D&D 5E Can you retry a failed skill check? How long?

[Psikerlord#]

I believe this sort of question is best left to DM discretion, depending on the circumstances and the preferences at the table. There is nothing I personally hate worse than a declaration from a player that "I keep searching till I roll a 20, just in case" scenario.

For me it all depends on the situation. If youre trying to bust a door down, I will let you keep making checks, usually because I like the idea of the door getting weaker and weaker as you hammer away on it. But say a knowledge check, or a search for secret doors check? I will only allow that once, until the circumstances change (they get new info, or whatever).

I suspect the passive check rule is might be supposed to cover the "I keep doing whatever till I roll a 20" situation. But for me I prefer leaving it to DM discretion, based on all the circumstances and table preference. That way the DM can choose whatever suits best for the scenario at hand - one roll only, a few rolls, take 10 (passive check), take 20, or whatever other resolution mechanic he wants to employ.

 

[Li Shenron]

I like that there is no official rule about this too.

Personally I often think that allowing retries "because they are realistic" conflicts with the total lack of realism of skill checks being random in the first place. If you are trying to break down a door or solve a puzzle in real life, there is really no randomness in whether you are inherently capable of succeeding or not. What is "random" from your perspective, is rather external factors you have no controls over. External factors include for example mental distractions, physical intrusions, and variations of the target (i.e. "this one wooden door happens to be casually sturdier/weaker than how wooden doors normally are").

Retrying cancels the effect of those external factors, so all you have to ask yourself is whether you want to have those factors in your game or not, and mostly it becomes a matter of narrative, because those factors randomize your story. This is what I believe was the original reason for including dice rolls in D&D 40 years ago: Gygax and Arneson decided they wanted the story to be (partially) randomized.

So if you allow infinite retries, just know that you are removing randomness from that type of story element, being it "breaking down doors", "picking locks", "solving puzzles" or whatever (most likely, you are not allowing retries on everything, so other story elements remain random of course). If this is your preference, then really you can save time and effort by just using Passive Checks (i.e. Take 10 or Take 20) all the time, without even asking for a dice roll.

Otherwise, if you do ask for dice rolls, it's better to disallow retries, and possibly even disallow other PCs to try the same thing. Once again, you'll save time and effort.

Intermediate solutions such as "only N retries" or cumulative penalties don't really remove randomness, they simply increase success chances, which is something you can also achieve by lowering DC if this is the intent... but it's not usually the intent, so IMHO intermediate solutions are again a waste of time and needless complexity.

Situational solutions instead can be good (e.g. conditioning retries to being unthreatened), but you have to really understand what you're doing, because inevitably you'll end up allowing these to some skills but not to others, and there are skills which you rarely use while threatened or it is at least uncommon to be threatened etc... That is just to say, you'll be effectively again removing randomness to some story elements while you still have it for other story elements!

And that's what you really should think about in the end if you want to end with satisfying results: what parts of the story you want random or deterministic, and not "realism".

EDIT: to rephrase my last sentence... you should ask yourself stuff like "do I want lockpicking to be a randomized part of the story, or do I want it to be deterministic?" and depending on your answer, see what rule you should use (do not instead ask your self "what is the most realistic rule for lockpicking")

 

[Evenglare]

A couple of people have said this I believe, but whatever, I'll say it again. I don't allow skill checks to be repeated in about 90% of cases. If you let someone repeat over and over there's really no point in having skills because they will eventually succeed at everything. Then the comeback to this argument is that you can have them get attacked or something because they ARE wasting time, my response is that you can throw monsters at them because of failed checks only so many times. It gets boring, and if you just let them take as much time as you want, we are back around to the first point. So whenever skill checks occur in my game, when the use that check that's it. That is as good as any character can do at that moment in time. The circumstances must be changed before another check can be attempted (for example if you fail lock picking you can come back when you have a spell cast on you or you find some crazy tools that help you etc etc.)

 

[Paraxis]

So it looks like we have 3 different camps on this.

1. Only roll if failure causes another negative effect, so repeating the roll is mostly pointless, you failed to pick the lock in time so the troll caught up to you, or you failed to pick the lock and the guards on the other side are alerted.

2. Allow only one check, the door is locked, the rogue tries to pick it and fails, he can't try again. So now the mage tries, then the cleric, then the fighter, if they all fail to pick the lock they all now stand in a conga line to try and smash down the door finally the wizard gets lucks and breaks the door down. This was all done for no real reason and took 30 minutes of time at the table and just upset everyone ohh and everyone laughs at the fighter for being weaker than the wizard.

3. Allow unlimited checks, the rogue just uses up time in game and in the real world waiting for the die roll to get to be lucky enough. What was the point of rolling again?

Obviously I think the first way is the best, I don't see how there could ever be a problem with it and see glaring issues with the other ways.

So roll only when it matters to the story, don't just put boring locks and walls in your adventure in the first place, and have failing the roll mean something.

 

[delericho]

I don't have a RAW answer. My inclination is to only allow retries if the situation has changed significantly (better tools, greater skill, more assistance, or something like that - with combat generally allowing retries because it's constantly changing).

However, I'm also of the view that you should only be rolling if it 'matters' anyway - if it's something the character could just try again and again until they get it, I wouldn't bother with the roll anyway. Give them the pass and move on.

 

[77IM]

How about this? The first attempt takes an "Action".
To try again takes 1 minute.
To try after that takes 5 minutes.
To try after that takes an hour.
To try after that takes 5 hours.

I did something just like this in 4e and it worked very well. The time required for many tasks is up to the DM already, and the justification is that subsequent attempts take longer because you're trying different things (naturally, you tried the quickest things first, but they did not work).

I think it strikes a nice balance between never retrying (which seems unnatural) and consequence-free retries (which are boring).

 

[jamesmanhattan]

FWIW the Playtest had some guidance on this subject that could be found in the 5E DMG; DM Guidelines Pg.4 Multiple Ability Checks: Sometimes a character fails an ability check and wants to try again. You have a couple of options in this case. In most cases, the character can simply try again. The only real cost is the time it takes. The character keeps trying and, after enough time passes, eventually succeeds. To speed things up, you can assume that a character can automatically succeed at a task if he or she spends twenty times the normal amount of time needed to complete it. This exception does not allow a character to turn an impossible task into a successful one. In other cases, the first failure renders subsequent ability checks impossible. For instance, a rogue tries to trick a town guard into thinking that the group members are undercover agents of the king. The rogue loses the contest of Charisma (Deception) against Wisdom. The same lie told again clearly won’t work.

Thanks, that makes sense and is easy. It seems there is a huge variance from individuals in answers to this.

I looked in the PHB at climbing, they make no mention of "if you fail a climb check by 5 or more, you fall". However, there is that line in the Starter Set adventure on one of the walls. It looks like skill checks are just really left to the DM or the adventure designer. They really did leave it all up to the situation. Guidelines, which often become "hard and fast rules", are left out of the PHB.

Thanks for the answers. I think if it is ever not specified I will simply use "You can retry. Which takes 20 times as long. If you fail by 5 or more something bad happens, lock jams and must be approached differently, you fall, etc." Hopefully adventures are written to not include too many boring old locked doors, and the only climb checks I'll ask for are on dangerous old crumbly walls above spikes, not plain old stone ones.

Maybe the DMG will have more answers.

 

[Manbearcat]

This is why I prefer conflict resolution systems now and pretty much universally avoid (process simulating) task resolution systems. "Fail and something interesting always happens (specifically related to the stakes or the overarching goal) to fill the PCs lives with adventure and conflict" (possibly or likely unrelated to the process of the task) is awesome. I think the base DCs are supposed to be 10, 15, 20? So long as the system/PC-build math holds together, they could create a Dungeon Worldesque conflict resolution module whereby on a 10 or less, * GMs invoke something severe as an immediate complication or something ominous lurks or is portended. ** An 11-15 means success with complications (a hard choice or some kind of fallout despite more or less accomplishing your goal, and *** 16+ means you get the job done, no questions asked.

Fail to pick a lock when trying to infiltrate the Duke's manor and seize incriminating documents:

* The door opens easily as you attempt to unlock it...a guard is dead in his own pooling blood on the floor just inside the chamber. When you enter, two shadowy figures throw daggers your way while a third slips out the window with the documents you seek!

or

** The sounds of the machinery of the pin tumbler lock are as thunder to your ears as you pick it, clearly a clever intent by the designer. You can almost hear it over the sounds of the dice-playing guards you bypassed on your way here. As you feel the lock give way, a voice behind you "did you hear that?..."

or

*** You're greased lightning. The lock melts away from the door with your adept quickness and care. There isn't a sound from the room beyond and you can see the torchlight on the wall within is undisturbed by any nearby motion.

Slightly more compelling gameplay than:

"I pick the lock."

"You fail."

"I try again...crap...ok, again...and again...4th times a charm?..."

"...oh hell you pass or just take 20...moving on..."

 

[evilbob]

If you allow constant rerolls, DCs for rope, manacles, locks, etc are far too low.

This is the main issue with the "taking 20" variants people are mentioning. If you just assume someone will get it eventually, then all manacles are 100% pointless, and the normal locks you can buy are 100% useless. Why would anyone use a normal lock on an unguarded dungeon door if literally anyone can pick it in about 20 minutes? Certainly there's a use for someone who has a good change to beat a DC 15 check in 6 seconds, but if anyone else can do the same thing in ~20 minutes, it definitely diminishes that role.

Also why would you ever tie up prisoners or capture anyone? Anyone can escape anything eventually so it just doesn't matter. It also means that all mundane jails are pointless and you should never happen upon a skeleton in manacles because they could have gotten out with half an hour's worth of effort. That's right: a halfling child can physically break manacles given 20 minutes.

While I agree with the idea that most things just don't need to be rolled if you expect the PCs will get it eventually and time isn't an issue, it also means that at least a few game mechanics are pointless.

Personally I am starting to like the idea mentioned in other threads and above that each retry takes a longer time increment (1 turn, then you have to try for another minute to roll again, then an hour, then a day, etc. - something like that). This is a complete houserule but it at least gives some meaning to locks and manacles, and it represents the increasing amount of time invested in the task. You'd have to have an end point as well, at which point it's simply impossible and cannot be done by that person, ever. (Or maybe the time increments simply increase to the point at which they would outlast that creature's lifespan.) The nice thing is that it also allows players to judge an increasing time risk vs. what they are trying to do.


Anyway, the point of this thread is that there is no one right way, and every idea has issues one way or another, so yeah: do what you think is best.

 

[Agamon]

So roll only when it matters to the story, don't just put boring locks and walls in your adventure in the first place, and have failing the roll mean something.

This isn't really about rerolling, it's more about to roll or not to roll. But I agree. I certainly strive to keep things interesting. If something interesting is on the other side of a door, and making it difficult to get through that door does nothing to make it more interesting, then they get through the door. If it does make things more interesting, then they roll.

I'd hate to play in a game were I was forced to pointlessly roll the dice.