Picking a lock typically requires a DC 15 Dexterity check (Basic rules, p. 49). Likewise for manacles (p. 50), which can alternately be escaped with a DC 20 Dexterity check or broken with a DC 20 Strength check. But what happens if you fail?

The general rules for ability checks state (p. 58):

If the total equals or exceeds the DC, the ability check is a success—the creature overcomes the challenge at hand. Otherwise, it’s a failure, which means the character or monster makes no progress toward the objective or makes progress combined with a setback determined by the DM.

As a DM, how can I handle this in practice? In our last session, the players wanted to get inside a locked building to rescue people trapped inside. Picking the lock seemed like an easy enough task for the rogue, but unluckily he failed the check, and I wasn’t prepared for the consequences. There wasn’t much time pressure and no obvious reason why he couldn’t try again, so I allowed another attempt, but it just didn’t feel right.

I’m sure that similar situations will come up again, so I’d like some advice on how to deal with checks so that they don’t just turn into a tedious exercise of rolling until you succeed. I presume that “take 20” was left out of the game for good reason, but I’m not sure what to do instead. I suspect that this was a good occasion to use “progress with a setback” but that’s a fairly new thing for D&D and I’m not sure how to go about it. I’m unsure when I should use setbacks versus lack of progress, and in the latter case I’m unsure when I should allow players to try again versus requiring a different approach.

For example, suppose that a manacled player tries to pick the lock and fails the DC 15 Dexterity check. Now, the player could try to pick it again, or he could try one of the other escape methods, or he could wait for somebody else to rescue him, or he could simply succeed with some setback like a minor injury. How do I choose how to proceed? Or do I let the players choose?


11 Answers 11


Don't ask your players to roll the dice unless failure has a consequence.

This mantra isn't particularly obvious in the rules of D&D as many times the checks are relatively pointless and failure at those checks doesn't really come with much cost. Failure is an opportunity both for you as the DM and for your players. Here are some ideas as to how to make failure have a cost in the situation you describe:

  • A second attempt at picking the lock costs time, and the life of a hostage.
  • The lock is ruined and the door must be forced causing noise and alerting the guards inside.
  • This door is not an option and you have to find another way into the building.

These are three ideas of things you can have in mind if the check fails. The important thing is that failure should cost something (not a lot, it's 1 d20 roll, but something). Make it something small, the next encounter is slightly harder, the opportunity for surprise is lost, a later plot consequence etc.

Nothing major should depend on your character succeeding a skill check. This is because skill checks fall afoul of the Goblin Dice concept, which is that important things should not be decided by a single swingy die roll.

Don't let failure stall the action. Use it as an opportunity to move it forward, make it more interesting and to raise the stakes.

  • 14
    \$\begingroup\$ I think your third bullet could use a tiny bit of explaining: I think the notion of "failure at DC15 lockpick check"="you've poked around and realize that this is a lock well beyond your capability" is not a widely-known interpretation of checks and their failures. (But I think it should be, so I hope you'll put a few more words in there.) \$\endgroup\$
    – nitsua60
    Commented Mar 16, 2017 at 14:40
  • 8
    \$\begingroup\$ @nitsua60 This. A character doesn't know they rolled a 14, they just know that what they tried didn't work. They could be an inch away from the solution, but they won't necessarily know that they are. This could (at the DM's discretion) be a place for a passive check like intuition, to see if they realize how they messed up and that they should just try again. \$\endgroup\$
    – thanby
    Commented Nov 10, 2017 at 13:25
  • \$\begingroup\$ This is a great answer but, I would also add that in the example you give above, failure to pass the DC doesn’t have to mean the lock picking itself fails, first of all set the DC in your head, allow the player to roll. The door opens, the player is happy, what they don’t know is that because they failed the DC somthing else results, maybe they have damaged the lock meaning that the door can’t be shut properly, maybe they made a noise as they did it that was heard, maybe they left marks that a passing guard will spot and investigate. \$\endgroup\$
    – Richard C
    Commented Jun 21, 2020 at 13:30
  • \$\begingroup\$ Eg failure should be interesting, best character development moments in one of my campaigns happened because of a failed stealth role. \$\endgroup\$ Commented Aug 31, 2023 at 15:36

Otherwise, it’s a failure, which means the character or monster makes no progress toward the objective or makes progress combined with a setback determined by the DM.

In the 5e Basic Rules, the 5e PHB, or the 5e MM progress combined with a setback is not specifically defined.

When this is the case, the first place to look for an answer is to look at why you make ability/skill rolls in the first place. For example do you make a dex check to pick something off the ground? Or only when in you are combat? Or perhaps when you are on a rope trying to get something when you swing past?

An ability check tests a character’s or monster’s innate talent and training in an effort to overcome a challenge. The DM calls for an ability check when a character or monster attempts an action (other than an attack) that has a chance of failure. When the outcome is uncertain, the dice determine the results

So an ability check is called for when

  • To overcome a challenge
  • Has a chance of failure
  • When the outcome is uncertain

Let's look at your example of picking a lock. I think definitely can be considered a challenge. And one is tempted to add in both there is a chance of failure and the outcome is uncertain. But is that always true?

I would say it is definitely true if one has limited time in which to pick the lock. Or one is trying to be silent while picking the lock. Or trying to avoid a trap on the lock.

But if a character has a whole day in a private room then it is all but certain that the character will pick the lock.

So using the guideline that failure means progress with a setback. Then failure will mean it take lot longer for the character to pick the lock perhaps incurring a wandering monster check. The character makes a lot of noise picking the lock.

I recommend that you use the resulting die roll as a guide to impose how bad the set back is. Perhaps natural one means that the lock and lockpick are ruined. A less than a 5 means the lock is ruined. Less than 10 means the lock takes a really long time (20 minutes or 2 wandering monster check). Less than the DC 15 means the lock takes 10 minutes or one wandering monster check.


I've been reading about this recently, both an AngryGM and in the XDM book by Tracy and Curtis Hickman.

The short answer is as follows.

Character attempts an action

  1. If there is no chance of success then the action fails and the character pays the appropriate cost.
  2. If there is no chance of failure then the action succeeds and the character pays the appropriate cost.
  3. If there is a chance of failure but failure has no meaningful effect then the action succeeds and the character pays the appropriate cost - possibly multiplied because the character repeated the action over and over until it succeeded.
  4. Roll the dice! The action succeeds or fails, based on the result of the dice, but the character still pays the appropriate cost.

The Cost

All actions have costs. This might be time (it takes 5 minutes for each attempt to pick the lock) or material components (each attempt to start the fire uses up some tinder) or reputation (each attempt to bribe the guard makes them more angry) or something else. Every time the character attempts an action, they must pay these costs.

However, if the character (not the player) knows the action cannot succeed then the player can choose that the character never even makes the attempt, in which case no costs are paid.

The Consequences

Everything has consequences, but (and here is the important part) the consequences come from the approach, not from the results. There might be five different approaches to a problem (some of which might involve the same skill roll) but because they have different consequences, then they are meaningful choices. If you get this right, then your players stop saying "I use Acrobatics" and start saying "I leap over the rail and grab the chandelier". The second phrase is roleplaying; the first is just rollplaying.

  • 5
    \$\begingroup\$ Good advice, although in practical terms I ran into the problem that D&D used to be very precise about things like costs and times for these things, whereas now they’re leaving it up to my judgment, and I’m just not sure how to apply it. \$\endgroup\$ Commented Oct 9, 2014 at 2:21
  • \$\begingroup\$ Concerning 3.: How do you determine whether the player has to repeat the same action and hence pay the cost multiple times if not by rolling dice? \$\endgroup\$ Commented Jan 8, 2020 at 19:42

One possible consideration is that the roll is not to see if the rogue lucked into opening it, but into whether the rogue found that the lock was the kind of lock he knows how to open (or, the internals are in such a way that he's able to manipulate them). This is effectively in the camp of 'fail once, don't reroll', but a rationalization behind that. Your rogue rolls, he fails, oops, that lock isn't a kind he knows how to open.

This largely depends on whether your group is the sort that likes to problem solve; in this case, for example, finding that the door isn't openable means they have to find a different solution, which might be interesting or might be tedious. If it's tedious to your group, then the kill-a-hostage answer is probably better.

  • 4
    \$\begingroup\$ I like that, in that it's more plausible that a rogue might be permanently stumped because a lock is outside his expertise than because the rogue broke the lock trying to pick it. On the other hand, one could combine the approaches, allowing each rogue one roll to see if he was fortunate enough to have training suitable for the lock in question, and allowing rogues who fail that the option of trying more "assertive" techniques which would have the possibility of jamming the lock. \$\endgroup\$
    – supercat
    Commented Oct 11, 2014 at 15:16
  • 1
    \$\begingroup\$ I like this, you're essentially turning the lock picking into something like a Knowledge check. Either you have the information or you don't, thinking about it for another hour isn't likely to cause you to suddenly know it. \$\endgroup\$ Commented Mar 9, 2018 at 14:29

The DMG is your friend.

Take a look at p.237, "Multiple Ability Checks." This describes multiple ways we might handle the player's desire to SPAM ability checks.

One is to gauge the time it might take to succeed by patient re-attempts, without actually rolling-until-success.

Another is to modify the approach, with the proviso that failed attempts may improve or drive down one's chances of later success.

Examples of each, with suggestions/examples, are given in the Dungeon Master's Guide.

The PHB is also your friend.

These answers/options don't have to seem to your players as if they've come out of nowhere (the DMG). They're also in the PHB, implicitly.

Remember that the fundamental flow of the game is "DM describes situation, player describes character's actions, DM adjudicates/describes results." (PHB p. 6, "How to Play") Also, "the DM calls for an ability check" and "the DM might call for an ability check using a particular skill." (PHB p.174, "Ability Scores")

Players don't call for rolls, they tell the GM what the character's up to. If they narrate a different approach or they say "I'm going to keep at this until I get it!" then that's where you get your answer.

  • \$\begingroup\$ My players make me want to take their dice away sometimes with their re-rolling. This makes it way better. Thank you. I have to find a better way to manage time in my campaigns so it can be used properly. \$\endgroup\$
    – Skathix
    Commented Jun 14, 2016 at 23:18

The Mythbusters technique

When players fail the first attempt at a skill check, give them a choice:

  • Try a different approach, or
  • Tell you how far they’re willing to go to make this approach work.

In the latter case, don’t roll the dice again. Just determine whether the approach is feasible within the players’ parameters. If so, then let them “fail forward” with a setback appropriate to the approach. Otherwise, the players will need to switch to a different approach, or accept worse consequences.

For example, if a player fails to wriggle out of manacles (DC 20 Dexterity check), they can:

  • Try to break the chains instead (DC 20 Strength check), or
  • Keep trying to wriggle free, so long as it doesn’t cause injury.

If you think the player could eventually get free safely, then you can just let them go, with some lost time as the price for failing forward. If not, then the player may need to accept a worse setback, like a minor wound, or an interruption from a guard on patrol.

Should the players try a few different things and fail at all of them, you can always return to the “fail forward” approach. For example, if the player in manacles fails the Dexterity check, then the Strength check, they can return to the “wriggle free, no matter what it takes” approach later.


As others have already said, rolling the dice should be interesting. This means that the consequence for failure should be meaningfully different to the consequence for success.

The issue you appear to be having is determining what that point of difference should be. And honestly? There's no "One True Way" to do that.

Some groups love the idea that, if they fail a lockpicking roll, a patrol of guards shows up to complicate things; It highlights that the character picking the lock is competent, but that sometimes things just don't go their way. Other groups feel that having a group of guards suddenly pop into existence just because the party locksmith has a headache stinks of railroading and breaks suspension of disbelief.

Some groups like the idea that failing a roll makes repeating that roll impossible, just because repeating rolls is boring; Others find that implausible, especially when success was theoretically possible, and the apparent difficulty of the task was readily apparent and within the actor's capabilities.

Some groups are willing to make exceptions for particular kinds of rolls, or even on a case-by-case basis, because it allows the group to define the setting in whatever way seems most plausible in the moment - and others value consistency, because it lets them make complex plans and have some guarantee that they'll succeed.

So, there's no single right answer. It depends on what the group prefers, and what the campaign's established conventions demand.

(Incidentally, it's worth noting that earlier editions of D&D do have the concept of "progress with a setback" - It's just that the setback is often "Trying again will require you to take another unit of time to do so." The "extra time" setback might not sound like it has any meaningful impact, but the risk that it might can be an interesting consequence in itself.)


Before you ask for a roll from a PC, have stakes in mind. (In some game styles, you actually communicate the stakes to the player).

For a combat roll, the stakes are easy. If they swing a weapon and miss, and/or don't kill the monster, the monster gets to attack the party back.

When climbing a cliff, having making a downside to failure is perhaps too easy (the character falls).

The player when they make a roll has a goal in mind. If the player succeeds, they should achieve that goal with minimal complications.

If they fail, you need to have a complication in mind.

The complication should ideally move the plot in an interesting direction. "It takes more time, but nothing much happens" is not interesting. "Reroll until you succeed" is also not interesting, even if there is time pressure, unless that time pressure is very visceral (like orcs shooting arrows at you, and you unlocking the door as you shoot back with your hand crossbow).

So failure should either be a fail forward, or force a different move from the players.

A fail forward is where you succeed, but some complication happens: You unlock the door just as some guards come around the corner, you unlock the door but your lockpick breaks off in the door, you unlock the door but it swings open and makes a loud noise, you unlock the door but slip & cut your hand leaving a gash, splattering blood around the door handle, you unlock the door but it takes longer than you expect, and you hear the sound of a hostage being killed inside, you unlock the door but it takes longer and the enemy reinforcements arrive.

Forcing a different move is the "no, you cannot unlock this door" option. You can phrase this as "there is something jammed in there. You think you could unlock it from the inside..." which suggests the move of "thief enters some other way, and sneaks around to the door from other side". Or "the door lock is jammed hopelessly. The wall does look climbable, however." which again suggests a move.

The suggesting of a move is optional, and may not go over well, but it does prevent failure from seeming to be insurmountable. You, as the DM, need to have 2+ alternative moves thought up if you want to force an alternative move on failure.

You can fold the two of them. Failure on the lock forces a different move. The different move is either the players thinking something up, or making a streetwise check to figure out a different way in. That streetwise check on success suggests X, which is a relatively problem-free way in (climb the side wall & sneak in that window, which from the bricks on the outside are positioned lead to a staircase to the inside of this door), and failure suggests Y, which is a way that is actually unwise (climb the back wall & sneak in that window, which actually has guards sleeping in it) but still a way forward.

They don't have to follow either suggestion from that second check. But by arranging it that way you can prevent the game from locking up due to a skill failure.

So, in actual play:

Player: I roll to unlock the door. Blast, a bad roll: 9?

GM: Sorry, the lock seems jammed. You cannot open it from this side. Make a streetwise check, DC 15?

Player: (happy to roll more dice) I got a 7. Streetwise is cha? A failure, even if I add proficiency.

GM: You from the layout of the wall, there is a staircase leading down from the roof in the SW corner of the building.

Player: looks at you suspiciously.

So the player's move was "unlock the door". It failed. You forced a different move: a streetwise check to find an alternative way in. You even told the player the DC.

They rolled and failed. You still provided an alternative way in, with no comments on it.

They can either go with your alternative way in (which obviously will lead to some issues), or they can think up yet a different way.

Some of this design comes from Dungeon World and other indie games.

For more "fail forward", "let it ride", "force a move", "setting stakes" are all strings that you can search for on google for more details about this kind of skill check model, many of which inspired D&D design.

  • \$\begingroup\$ Mm, while these techniques work well with some campaigns and playstyles, there are other playstyles and campaigns they won't work quite so well with. Some discussion of the pros and cons of each one might help the OP in deciding whether and when to use them. \$\endgroup\$
    – GMJoe
    Commented Oct 10, 2014 at 2:07

Sometimes it is helpful to attach consequences to an action, such as getting caught for taking to long to pick the lock or breaking the lock. However, I've found that other situations don't work well with this solution, if you maybe don't want the consequences to affect your players but don't want to allow indefinite checks.

I have taken to ruling that an ability check measures a PC's ability to accomplish something, instead of a single roll representing a single attempt. To use the lock-picking example, rolling a 10 on a DC 15 lock means that the sum total of the character's attempts to pick the lock were not sufficient. This works well with the idea of environmental change- finding a better set of lock picks or getting someone else's help would change the PC's ability enough to warrant another check.

There are also cases where this isn't necessary. For conversational skill checks, good role-playing by the DM will render repeated checks useless or impossible after several failures, either by having the PC leave the conversation or by no longer calling for persuasion rolls. In battle, the penalty of an action is sufficient to allow for repeat attempts. A player engaging in a Strength contest to grab an opponent's weapon could be repeated for multiple rounds, as the player would be losing their action for each of those turns.

No matter what, try to establish proper discourse in your campaign. Instead of players asking to roll persuasion or slight of hand, have them describe their course of action to you. You can start by asking for skill checks, but, after a failure or two, feel free to just describe continual failure without calling for a roll.

  • 2
    \$\begingroup\$ Welcome to rpg.se! Take the tour and visit the help center when you get a chance. This is a pretty good first answer. Perhaps it could be improved by giving example of situations where you would use each of the techniques you describe. Thanks for participating and happy gaming! \$\endgroup\$
    – linksassin
    Commented Jan 15, 2019 at 5:12

For me keeping the target number hidden from the players works. They still feel in control when I tell them to roll the dice, even though the outcome could already have been made up. If they roll low I could still tell them they successfully picked the lock, because it was fortunately very easy to pick. This way I keep control over the outcome, while the players feel they have influence.

Basically I'd try to avoid the metagame. In character, they can't know what the DC of an action is.

The fact they feel in control though they have none was proven to me in one of the first sessions we played from the red starter kit. I let them choose between two doors and they really spent 10 minutes arguing and using skill checks to figure out which door they wanted to open, which didn't really accomplish anything because door B led to an encounter I hadn't prepared, while the door A encounter was prepared. Whatever they picked, they would always get door A. They didn't know and had fun deciding anyway.

  • 4
    \$\begingroup\$ What's the point of having them roll at all though if you're just going to give the success no matter the result? \$\endgroup\$
    – wax eagle
    Commented Oct 9, 2014 at 13:25
  • \$\begingroup\$ The "red starter kit" is for a different game with different rules; crucially, neither edition I know of that has a "red starter kit" has a rule about progress with a setback. \$\endgroup\$ Commented Oct 9, 2014 at 13:44
  • 3
    \$\begingroup\$ @MeanGreen Do you have a suggestion for what to do when the player rolls an extremely low number, like 1 or 2? \$\endgroup\$ Commented Oct 9, 2014 at 13:53
  • 3
    \$\begingroup\$ The magician’s choice is a useful GM technique, but it isn’t always appropriate. \$\endgroup\$ Commented Oct 9, 2014 at 18:53

In this case I'd let them roll a bunch of times over and over until they succeed if there is no consequence, but you still want them to roll.

You might want to make them roll just to spook them. You might want to trigger a trap if they fail.

Historically this has been treated with the wandering monster approach, whereby they will be interrupted by a random creature nearby that is roaming the dungeon.

Assuming the lock is very high DC it's also possible the rogue will fumble and damage the lock, or his or her tools, which is another reason to want them to roll.

Lastly, they can always take 20, which has a much bigger time penalty, whereby they get a guaranteed 20 at the cost of a specific amount of time elapsing.

  • 3
    \$\begingroup\$ The question literally says there is no such thing as take 20 in 5th edition and that "rolling until you succeed" is to be avoided because it's a tedious exercise. \$\endgroup\$
    – user4000
    Commented Oct 10, 2014 at 17:54
  • 1
    \$\begingroup\$ Thanks for the input, and welcome to the site. Unfortunately, while this advice might work well for D&D3 players, it’s not in line with the current version of the game or what I’m looking for. \$\endgroup\$ Commented Oct 10, 2014 at 21:39
  • 2
    \$\begingroup\$ That's unfortunate. Rolling until you succeed isn't tedious if there's a consequence, e.g. wandering monster, which my answer quite literally says. Sorry I wasn't aware things changed in 5th edition, but who said you have to stick to the rules?! You can take 20 if you want to. You're the DM, you can make up the rules if they're sensible, override them, change them and throw them away. The last thing you want to do in this situation is spend 15 minutes deciding how to proceed. This community isn't for me if it's like a lawyer's office calculating what is acceptable or not. Just enjoy the game! \$\endgroup\$
    – NibblyPig
    Commented Oct 13, 2014 at 8:50
  • \$\begingroup\$ It seems to me like the first sentence doesn't jive with the rest of the answer. \$\endgroup\$ Commented Mar 27, 2018 at 1:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .