YARACT (yet another repeated ability checks thread)

[terodil]

Hi folks,

I'm once again stuck on the 'repeated ability checks' issue. There have been a lot of great threads on here already touching on this issue, but I don't want to cast Raise Dead Thread, so apologies for this new one.

We all know that the DMG says you can take 10x the usual time to auto-succeed on an ability check, provided that you could beat it with a perfect roll. It's very similar to the old 'take 20' rule -- a term which I may use here for brevity, even if it's not exactly the same. We all know what we're talking about.

Note that taking 20 is, unlike the 'ability score-5' rule, not optional. It effectively means that even a completely unskilled person with a governing attribute modifier of -1 will auto-succed at an almost-difficult task, provided there is no pressure of any kind (time, danger involved in failing, etc.). I think this is ridiculous for three reasons: Firstly, it makes no narrative sense that Mr. Klutz, who has only ever used keys to open locks, can suddenly pick an elaborate lock -- reliably! -- by taking 10x the time; more time to poke his index finger at the lock won't help. Secondly, by comparing 20+mods to a fixed DC, the relative weight of the mods (usually significantly <10) is cheapened. Thirdly, it eliminates the idea of luck entirely while, at the same time, taking the sting out of failed rolls to the point of almost making them inconsequential.

Now there are several ways to solve this. A DM can always introduce a chance-to-fail that prevents taking 20, i.e. a lock can jam, or a ledge may become brittle and collapse; or he may introduce a doom clock. Given that there is a non-optional rule regarding repeated tries, though, this seems to be intended as an exceptional case, not as the default.

So I'm considering:

Option A: Players do not get to take 20. Instead, they may expend 5x the usual time to try once more to beat the DC, this time with advantage. This roll represents the best they can do given current circumstances. This would reintroduce a modicum of luck. Unfortunately, it does nothing to improve Ms. Expert's expected performance relative to Mr. Klutz's. I'm also not sure if this would negatively impact (cheapen) any class features. Does it?

(Additional) Option B: Instead of rolling 1D20, roll 2D10. This would change the probability distribution to favour average results over extreme ones, and thus increase the impact of character builds on expected performance. It would also make high DCs even more difficult to beat, though -- they should probably be reduced a bit to compensate.

I'm sure there's a perfect solution around somewhere that I've not thought of. I've read one suggestion to reduce all DCs by 5 and to roll 1D10 instead of 1D20, which was interesting but introduces too many (new) problems, imo. How would you handle this, given the three issues I have with repeated ability checks per RAW?

We have one more thing to consider: What exactly does failing a check mean, in this context? Obviously the person doing the check can't do 'it'. But why? The official material generally seems to singularly point at the individual attempting the check: At this very moment, the attempt simply was not good enough, but the person could have beaten it (this is very apparent in the knowledge proficiency descriptions: '[Your] ability to recall lore about X', not 'your knowledge of X'). Another position to take, however, would be that the world state simply didn't allow success. The lock the rogue has failed to pick is one-of-a-kind. By eliminating the roll-20, a failed roll could effectively change the world state from 'lock that is pickable' to 'lock that is not (and was never) pickable for the rogue'. Similarly, a failed knowledge check (yes, I know, more precisely: an ability check with an applicable knowledge proficiency) could change the world state from 'the priest knows this subject matter, but may or may not recall it' to 'the priest doesn't know the subject matter at all, so he cannot recall it'. I think this would be a fair and narratively interesting re-interpretation of failed rolls. What do you think?

[MaxWilson]

The way I run lockpicking is that:

(1) If you beat the DC, you open the lock.

(2) If you fail by 10 or more, you are stymied and cannot try any more until you learn more about locks (e.g. talk to a locksmith). This is similar to what you talk about in your final paragraph, about world states. The interpretation here is that you just have no idea what to do about this kind of lock and you have exhausted all of your best ideas.

(3) If you are not stymied, you can retry at the next opportunity (e.g. next round).

I like the probability curve this gives for skilled and unskilled lockpickers against locks of DC 10, 15, 20, 25.


Similarly, a failed knowledge check (yes, I know, more precisely: an ability check with an applicable knowledge proficiency) could change the world state from 'the priest knows this subject matter, but may or may not recall it' to 'the priest doesn't know the subject matter at all, so he cannot recall it'. I think this would be a fair and narratively interesting re-interpretation of failed rolls. What do you think?

Yup, sounds reasonable, and very much in keeping with how random rolls were viewed in AD&D: it's not just "did you succeed on this attempt?", it is "can you succeed in this endeavor given who you are right now?" I didn't play 3E but I gather the idea of re-rolling repeatedly until you succeed was pretty much a 3E-ism which has continued into 5E. It's not an entirely bad approach, but sometimes it's appropriate, sometimes it's not.

[terodil]

The way I run lockpicking is that: [...] (2) If you fail by 10 or more, you are stymied and cannot try any more until you learn more about locks (e.g. talk to a locksmith). This is similar to what you talk about in your final paragraph, about world states. The interpretation here is that you just have no idea what to do about this kind of lock and you have exhausted all of your best ideas.

Thanks for your reply, MaxWilson. Just to clarify, are you not using the take 20 rule at your table then? I very much like (and use) the idea of degrees of failure and success for ability checks already, so it makes sense to tie this into it.

Note, though, that I was also particularly thinking of knowledge checks (language!). I really dislike 'unlimited attempts to remember'. Heh. Crossed the time streams. Thanks for your reply!

[Mr.Spastic]

I always played it that you had to be proficient in the tool or skill that the check is being made for. So if you don't have a proficiency you can't do the thing. This prevents people from doing the Mr. Kluts picks the DC fifteen lock with his -2 dex bonus and no proficiency.

I also run that you have to be proficient with the most tools to even be able to make the check. This means that you have to have a thieves tool's proficiency to attempt lock picking. If you don't there are other things you can do like chick the door down or going through a window. This works in my games because most of the people who want to do criminal stuff find a way to get thieves tools anyways.

The few exceptions that I have are...

Cook's utensils, because it's just cooking.

Painter's Supplies, same reason.

Land Vehicles, because it's not super hard to direct a cart

Games and Instruments, because just because you aren't proficient doesn't mean you can't play.

[Ganymede]

Note that taking 20 is, unlike the 'ability score-5' rule, not optional. It effectively means that even a completely unskilled person with a governing attribute modifier of -1 will auto-succed at an almost-difficult task, provided there is no pressure of any kind (time, danger involved in failing, etc.).

This is not true.

The quasi take 20 mechanic only comes into play if a PC has an opportunity to make a check to begin with, which is gated by the DM. Before any of this happens, the DM must determine whether a task automatically succeeds, automatically fails, or if there is uncertainty in the outcome. If the DM says, "you fumble around with this unfamiliar set of picks and files and, ultimately, are unable to make any progress" and doesn't call for a check, the take 20 option is no recourse for that PC.

[solidork]

An easy fix to the "Anyone can open this lock with enough time" is to restrict the ability to take 20 to people have proficiency in a relevant skill/tool.

This doesn't help for situations where there really isn't an applicable proficiency, but it still makes sense for someone who spends enough time should be able to succeed.

[mer.c]

An easy fix to the "Anyone can open this lock with enough time" is to restrict the ability to take 20 to people have proficiency in a relevant skill/tool.

This doesn't help for situations where there really isn't an applicable proficiency, but it still makes sense for someone who spends enough time should be able to succeed.
I've been using this as well. It's really nice in that it emphasizes the utility of proficiency (which is quite low at lower levels) at the same time.

[terodil]

This is not true.

The quasi take 20 mechanic only comes into play if a PC has an opportunity to make a check to begin with, which is gated by the DM. Before any of this happens, the DM must determine whether a task automatically succeeds, automatically fails, or if there is uncertainty in the outcome. If the DM says, "you fumble around with this unfamiliar set of picks and files and, ultimately, are unable to make any progress" and doesn't call for a check, the take 20 option is no recourse for that PC.
Thank you for this comment, but it's redundant. I knew this pedantry was coming (after all, I tend to do it too), so I wrote right in the beginning:

We all know that the DMG says you can take 10x the usual time to auto-succeed on an ability check, provided that you could beat it with a perfect roll. It's very similar to the old 'take 20' rule -- a term which I may use here for brevity, even if it's not exactly the same. We all know what we're talking about.
Emphasis added. If the DM decided that there is no roll, you could not beat 'it' with a perfect roll, ergo there is either auto-success or auto-fail, and take 20 doesn't even enter the picture. Please, let's explicitly look at the cases where the DM decided to allow rolls to begin with. This thread is about such cases and repeated attempts.

[Ganymede]

Please, let's explicitly look at the cases where the DM decided to allow rolls to begin with. This thread is about such cases and repeated attempts.

LOL, those situations are your own damn fault.

You want a thread talking about how DMs can walk back screw-ups? I mean, come on: the parameters of this thread are to deal with the paradoxical situation where the DM says "Yeah, I guess that's possible. Roll" and "Wait, that shouldn't be possible." ?

[terodil]

LOL, those situations are your own damn fault.

You want a thread talking about how DMs can walk back screw-ups?

What?

How is it a screw-up to put something into the party's way that has a DC at least one of them could theoretically beat and to allow them to roll? Because that's when the take 20 rule possibly enters the picture, and it's about this that I want to talk.

How you come up with the idea that such a situation is per se a DM screw up is beyond my current understanding.

Edit to account for your edit: You are misunderstanding the situation. We are not talking about flat-out impossible challenges, we are talking about challenges that ARE possible to beat. The question is if you restrict the players to just roll once or twice and live with the result, or if you allow them to roll-20 to force success as the DMG seems to demand.

[Pex]

LOL, those situations are your own damn fault.

You want a thread talking about how DMs can walk back screw-ups? I mean, come on: the parameters of this thread are to deal with the paradoxical situation where the DM says "Yeah, I guess that's possible. Roll" and "Wait, that shouldn't be possible." ?

There's also the problem of my character being able to do stuff depends on who is DM that day, but some people would yell at me for bringing it up again.

[Ganymede]

What?

How is it a screw-up to put something into the party's way that has a DC at least one of them could theoretically beat and to allow them to roll? Because that's when the take 20 rule possibly enters the picture, and it's about this that I want to talk.

How you come up with the idea that such a situation is per se a DM screw up is beyond my current understanding.

It is actually really easy to understand. By allowing the PC to roll in one of those situations, you've put yourself in a situation that you yourself describe as "ridiculous."

Why would you do that to yourself? Why would you do that to your group's game?


Edit to account for your edit: You are misunderstanding the situation. We are not talking about flat-out impossible challenges, we are talking about challenges that ARE possible to beat. The question is if you restrict the players to just roll once or twice and live with the result, or if you allow them to roll-20 to force success as the DMG seems to demand.

"The D&D rules help you and the other players have a good time, but the rules aren't in charge. You're the DM, and you are in charge of the game." - Dungeon Master's Guide.

[terodil]

It is actually really easy to understand. By allowing the PC to roll in one of those situations, you've put yourself in a situation that you yourself describe as "ridiculous."

Why would you do that to yourself? Why would you do that to your group's game?

Could you please define what you mean by 'those situations'? I think there's a rather significant misunderstanding here.

Edit: And I'd be grateful if you took edits to new posts instead, it's making replying really difficult.

[Willie the Duck]

Thanks for your reply, MaxWilson. Just to clarify, are you not using the take 20 rule at your table then? I very much like (and use) the idea of degrees of failure and success for ability checks already, so it makes sense to tie this into it.

Not MW, but I see where it is not pertinent. In this situation, the take 20 rule would no longer apply, because the situation no longer meets the 'provided there is no pressure of any kind' clause (consequence of failure being a form of pressure).

Honestly, the idea of a single, simple ruleset being able to arbitrate (almost) all non-combat, non-magic task resolutions is pretty pie-in-the-sky. Especially if trying to craft a formulaic multiplier of the duration of 'single attempts' that would equate to having exhausted one's options. If it is possible at all for me to dead lift a given weight, I probably should be able to do it within x multiples of a single attempt (maybe not if I might strain myself, but there's another example of consequences invalidating the take20 rule), while picking a lock... I'd say if it is possible at all, it would be a completely different y multiple of a single attempt.

[ChildofLuthic]

Hi folks,
Note that taking 20 is, unlike the 'ability score-5' rule, not optional. It effectively means that even a completely unskilled person with a governing attribute modifier of -1 will auto-succed at an almost-difficult task, provided there is no pressure of any kind (time, danger involved in failing, etc.). I think this is ridiculous for three reasons: Firstly, it makes no narrative sense that Mr. Klutz, who has only ever used keys to open locks, can suddenly pick an elaborate lock -- reliably! -- by taking 10x the time; more time to poke his index finger at the lock won't help. Secondly, by comparing 20+mods to a fixed DC, the relative weight of the mods (usually significantly <10) is cheapened. Thirdly, it eliminates the idea of luck entirely while, at the same time, taking the sting out of failed rolls to the point of almost making them inconsequential.

That's not what that means at all. The idea is if a perfect roll would let the character succeed, than you can just handwave all that rolling by saying "your character tries until he succeeds." So Mr. Klutz doesn't get to roll, because even on a nat 20 - even if he does the best he could possibly do - he would still fail.

[Cynthaer]

[...]I think this is ridiculous for three reasons: Firstly, it makes no narrative sense that Mr. Klutz, who has only ever used keys to open locks, can suddenly pick an elaborate lock -- reliably! -- by taking 10x the time; more time to poke his index finger at the lock won't help. [...]

It's probably too late, but I'd like to establish something before the discussion goes off the rails.

Lockpicking is a bad example to use here, because it has special restrictions related to possessing and being proficient with thieves' tools in order to even attempt picking a lock. There are arguments to be had over the precise nature of these restrictions, but we can avoid that argument entirely if you choose a more generic scenario.

Regarding the substance of the OP, I feel you may be coming at this backwards.

You're starting with "players are allowed to 'take 20' under certain circumstances, but sometimes that doesn't make sense"—and concluding that the "take 20" clause is the problem.

But you're ignoring the second half. To quote the DMG directly:


Using Ability Scores

When a player wants to do something it's ofter appropriate to let the attempt succeed without a roll or a reference to the character's ability scores. [...] Only call for a roll if there is a meaningful consequence for failure.

When deciding whether to use a roll, ask yourself two questions:

Is a task so easy and so free of conflict and stress that there should be no chance of failure?
Is a task so inappropriate or impossible — such as hitting the moon with an arrow — that it can't work?

[...]

Multiple Ability Checks

Sometimes a character fails an ability check and wants to try again. In some cases, a character is free to do so; the only real cost is the time it takes. With enough attempts and enough time, a character should eventually succeed at the task. To speed things up, assume that a character spending ten times the normal amount of time needed to complete a task automatically succeeds at that task. However, no amount of repeating the check allows a character to turn an impossible talk into a successful one.

In other cases, failing an ability check makes it impossible to make the same check to do the same thing again. For example, a rogue might try to trick a town guard into thinking the adventurers are undercover agents of the king. If the rogue loses a contest of Charisma (Deception) against the guard's Wisdom (Insight), the same lie told again won't work. The characters can come up with a different way to get past the guard or try the check again against another guard at a different gate. But you might decide that the initial failure makes those checks more difficult to pull off.

So you've got a number of tools available as a DM to determine how a scenario plays out. You're not at the mercy of the "take 20" rules, because you are the one who decides if taking 20 is appropriate!


Firstly, it makes no narrative sense that Mr. Klutz, who has only ever used keys to open locks, can suddenly pick an elaborate lock -- reliably! -- by taking 10x the time; more time to poke his index finger at the lock won't help.

If it truly "makes no sense" for Mr. Klutz to succeed at picking a lock, then it's an automatic failure.


Secondly, by comparing 20+mods to a fixed DC, the relative weight of the mods (usually significantly <10) is cheapened. Thirdly, it eliminates the idea of luck entirely while, at the same time, taking the sting out of failed rolls to the point of almost making them inconsequential.

If the "take 20" approach is coming up so much that it's devaluing the impact of mods, then it's worth asking why you keep presenting obstacles to your players where (A) they require ability checks to solve, but (B) there is "no pressure of any kind" to actually solve them. If there's no stress or consequence for failure, should these just be automatic successes so the party can get to the real challenges? If all the challenges are like this, wouldn't you be better off creating obstacles and scenarios with actual stakes?


We have one more thing to consider: What exactly does failing a check mean, in this context? Obviously the person doing the check can't do 'it'. But why? The official material generally seems to singularly point at the individual attempting the check: At this very moment, the attempt simply was not good enough, but the person could have beaten it (this is very apparent in the knowledge proficiency descriptions: '[Your] ability to recall lore about X', not 'your knowledge of X'). Another position to take, however, would be that the world state simply didn't allow success. The lock the rogue has failed to pick is one-of-a-kind. By eliminating the roll-20, a failed roll could effectively change the world state from 'lock that is pickable' to 'lock that is not (and was never) pickable for the rogue'. Similarly, a failed knowledge check (yes, I know, more precisely: an ability check with an applicable knowledge proficiency) could change the world state from 'the priest knows this subject matter, but may or may not recall it' to 'the priest doesn't know the subject matter at all, so he cannot recall it'. I think this would be a fair and narratively interesting re-interpretation of failed rolls. What do you think?

This is a pretty standard approach. The DMG doesn't really take a strong position, and the vast majority of DMs don't really use a singular Theory of Ability Checks such that they always have the same narrative implications.

More common is an ad hoc approach that interprets rolls in whatever way makes the most sense (or is the most interesting) in the moment. Sometimes a falied Intelligence (History) check means you've never heard of the lost orc city. A failed lockpick check to open manacles in mid-combat probably means you can try again next turn using your Action, while a failed check to open a door out of combat might mean you broke the lock. Sometimes a low roll to kick down a door means you had a bad angle, while sometimes it means the door was reinforced with steel.

Many people feel very strongly about precisely how checks should be used and interpreted, and more power to them. My point is that the DMG doesn't take a hard position beyond the guidelines I quoted above, and those are mostly focused on how to use them to make the game flow well.

TL;DR: As a DM, use your gut. If a check seems pointless, skip it. If success doesn't make sense, don't allow it. If your obstacles lack time constraints or consequences for failure, fix the obstacles, not the skill check rules.

[Keravath]

My approach is the following ..

1) If the task requires specific tools or skills then the character needs to have those pre-requisites in order to attempt the task (e.g. opening a lock may require proficiency with thieves tools AND having a set of tools ... otherwise you can't even attempt to open the lock ... if you have proficiency but only improvised tools then I might apply disadvantage ... similar if you have tools but no proficiency ... depending on the lock or task)

2) If the character is capable of attempting the task and if there is no limiting factor (time, guards, ticking clock, consequence of failure), I check the characters passive skill first (equivalent to take 10) ... if the passive skill is high enough then I usually narrate the success of the action they take without requiring any die rolls.

3) If there are limiting factors, then depending on the nature of the factors, I will either check the passive skill or have the player roll a skill check.

4) If there is a task for which the character has the necessary skills and for which there are no time limits or consequences of failure bu the task is more difficult than their passive skill then I will just narrate the successful result as taking a very long time making sure that the character actually wants to spend what could turn into hours completing the task.

Basically, if the DM puts in a task which the character will be able to succeed on eventually, for which they have the needed skills/proficiency/tools and for which there is no time limit and no significant consequences for trying over and over ... then two things happen ... the character succeeds and you ask the DM why they put in a challenge that the character would succeed on eventually ... unless the whole reason of the task was to specifically delay the character/party in terms of game time.

[terodil]

That's not what that means at all. The idea is if a perfect roll would let the character succeed, than you can just handwave all that rolling by saying "your character tries until he succeeds." So Mr. Klutz doesn't get to roll, because even on a nat 20 - even if he does the best he could possibly do - he would still fail.
My assumption/example was: Mr. Klutz has -1 on his governing attribute, which would mean that if he rolled 20, he could beat DC19. If the conditions for the take-20 rule in the DMG are met (i.e., he gets to roll in the first place, there's no time pressure or significant penalty for failure, yadda yadda), then I'd have to handwave it and give him an autosuccess -- reliably, every time as long as the take-20 rule applies and the DC does not exceed 19.

So, looking back on this thread so far, I'd like to thank everyone for their input! I've picked up the following additional ideas from this thread:
* Be restrictive with when you allow characters to roll at all vis-a-vis to proficiencies. I.e., don't allow them to roll at all if you as DM think that proficiency is required.
* Be restrictive with take-20. Even if you allow characters to roll without proficiency, only allow them to take 20 if they do have it.

Thank you, Mr.Spastic, solidorc, and mer.c for this input. I hadn't really considered gating take-20 via proficiency, but it seems logical to me. I feel a bit more confident now that you all say you deviate to some degree from RAW. I know it's in the purview of the DM, but still... it takes a bit to just say 'Here's how I'm gonna do this...'

I still don't like the rule, though. The result of dice rolls should matter and not be replaced by the slot machine mentality of 'I know I'm going to hit the jackpot some time, so might as well award it to me now', even if you possess the applicable proficiency. I think I'll do the following:

* I'll be more restrictive in allowing characters to roll in the first place, and emphasise proficiency in this decision.
* Under the same conditions as the take-20 rule in the DMG, characters will be allowed one reroll with advantage, and will then have to live with that roll. No auto-success. If you fail a knowledge check, you get to try your best to see if you remember (advantage), but if that fails you will need to seek out a library or just get that knowledge from somewhere else. Similar for lockpicking etc.

[strangebloke]

The DM has to decide when a check is called for at all. There are a few defined calls he can make:

1. Automatic Success. Its a normal padlock. The character has the tools and the proficiency. Clearly a normal padlock isn't a real test for a skilled thief.
2. Impossible. It's a complex combination lock, and the character doesn't have the tools and/or the proficiency. What, is he accidentally a safecracker all the sudden?
3. Automatic, given enough time(take 20): If time matters, the PC has to make a check to rush it. If time doesn't matter, no check required. It helps to introduce some kind of time tracker in a dungeon environment.
4. Possible, but only one chance: The PC makes a check, but if he fails the lock breaks. He then cannot pick it.

You must communicate the results of failure or success as well as the task's approximate difficulty to your players before they commit to the action.

There's this idea that you have to roll for everything, and that you can roll for everything, that's completely toxic to the game. Let your players climb a gosh-darned tree.

[terodil]

Good lord, swordsaged by two excellent replies -- thank you, Cynthaer and Keravath!


Lockpicking is a bad example to use here, because it has special restrictions related to possessing and being proficient with thieves' tools in order to even attempt picking a lock.
You are completely correct, I had actually forgotten about the requirement for proficiency with thieves' tools. Apologies.


But you're ignoring the second half. To quote the DMG directly [...]

If it truly "makes no sense" for Mr. Klutz to succeed at picking a lock, then it's an automatic failure.
A valid point!

I think the issue is with 'impossible'. To belabour the rather unfortunate lockpicking example again and trying to salvage it just for the discussion of this word 'impossible', I'd be happy to allow an exceptionally bad rogue to open an extremely difficult lock if he hit the nat 20 on his one-and-only roll. It's possible for him to open it, but it takes a LOT of luck. Basically, fate intervened in his favour. However, I dislike the take-20 rule because it suggests that as long as he can beat the DC with a nat 20, you just pluck him in front of the lock and let time sort it out -- the problem is that this kind of resolution is 100% reliable. This is not an extraordinary stroke of luck like the nat 20 he rolled, it's an automatism that kicks in as soon as the situation qualifies for take-20. Perhaps I'd be happier with take-20 rather being a take-18 or so...

That said, I'm thankful for the reminder that it's up to me to decide if the situation warrants take-20. I'm just not looking forward to argue this with players that point at that passage in the DMG and go 'but but everybody does it this way! You're a **** DM, I'm leaving and I hope nobody ever plays at your table again'.


If the "take 20" approach is coming up so much that it's devaluing the impact of mods, then it's worth asking why you keep presenting obstacles to your players where (A) they require ability checks to solve, but (B) there is "no pressure of any kind" to actually solve them. If there's no stress or consequence for failure, should these just be automatic successes so the party can get to the real challenges? If all the challenges are like this, wouldn't you be better off creating obstacles and scenarios with actual stakes?
That's all very valid. I'm asking chiefly for clarity and peace of mind, not because I keep running into this. In fact, I hope not to.


TL;DR: As a DM, use your gut. If a check seems pointless, skip it. If success doesn't make sense, don't allow it. If your obstacles lack time constraints or consequences for failure, fix the obstacles, not the skill check rules.
Aye. Thank you. As I said though, if you are a relatively inexperienced DM like I am and face players who know the material inside-out by heart, the 'gut' is a rather unreliable advisor >.<



My approach is the following ..

1) If the task requires specific tools or skills then the character needs to have those pre-requisites in order to attempt the task (e.g. opening a lock may require proficiency with thieves tools AND having a set of tools ... otherwise you can't even attempt to open the lock ... if you have proficiency but only improvised tools then I might apply disadvantage ... similar if you have tools but no proficiency ... depending on the lock or task)

2) If the character is capable of attempting the task and if there is no limiting factor (time, guards, ticking clock, consequence of failure), I check the characters passive skill first (equivalent to take 10) ... if the passive skill is high enough then I usually narrate the success of the action they take without requiring any die rolls.

3) If there are limiting factors, then depending on the nature of the factors, I will either check the passive skill or have the player roll a skill check.

4) If there is a task for which the character has the necessary skills and for which there are no time limits or consequences of failure bu the task is more difficult than their passive skill then I will just narrate the successful result as taking a very long time making sure that the character actually wants to spend what could turn into hours completing the task.

Thank you, too, for your input. I especially like how you worked passives into your algorithm. I'm just weary of step 4 -- I don't think I can think of one situation where I'd actually prefer time to guarantee beating a 20+mod DC. Do you have any examples? All I can come up with is: Knowledge checks: No. More time doesn't help. Go find a library or an expert. In the case of lockpicking: No. Need a specially-made lockpick or extra research, go find an expert or craft that tool. In the case of social checks: Well mostly failure degrades the relationship so take-20 is only rarely, if ever, applicable in the first place. Survival? Don't know. Maybe?

Thanks again for your time and thoughts!

[terodil]

Hey strangebloke,

thank you for your reply as well!

Just to clarify, regarding climbing trees: Yeah. I'm not about to ask for rolls for trivial or rather easy tasks. I'm just curious about the take-20 rule, which really only affects high-DC challenges.

Now I do have one question. You say:

You must communicate the results of failure or success as well as the task's approximate difficulty to your players before they commit to the action.
Could you give me an example for this? How do you communicate this? Do you really telegraph to the players that a lock would jam if they failed the DC by, say, 10 or more?

[Trustypeaches]

In situations where success is guaranteed, eventually, you can still use a dice roll to determine how effectively the character performs the check. Basically, tiered DCs that indicate different levels of success.

This can have mechanical implications or be purely flavorful, depending on the situation.

[strangebloke]

Hey strangebloke,

thank you for your reply as well!

Just to clarify, regarding climbing trees: Yeah. I'm not about to ask for rolls for trivial or rather easy tasks. I'm just curious about the take-20 rule, which really only affects high-DC challenges.

Now I do have one question. You say:

Could you give me an example for this? How do you communicate this? Do you really telegraph to the players that a lock would jam if they failed the DC by, say, 10 or more?

Well, as to the climbing trees specifically, I bring that up because DMs often ask for athletics to climb, when the rules explicitly give every character a climb speed. Even Centaurs don't need checks to climb ropes. Its only when there's some risk like loose handholds that you need to make a check.

As to conveyance, well, here's a good example:

Player: "I climb the wall."
DM: "It's raining, and the handholds are questionable. A normal person would probably hurt themselves if they tried to get over it. You can attempt it, but it will require an athletics check."
P1: "How hard do I think it would be for me?"
DM: "You're strong and know how to handle yourself. It'd be risky, but you've definitely done harder climbs."

What do you think the DC here would be?

The trick is that when a PC declares an action, you have to be willing to interrupt them. If they say "I cast fireball!" in the middle of a conversation, interrupt them with initiative. IF they say "I steal his wallet" interrupt them with information.

Then too, sometimes due to the fiction you're trying to convey, you'll want to surprise a player with a check partway through an action. For example:

Player: "I climb the wall."
DM: "Alright. You're strong and you know how to move. You get halfway up the wall without issue, climbing hand-over-hand, when suddenly a stone gives way under your grip. Make an athletics check."
P1: "crap! Can I make an acrobatics check instead?"
DM: "Sure, but the check will only make you take no damage from the fall."
P1: "Fine. Athletics it is." *rolling* "15!"
DM: "you manage to recover, and get over the rest of the wall without issue."

[mer.c]

In situations where success is guaranteed, eventually, you can still use a dice roll to determine how effectively the character performs the check. Basically, tiered DCs that indicate different levels of success.
And/or the time taken for completion, where appropriate

[MaxWilson]

Thanks for your reply, MaxWilson. Just to clarify, are you not using the take 20 rule at your table then?

It's never come up in any kind of a formal sense as a "rule". If someone attempts something that I know they can succeed on eventually, and there is no time pressure, I just tell them, "Okay, you succeed," because I am not ignorant of basic probability math.

The take 20 rule from the DMG would only apply if there are no consequences, so it wouldn't apply to lockpicking unless you are so good that you cannot be stymied, e.g. lockpicking +5 vs. a DC 15 lock. I would only ask for rolls in that case if there is some kind of time pressure, e.g. "can you open the lock before the goblin guards respond to the cries of alarm"? If I don't care how long it takes to succeed, why bother rolling?

[Cynthaer]

I think the issue is with 'impossible'. To belabour the rather unfortunate lockpicking example again and trying to salvage it just for the discussion of this word 'impossible', I'd be happy to allow an exceptionally bad rogue to open an extremely difficult lock if he hit the nat 20 on his one-and-only roll. It's possible for him to open it, but it takes a LOT of luck. Basically, fate intervened in his favour. However, I dislike the take-20 rule because it suggests that as long as he can beat the DC with a nat 20, you just pluck him in front of the lock and let time sort it out -- the problem is that this kind of resolution is 100% reliable. This is not an extraordinary stroke of luck like the nat 20 he rolled, it's an automatism that kicks in as soon as the situation qualifies for take-20. Perhaps I'd be happier with take-20 rather being a take-18 or so...

Let's put some hard numbers down as an example, because I think you're psyching yourself out over a case where take-20 doesn't actually matter.

The "normal" scenario that the take-20 rule is meant to expedite is this: You have a "normal" level 1 Rogue (Dex +3, Proficiency +2, overall +5 to pick locks) and a door with a Very Hard lock (DC 25). As the DM, you've determined there are no time constraints or consequences for failed attempts. This is basically within this Rogue's normal capabilities, so you might as well skip the rerolls and just say, "it takes a minute, and it's very frustrating, but eventually you get it".

Note: As the DM, you decide how the lock works and what the consequences are! You might determine that this lock breaks if the first roll fails, so it's a one-roll pass or fail test. Or maybe you give three tries at increasing DCs, and it breaks if you fail all three. There's no universal rule for how locks work, because every lock is different, and every lock serves a different narrative purpose.

Now, here's the scenario you've described: You have a "bad" level 1 Rogue (Dex +1, Proficiency +2, overall +3 to pick locks), and a door with a mundane but Very Hard Lock (DC 25). As the DM, you've determined there are no time constraints or consequences for failed attempts. This is not within this Rogue's normal capabilities, so the take-20 rule doesn't apply—they simply can't hit the DC. However, as the DM, you've also decided that you're willing to let them try once for a "critical success".

Note: As the DM, you don't actually need to be perfectly consistent if something else improves the flow of the game. You don't need a general houserule about "critical successes" to justify letting a natural 20 succeed if it feels right in the moment, and you have no obligation to treat the ability to "take 20" the same way as a natural roll.


That said, I'm thankful for the reminder that it's up to me to decide if the situation warrants take-20. I'm just not looking forward to argue this with players that point at that passage in the DMG and go 'but but everybody does it this way! You're a **** DM, I'm leaving and I hope nobody ever plays at your table again'.

[...]

That's all very valid. I'm asking chiefly for clarity and peace of mind, not because I keep running into this. In fact, I hope not to.

[...]

Aye. Thank you. As I said though, if you are a relatively inexperienced DM like I am and face players who know the material inside-out by heart, the 'gut' is a rather unreliable advisor >.<

I think I have a better idea where you're coming from now. In that context, let me assure you: The take-20 clause rarely comes up in actual play.

If you consider the ability check rules as a whole, you can see why. In most scenarios where taking 20 yields certain success, you wouldn't have called for a roll in the first place: "Only call for a roll if there is a meaningful consequence for failure."

Now, you might ask: If the "meaningful consequence for failure" clause makes taking 20 pointless, then why did they include the bit about taking 20 at all?

One reason is that it quantifies the time cost of an activity when time is the only source of pressure — but in my opinion, mostly it's just to help new DMs avoid getting bogged down by obstacles with explicit DCs that only matter under certain circumstances.

Here's a realistic example. Suppose you're fighting through a kobold den, and one room has kobolds shooting darts at you from a secret room through tiny slits in the wall. There is a locked door (DC 15 lock) that would let your party get into the secret room.

While in combat, rolling for this is a very interesting tactical question: Should the rogue use her Action to gamble on a successful lockpick attempt? If she fails, the consequence is a round of combat wasted, and she must make the decision again the next turn.

Out of combat, well, the kobolds are already dead; who cares how many times you fail? But the new DM may feel like she is getting mixed signals — there's no obvious reason to roll, but the lock has a DC, so...? She asks for a roll on principle, the first attempt fails, and now it's clear that this is basically pointless. Does she keep asking for rolls, or does she say "uh, I guess you unlock it anyway"?

Really, she should have skipped the check altogether (no "meaningful consequence for failure"), but it's easy to see why she ended up on this path. You see a DC, you call for a check. It's a reflex. The take-20 rule is the official way to smoothly resolve this once you're in the middle of it. You ask for a check, the check fails, you realize rolling again is a waste of time, and you just say "it takes a while but eventually you get it".

This is actually why the take-20 rule is in the Dungeon Master's Guide, and not the Player's Handbook. It's not a tool for players to optimize their skill checks; it's a tool for dungeon masters to keep gameplay moving!

[terodil]

Thank you all, once again, for your comments.

@Strangebloke: Cool, I understand what you meant now -- funnily enough, it's what I'd usually have told the players anyway. When you talked explicitly about difficulty and consequences for failure, I was wondering if you'd actually mention the DC, and perhaps even enumerate the tiers of success you might have assigned to the challenge. I remember reading posts in a related thread where people suggested to actually, literally, do that ("You'll need an 18 to beat the check."), both for clarity of communication and for suspense. I'm not sure which way I prefer -- I think I'd lean more towards your way of doing it, because it maintains some semblance of a fourth wall, though I can see the merits of the more literal approach as well.

@Cynthaer: Wow. Thank you so much for the time and thought you put into your post! I really appreciate it. I tend to overthink and overprepare by a ridiculous degree still, trying to think of ways to approach situations, setting DCs, etc. even though I *know* that most DM primers tell you to exactly not do that: the curse of being a perfectionist. I really need to take it easier and 'go with the flow'. Your post was a great reminder and a tremendous inspiration. :smallsmile: