makes no sense. Someone else searching will find same.


The items on the monster existed before any PC decided to loot them. Therefore, the roll for which items are present is a DM roll and should be hidden.

Unless thoroughly looting that specific monster is somehow more difficult than walking, no PC roll is necessary.


The GM is role playing too

Sometimes it's easy for us GM's to forget that we're role playing too. Even when you're narrating you're role playing as the narrator, you're not you, you're an omniscient all-knowing being known as the Game Master.

Often a simple change in my tone of voice can cue the players into everything they need to know. For example if they start rolling to search every small rocky crevice of a cave wall I'll stare at the first player who rolls and without even looking at the result of the die I'll say in my flattest most monotone voice, "you don't find anything."

This let's the players know that their die roll is irrelevant and the GM (the person I'm role playing as) is not pleased.

This is not to say I always employ this technique when there's nothing to find, more often I improv, sometimes the whole course of the days adventuring is changed and occasionally I let them search as long as they want if they seem to be having fun with it. Throwing in little rings and trinkets seems to make many players very happy so I give them that occasionally. They especially like trinkets with details, like little bone carvings of the family of the goblin they just slaughtered.

Look through all these answers and try different techniques. Switch it up and see what gets good results with your particular group of players. The main thing is to keep your players interested and excited so don't get stuck in just one method.

  • \$\begingroup\$ The GM roleplays too, but I wouldn’t say the GM is itself a fictional role, as there is no separation between the person acting as GM and the “character” of the GM. Thus, this won’t be received by anyone as the GM playing an annoyed GM, it will be received as the person GMing being annoyed (and passive-aggressive about it). There’s definitely ways to verbally and non-verbally convey information like this, but nobody will experience this as leveraging a “character”. \$\endgroup\$ – SevenSidedDie Jan 12 at 18:21

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.