D&D 5E Consequences of Failure

[Lanefan]

Actually, I’d say it’s very much in line with 3e design, and isn’t terribly far off from 4e design (which despite popular conception was very much a direct continuation of 3e design, only further evidenced by the abundance of 4e style design found in PF2). What it’s not in line with is pre-3e design, or 5e design which revived a lot of pre-3e thinking.

Sorry, old chum, but you've got this backwards.

In 1e your roll was always assumed to represent your best attempt under the circumstances, and a re-roll was not allowed unless something materially changed in the fiction. For example, if you try to pick a lock and fail you can't just try again; the assumption is that you already did try again and this is the best you can do. But if something changes, e.g. someone hands you a better set of thieves' tools, you can try again and get another roll.

3e gave us 'take-20'; which changed the underlying assumption from "one roll represents your best attempt, no re-rolls" to "you can (and thus would) keep rerolling until you succeed, so let's just jump to when you roll a 20 and have done with it".

4e, while dropping the 'take-20' terminology, kept with allowing rerolls until you succeed. 5e kinda kept it but is even less clear, as far as I've seen.

 

[Ashrym]

So, meaningful consequence for a failed “knowledge check”...

Is “not knowing” really enough? That seems to encourage the “I try again” scheme or the next player declaring “I try!” In which case, the meaningful consequence for failure materializes: “after a group discussion that lasts X minutes, you cobble together the facts you were looking for.” And now we’re that much closer to @Bawylie’s potential wandering monster threat. Time can be a dangerous cost. Unless it isn’t - in which case, no roll - just Yes or No.

Sorry, kinda thinking out loud here... thoughts?

I found knowledge skills and failed check rather meaningless given the lack of inherent danger. In combat a person is distracted so I have players roll if it's relevant at the time. If it's out of combat I usually stick to passive checks unless they have resources allowing them to research the lore, such as a library or documents; if they have meaningful access to such information I let them "take 20". Multiple characters attempting to piece information is the help action and would give +5 bonus to passive checks or advantage on rolls.

That's how I choose to deal with arcana, history, nature, and religion to recall lore. It's often a know it or don't and research grants knowledge as an action PC's can take.

Side note: I rolled 5 1's in a roll Sunday night. The consequences of failure was mockery at doing 0 DPR before the encounter ended with many "rogues suck" jabs. I was impressed with the odds. I've never rolled 4 1's in a roll let alone 5 in decades of gaming.

 

[5ekyu]

I found knowledge skills and failed check rather meaningless given the lack of inherent danger. In combat a person is distracted so I have players roll if it's relevant at the time. If it's out of combat I usually stick to passive checks unless they have resources allowing them to research the lore, such as a library or documents; if they have meaningful access to such information I let them "take 20". Multiple characters attempting to piece information is the help action and would give +5 bonus to passive checks or advantage on rolls.

That's how I choose to deal with arcana, history, nature, and religion to recall lore. It's often a know it or don't and research grants knowledge as an action PC's can take.

Side note: I rolled 5 1's in a roll Sunday night. The consequences of failure was mockery at doing 0 DPR before the encounter ended with many "rogues suck" jabs. I was impressed with the odds. I've never rolled 4 1's in a roll let alone 5 in decades of gaming.

My first roll20 online gsme was with new people. I rolled 5 nat-1 in a row on my first 5 attacks - in a game where gm ruled that meant hitting allies.

After the 5th I had shot my "allies" (that we had just met - no bonds yet) 5 of 5 which was better than any "enemy" do far. I asked aloud "why arent you fighting back."

 

[MarkB]

My first roll20 online gsme was with new people. I rolled 5 nat-1 in a row on my first 5 attacks - in a game where gm ruled that meant hitting allies.

After the 5th I had shot my "allies" (that we had just met - no bonds yet) 5 of 5 which was better than any "enemy" do far. I asked aloud "why arent you fighting back."

Wow, I thought the time my character with Elven Accuracy managed to roll a triple 1 on his first ever attack was as bad as it could get.

 

[5ekyu]

Wow, I thought the time my character with Elven Accuracy managed to roll a triple 1 on his first ever attack was as bad as it could get.

Honestly, if I had been them by the third shot, I would have been targeting the "sharpshooter".

 

[Guest 6801328]

But that's what doesn't make sense to me. If you eliminate all checks that you consider zero cost there are so many things that you cannot adequately represent in your game.

Most knowledge checks get thrown out the window. Insight? Never used. Using perception to search? It seems like just about anything not related to a physical activity goes away.

It's your game, do it how you want, I think you still dismissing the cost of not gaining useful information.

Maybe. The point of this thread is to explore exactly that question.

I'll admit (and have been admitting) that I wrestle with the knowledge check. I don't think it's a terrible crime to just ask for a check, and then say "Yeah, you know..." or not. But I also don't think that's very interesting or particularly fun. As I expressed in another thread, you're unlikely to ever recall with your friends, "Remember that time I made that History check?"

So if it's possible to make "History checks" more exciting, in a way that can be applied in various situations, I'd like to figure it out.

Insight is somewhat easier. Want to figure out if somebody is lying? As I've stated elsewhere (in many threads) I simply don't have "lie detection" in strangers in my game. Not without magic. But please let's not digress into whether that's a good idea or bad idea, and just start with that baseline. If you want to "search out a lie" in my game I want you to actually search out a lie, and try to push/trick/provoke/lure the subject into making a mistake or otherwise showing their hand. And doing so carries a risk of them figuring out what you're up to. That might involve History, Sleight-of-Hand, Stealth, Animal Handling ("It's clear the horse has been ridden recently, even though he says he's been here the whole time.") or prac

In the thread about Dragon Heist there is a link to a guide about how to DM, which in turn introduces "Response Teams", which are basically groups of thugs who might show up if you start asking the wrong question. Now there's a great consequence to failing skill checks while trying to detect lies. "Hey, Guido, these adventurers dropped by the shop today and were asking all kinds of questions..."

 

[Guest 6801328]

That's fine and I agree. But that's not what the OP wants (@Elfcrusher correct me if I'm wrong). If there's not a cost to failure then he doesn't want the check.

That's the (theoretical) goal. Not that it's bad to have zero-cost failures, but I find them uninteresting, and I'm curious if it's possible to excise them from the game without losing anything.

Rather than try to give knowledge checks a cost, what if we just never roll? What if we always just state it as an action ("I try to recall from the time my monastery as ransacked by Flumphs whether they are immune to thunder damage or not") and let the DM decide. Sometimes he/she says yes, sometimes no. Assuming we trust our DM, we can trust that our background and skill choices and attribute allocation and backstory will all factor into the decision.

Does taking out the dice negatively affect the game?

I think my choice would be to let the DM decide. The way I see it, the DM is going to decide the DC anyway, and in the absence of it being a meaningful decision point to try (by which I mean risk:reward, a.k.a. meaningful consequences of failure) given the choice between leaving it to RNG, or leaving it to a decision by somebody with a vested interested in an exciting adventure, I'll choose the latter.

 

[Oofta]

Maybe. The point of this thread is to explore exactly that question.

I'll admit (and have been admitting) that I wrestle with the knowledge check. I don't think it's a terrible crime to just ask for a check, and then say "Yeah, you know..." or not. But I also don't think that's very interesting or particularly fun. As I expressed in another thread, you're unlikely to ever recall with your friends, "Remember that time I made that History check?"

So if it's possible to make "History checks" more exciting, in a way that can be applied in various situations, I'd like to figure it out.

Insight is somewhat easier. Want to figure out if somebody is lying? As I've stated elsewhere (in many threads) I simply don't have "lie detection" in strangers in my game. Not without magic. But please let's not digress into whether that's a good idea or bad idea, and just start with that baseline. If you want to "search out a lie" in my game I want you to actually search out a lie, and try to push/trick/provoke/lure the subject into making a mistake or otherwise showing their hand. And doing so carries a risk of them figuring out what you're up to. That might involve History, Sleight-of-Hand, Stealth, Animal Handling ("It's clear the horse has been ridden recently, even though he says he's been here the whole time.") or prac

In the thread about Dragon Heist there is a link to a guide about how to DM, which in turn introduces "Response Teams", which are basically groups of thugs who might show up if you start asking the wrong question. Now there's a great consequence to failing skill checks while trying to detect lies. "Hey, Guido, these adventurers dropped by the shop today and were asking all kinds of questions..."

Have you ever watched a medical drama? You know, the doctor has to figure out some disease, or try to sew somebody up who's been practically cut in half or do emergency surgery? Do you notice the part they don't show except in passing? Those doctors filling out paperwork. Yep, filling out forms, filing insurance claims, recording what meds were given is all part of being a doctor. Same with cop shows - they may have the character groan for a moment or two about it, but it's all handled off screen.

It's just not exciting so the shows don't spend any time on it. That's the same way I handle history checks. The check itself is, in a word, boring. Hopefully the information it relays can be interesting (I love it when I can make that happens) or it at least moves the plot forward. Big reveals that I have planned ahead of time are one of the few times I type things up ahead of time and pass notes. Because people enjoy relaying plot twists to their friends.

So I do feel your pain. Even if we have different styles, I just can't think of a way of making knowledge checks interesting. The knowledge gained from those checks? If we do it right, that can be it's own reward.

 

[Guest 6801328]

Weird posting errors. Ignore.

 

[FrogReaver]

Nice. At least there is some acknowledgement in this thread about the shortcomings of goal and approach.

So can we make goal and approach (as meant here) work for knowledge checks.

I think it's possible to make anything have a meaningful consequence of failure so long as you aren't too picky about how that failure looks.

Consider an arcana check to learn how to shut off the magical portal. Perhaps a failed arcana check gives you the information but also has an enemy come through the portal or gives true info that the portal is (now) more difficult to shut off than it would have been if you succeeded.

In other words, if you don't maintain that the consequence of failure must directly follow from whatever your character is trying to do then you can have exactly what you are looking for. Personally I dislike that technique, though I still like having it in my toolbox just in case.