Opinion: will you be reluctant or eager to switch from 5E to 6E?

[OldTrees1]

Ability scores have this problem. Some tasks are virtually impossible for a person not primed with some basic knowledge or aptitude, but more manageable for another person who has specialized education, training, or similar experience with it. We're not supposed to use proficiency as a gate, and we're not supposed to have different DCs for different characters attempting the same action. So we're left with setting a really high DC. It's clunky.

I have considered using 1d10+5 instead of 1d20 for ability checks.

I found the limited tools of "automatic fail & automatic success" struggled to represent growth. I found it tempting to round "need a 5+" to "automatic pass" and "needing a 16+" to "automatic fail". Reducing the size of the RNG by changing 1d20 to 1d10+5 seems reasonable if 5E encourages me to delete the extremes on the d20 anyways.

[Eldariel]

I have considered using 1d10+5 instead of 1d20 for ability checks.

I found the limited tools of "automatic fail & automatic success" struggled to represent growth. I found it tempting to round "need a 5+" to "automatic pass" and "needing a 16+" to "automatic fail". Reducing the size of the RNG by changing 1d20 to 1d10+5 seems reasonable if 5E encourages me to delete the extremes on the d20 anyways.

3d6 works well to that end: extremes exist but are exceedingly rare. Vast majority of the rolls in the 6-15 region (by comparison, whole 16-18 is less likely than natural 20). The other option is using some "degrees of success"-system, which I tend towards since it maintains the d20 core resolution mechanic (though 3d6 core resolution works pretty well for the whole system as well): use "ability modifier/2 rounded up" as base success and then have the roll modify that (so a +0 character has level 0 success, +1 has level 1 success, +3 has level 2 success, etc. and then roll = DC gives you your base level success while +5 gives you +1 tier result, +10 gives you +2 tier result, etc.).

You could remove the base success level but of course that means your ability bonus barely matters. Either way, this de-emphasizes the die roll but everyone can still reach a reasonable degree of success. Of course, producing tables of the degrees for an event on the spot can be rough.

[Telok]

I think the straightforward truth is that in the end, DMing can't be taught. I don't mean you can't learn to be a DM. What I mean is, you learn to be a DM by DMing. You need to be half game designer. More than half, maybe.

No. You absolutely can be taught to DM. I have taught people how to run Paranoia in an hour. The difference is Paranoia's blackjack d20 mechanic and comedy style mean the DM never sets dcs and there are no inappropriate results. Contrast with D&D where the DM has three+ decisions before even stating what the player should roll for all actions.

And that's something I (hope|think) we can all agree on--the DMG doesn't do a great job of helping us understand what the developers were thinking and doesn't give enough tools to understand game design. This is orthogonal to "what should the DC be in this case". More "how we expect you to think about target numbers and checks."

For example, I could get behind a set of worked examples, where the various considerations are called out. This doesn't have to be exhaustive or even representative for all the possibilities, but can illuminate what they think are the important parts.

Dear gods yes. So bloody much this. People (in general typically) learn & respond better to anecedotes & examples than they do to data & no explanation rules.

[KorvinStarmast]

Segev, does this mean if a player tried to read some arcane runes you'd call for a DC 30+ Arcana check because you personally can't read magical runes?

Honestly it would make some sense: either you know the language a bit or you do not and if you do not know the ancient language at all due to it being ancient and you never having learned it, no sort of dice will help you.
Making skills replace language is definitively not raw either, languages are their own thing.

The check represents the abstraction of "did the PC learn a bit about this at some point in their lives?" which is a heck of a lot of detail that the player and DM do not know in the first place. And, the check is a short hand for "can you remember learning this?" or "did you ever learn enough about this to make an educated guess?"
We discover who the PC is through play, we don't actually have to immerse ourselves into the entirety of their existence before adventure 1 at level 1 during session 1.

I'd love to know what kind of table or text would be expected to help with adjudicating "over-seasoned tacos."

At my table, serving over seasoned tacos might get you disadvantage on the following social skills check. My wife, however, may wonder if it is possible to over season tacos.

Specifically, the "does this take a check at all" steps that come before setting a DC. A certain task may be possible but not guaranteed for one person and impossible for another. Or guaranteed for one person and impossible for another. That happens based on the fiction and the characters, without reference to things like proficiency. because proficiency is too broad to cover individual cases.

Spoiler: good example

Show

Take, for instance, the task "respond appropriately with the hand sign when challenged by a member of the Thieves Guild of Thieftown". If you're not a member of that guild, you don't know the right sign. If you are a member, you do know it. You might have two rogues with identical proficiencies, but if one had the backstory/adventuring statement[1] "I am a member of the Thieves Guild of Thieftown" and the other doesn't, one gets an auto-pass and the other an auto-fail.

Spoiler: another good example

Show

The barbarian from the howling wilderness does not know the secret symbology of the Cult of the Obscure. The scholar of religions might. One gets a check, the other doesn't.

And since players don't call for checks, it all works.

Bingo, the bolded part. This is where 3.x norms are harmful if carried over into 5e. (IME)

Which raises a point--we never have to decide check/no-check or DCs in the abstract. They're always specific to one particular character in one particular situation in one particular world.

Except perhaps when trying to make a dubious point during a forum argument.

[PhoenixPhyre]

Except perhaps when trying to make a dubious point during a forum argument.

But no one would do that, right?

[KorvinStarmast]

But no one would do that, right?

Perish the thought.

[OldTrees1]

3d6 works well to that end: extremes exist but are exceedingly rare.

The other option is using some "degrees of success"-system (snip a "damage" step implementation)

Depending on what exactly bothers someone about 5E's ability check math, these can be anywhere from perfect to orthogonal.

For me 1d10+5 works better than 3d6 because I am more worried about the range (compared to improvement) than about the frequency of the extremes. So deleting the extremes might be better than keeping them around at low frequency. On the other hand some players really love their bell curves as ideal solutions.

There are many types of degrees of success systems (some are just "beat DC by N"). Your example sounds like a damage check step. This is great for opposed checks or other cases where you want to minimize improvement on the initial check without minimizing overall improvement. It is another solution I have considered.

[Psyren]

The check represents the abstraction of "did the PC learn a bit about this at some point in their lives?" which is a heck of a lot of detail that the player and DM do not know in the first place. And, the check is a short hand for "can you remember learning this?" or "did you ever learn enough about this to make an educated guess?"

No one is against there being a check to decipher runes (at least I'm not, as I explained to noob). The issue rather is when the DM is basing DCs on their own IRL capabilities, rather than considering what talent and training (and magic and luck) might look like for an adventurer in a D&D world. If we use ourselves as the baseline, then a lot of tasks are going to have DCs that can only be hit by high-level PCs, if they're achievable at all.

Specifically, the "does this take a check at all" steps that come before setting a DC. A certain task may be possible but not guaranteed for one person and impossible for another. Or guaranteed for one person and impossible for another. That happens based on the fiction and the characters, without reference to things like proficiency. because proficiency is too broad to cover individual cases.

Take, for instance, the task "respond appropriately with the hand sign when challenged by a member of the Thieves Guild of Thieftown". If you're not a member of that guild, you don't know the right sign. If you are a member, you do know it. You might have two rogues with identical proficiencies, but if one had the backstory/adventuring statement[1] "I am a member of the Thieves Guild of Thieftown" and the other doesn't, one gets an auto-pass and the other an auto-fail.

While this is indeed a good example of "no roll needed" in a vacuum, I think it's also possible for this challenge to involve a roll before the moment of the challenge. Like if you successfully spied on suspected Thieves' Guild members beforehand, or successfully befriended/intimidated/tricked one into teaching you the particulars, you might have been able to learn the secret hand sign needed for this test.

Some games might even let you perform this kind of roll retroactively (e.g. Blades In The Dark) and if you don't mind a tiny bit of narrative absurdity, D&D could model this as well.

And such things are very common, especially for Intelligence (aka knowlege)-based checks. The barbarian from the howling wilderness does not know the secret symbology of the Cult of the Obscure. The scholar of religions might. One gets a check, the other doesn't. And since players don't call for checks, it all works.

Agreed, but I'll also point out that when placing challenges in the world, another good piece of advice is the Rule of Three:

"If an obstacle stands between your players and your plot's critical path, try to include at least three ways to overcome it."

For example, knowing the symbol of the Cult of the Obscure is one way to tell where their hideout is. It's also true the barbarian from the howling wilderness is unlikely to know it.
But what he MIGHT be able to do, is track one of them across said wilderness, or even track their horses.
Alternatively, he might be able to ambush one outside of said hideout, beat the gently caress out of him, and make him spill the beans.

This also helps you devise meaningful consequences for failure. If your party doesn't know their symbol and can't track one down, they might be forced to do option 3, which involves a dangerous fight, or could even alert the cult that you're coming for them when the cult's ambushed members fail to return on time.

[PhoenixPhyre]

Depending on what exactly bothers someone about 5E's ability check math, these can be anywhere from perfect to orthogonal.

For me 1d10+5 works better than 3d6 because I am more worried about the range (compared to improvement) than about the frequency of the extremes. So deleting the extremes might be better than keeping them around at low frequency. On the other hand some players really love their bell curves as ideal solutions.

There are many types of degrees of success systems (some are just "beat DC by N"). Your example sounds like a damage check step. This is great for opposed checks or other cases where you want to minimize improvement on the initial check without minimizing overall improvement. It is another solution I have considered.

Personally, I find the extremes only an issue on one side. It's not symmetric. I'm totally fine with people being able to try something they only have a 5% chance at. Because heroic people are heroic and beat the odds. That's in their job description. And if they shouldn't be rolling at all because it's not fictionally appropriate, the opportunity to stop that is much earlier (before setting a DC at all). On the other side, having to roll for things with a very low chance of failure often (but not always) is just an invitation for slapstick/failure theater.

While this is indeed a good example of "no roll needed" in a vacuum, I think it's also possible for this challenge to involve a roll before the moment of the challenge. Like if you successfully spied on suspected Thieves' Guild members beforehand, or successfully befriended/intimidated/tricked one into teaching you the particulars, you might have been able to learn the secret hand sign needed for this test.

Some games might even let you perform this kind of roll retroactively (e.g. Blades In The Dark) and if you don't mind a tiny bit of narrative absurdity, D&D could model this as well.

I'd be opposed to doing it retroactively. If you had already spied on them and knew the hand sign, then no one needs to make the check to know the hand sign at all. And that other check is going to be a completely different one with a completely different situation and considerations.

Agreed, but I'll also point out that when placing challenges in the world, another good piece of advice is the Rule of Three:

"If an obstacle stands between your players and your plot's critical path, try to include at least three ways to overcome it."

For example, knowing the symbol of the Cult of the Obscure is one way to tell where their hideout is. It's also true the barbarian from the howling wilderness is unlikely to know it.
But what he MIGHT be able to do, is track one of them across said wilderness, or even track their horses.
Alternatively, he might be able to ambush one outside of said hideout, beat the gently caress out of him, and make him spill the beans.

This also helps you devise meaningful consequences for failure. If your party doesn't know their symbol and can't track one down, they might be forced to do option 3, which involves a dangerous fight, or could even alert the cult that you're coming for them when the cult's ambushed members fail to return on time.

I try not to plan solutions at all. I plan situations and adjudicate on the fly. I let the players tell me how they approach things, and then draw on my knowledge of the situation.

But I also don't make plot-critical, single-point-of-failure situations as a general rule. Mostly because I don't have a plot to fail. There are just a bunch of things happening that they might be interested in. So my examples were in the moment, when they're in somewhere where knowing the symbolism used by the Cult of the Obscure will help, do they need to roll for that. Or should I just flat out tell them. And the results may be small or large. Depending on the situation.

But yes, generally, single points of failure are bad and should be avoided unless the PCs have really backed themselves into the corner and only have that one slim reed left. In which case, failure is on them.

[OldTrees1]

Personally, I find the extremes only an issue on one side. It's not symmetric. I'm totally fine with people being able to try something they only have a 5% chance at. Because heroic people are heroic and beat the odds. That's in their job description. And if they shouldn't be rolling at all because it's not fictionally appropriate, the opportunity to stop that is much earlier (before setting a DC at all). On the other side, having to roll for things with a very low chance of failure often (but not always) is just an invitation for slapstick/failure theater.

Normally I would agree, but 5E pushed the math far enough that I personally found the majority of the RNG to be an issue. I could nitpick and say it was the bottom 2/3rds rather than the extremes, but it is mathematically similar when not using a bell curve. As you know, in 5E if a character starts with a 5% (+2 Prof, 14 Stat) and grows to (+6 Prof, 20 Stat) then they would only have a 40% chance despite 19 levels of growth. Personal preference, I feel "beating the odds at 1st level in your specialty" should eventually grow into not having to roll anymore by 20th. However it is hard for me to draw a line saying 35% rolls but 40% is now an auto pass. Something like 1d20 (minimum 10) but more granular might be worth me looking into. Or maybe just the right side of a bell curve?

output 1d{11, 11, 11, 12, 12, 12, 13, 13, 13, 14, 14, 15, 15, 16, 16, 17, 17, 18, 19, 20}
5%->40% story upgrades to 5%->70%
output 1d{8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14, 14, 15, 16, 17, 18, 19, 20}
5%->40% story upgrades to 5%->50%

[PhoenixPhyre]

Normally I would agree, but 5E pushed the math far enough that I personally found the majority of the RNG to be an issue. I could nitpick and say it was the bottom 2/3rds rather than the extremes, but it is mathematically similar when not using a bell curve. As you know, in 5E if a character starts with a 5% (+2 Prof, 14 Stat) and grows to (+6 Prof, 20 Stat) then they would only have a 40% chance despite 19 levels of growth. Personal preference, I feel "beating the odds at 1st level in your specialty" should eventually grow into not having to roll anymore by 20th. However it is hard for me to draw a line saying 35% rolls but 40% is now an auto pass. Something like 1d20 (minimum 10) but more granular might be worth me looking into. Or maybe just the right side of a bell curve?

output 1d{11, 11, 11, 12, 12, 12, 13, 13, 13, 14, 14, 15, 15, 16, 16, 17, 17, 18, 19, 20}
5%->40% story upgrades to 5%->70%
output 1d{8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14, 14, 15, 16, 17, 18, 19, 20}
5%->40% story upgrades to 5%->50%

I guess...I just don't care about bell curves at all. Because that's not what I want from a resolution system. And assumes a false precision in deciding what the success probability should be.

Basically, I see ability checks as answering the question "which of these two, roughly-equally-probable, equally-acceptable-and-interesting outcomes actually happen" when I'm not sure. If I know what the result should be, there's no reason to roll at all. Foregone conclusions either way don't need rolls. So once we take off the auto-success and auto-fail cases[1], the remaining distribution is roughly flat. Because we've chopped the curve down to the middle. And it's not a sharply-peaked one (otherwise there's no uncertainty anyway).

[1] which must be determined on a case-by-case basis without trying to set an arbitrary cutoff based on DC/success chances. Because there are things that are in the abstract very difficult that the fiction demands must be trivial for a specific character. And things that are trivial in the abstract...but impossible for one specific character. Examples

a) The task is passing oneself off as the True Heir to the Kingdom when presented with a magical trial (sword in the stone style). It's impossible...unless the character happens to actually be that true heir. In which case it's trivial. Or figuring out what the hand sign is for a secret cult. Hard from Intelligence (Religion)...unless you've seen it before. In which case it's trivial.
b) Answering the question "did you kill the king" honestly is really trivial for everyone who didn't kill the king. But impossible for someone who did do that. Etc.

[EggKookoo]

Basically, I see ability checks as answering the question "which of these two, roughly-equally-probable, equally-acceptable-and-interesting outcomes actually happen" when I'm not sure.

Fundamentally, dice aren't about modeling effort or difficulty or anything like that. Dice are about taking the decision whether something succeeds or fails out of the players' hands (including the DM).

D&D would play just fine if everyone flipped coins. People just like that illusion of simulation.

[PhoenixPhyre]

Fundamentally, dice aren't about modeling effort or difficulty or anything like that. Dice are about taking the decision whether something succeeds or fails out of the players' hands (including the DM).

D&D would play just fine if everyone flipped coins. People just like that illusion of simulation.

I agree. Dice exist to resolve uncertainty. And do so in a way that feels like it's fair and just.

[Mjolnirbear]

I guess...I just don't care about bell curves at all. Because that's not what I want from a resolution system. And assumes a false precision in deciding what the success probability should be.

Basically, I see ability checks as answering the question "which of these two, roughly-equally-probable, equally-acceptable-and-interesting outcomes actually happen" when I'm not sure. If I know what the result should be, there's no reason to roll at all. Foregone conclusions either way don't need rolls. So once we take off the auto-success and auto-fail cases[1], the remaining distribution is roughly flat. Because we've chopped the curve down to the middle. And it's not a sharply-peaked one (otherwise there's no uncertainty anyway).

[1] which must be determined on a case-by-case basis without trying to set an arbitrary cutoff based on DC/success chances. Because there are things that are in the abstract very difficult that the fiction demands must be trivial for a specific character. And things that are trivial in the abstract...but impossible for one specific character. Examples

a) The task is passing oneself off as the True Heir to the Kingdom when presented with a magical trial (sword in the stone style). It's impossible...unless the character happens to actually be that true heir. In which case it's trivial. Or figuring out what the hand sign is for a secret cult. Hard from Intelligence (Religion)...unless you've seen it before. In which case it's trivial.
b) Answering the question "did you kill the king" honestly is really trivial for everyone who didn't kill the king. But impossible for someone who did do that. Etc.

Too many DMs forget the first step to resolving any character action is to determine if it's impossible to succeed or to fail. You only roll if success in the outcome is both possible and uncertain. And personally, even then, I only ask for a roll if the reason they might fail, the obstacle preventing success, is significant or epic.



The rogue climbs the roof because his favourite dagger is up there. I tell him he succeeds. No roll.

The warlock tries to haggle with a merchant. Success is possible but not guaranteed. If he's buying something worth pocket change, he succeeds. If he's buying a magic item though, I'll call for a roll. Because while he *could* fail with pocket change, failure comes with negligible consequences. But the magic item is expensive. A fail here could put the item out of reach. The stakes are significant. Failure *matters*.

When the party says they proceed stealthily, I describe their approach as stealthy, and won't call for a roll until the roll matters, such as when they notice approaching footsteps. I might instead use passive stealth just to draw out the tension.

Narratively, I also prefer describing failures as circumstances outside the character's control. The description assumes competence, not clown car tomfoolery. You didn't fail because you suck at whispering or comically fell over in a clatter; you failed because the guard had a mastiff that caught your scent, or your stomach rumbled really loudly at precisely the wrong time, or light glinted off a weapon.

I think a lot more people would engage with immersive actions instead of skills-as-buttons-to-press, and try thinking outside the box in a more freeform way, and find both the social and exploration pillars more fun and engaging, if DMs had more training.

[Telok]

I guess...I just don't care about bell curves at all. Because that's not what I want from a resolution system. And assumes a false precision in deciding what the success probability should be.

Basically, I see ability checks as answering the question "which of these two, roughly-equally-probable, equally-acceptable-and-interesting outcomes actually happen" when I'm not sure. If I know what the result should be, there's no reason to roll at all. Foregone conclusions either way don't need rolls. So once we take off the auto-success and auto-fail cases[1], the remaining distribution is roughly flat. Because we've chopped the curve down to the middle. And it's not a sharply-peaked one (otherwise there's no uncertainty anyway).

[1] which must be determined on a case-by-case basis without trying to set an arbitrary cutoff based on DC/success chances. Because there are things that are in the abstract very difficult that the fiction demands must be trivial for a specific character. And things that are trivial in the abstract...but impossible for one specific character.

Ok, you need to understand that this comes across as "don't set DCs, just gut feeling based on narrative & what remember of the character sheets to choose between success, failure, and a coin flip". You're saying you'd be better off without an attribute/skill system because you can wing it all, and that's basically what you do.

That's fine and all for you. Ditching subsystems you don't want and winging it goes all the way back and works... to varying degrees for different people. But you do understand there are DMs out there that want more sttucture because they don't have your

[PhoenixPhyre]

Ok, you need to understand that this comes across as "don't set DCs, just gut feeling based on narrative & what remember of the character sheets to choose between success, failure, and a coin flip". You're saying you'd be better off without an attribute/skill system because you can wing it all, and that's basically what you do.

That's fine and all for you. Ditching subsystems you don't want and winging it goes all the way back and works... to varying degrees for different people. But you do understand there are DMs out there that want more sttucture because they don't have your

I'm just going to say "no, that's not what I mean at all" and leave it at that.

[Goobahfish]

I have considered using 1d10+5 instead of 1d20 for ability checks.

I found the limited tools of "automatic fail & automatic success" struggled to represent growth. I found it tempting to round "need a 5+" to "automatic pass" and "needing a 16+" to "automatic fail". Reducing the size of the RNG by changing 1d20 to 1d10+5 seems reasonable if 5E encourages me to delete the extremes on the d20 anyways.

I was playing with a similar idea. I really like the "if you only need a 5" you don't need to roll. Sort of like a weird version of taking 10/taking 20. The non-linearity is kind of neat because you can imagine a +5 basically makes you immune to failure of DC=10 checks and below. The fighter climbs the rope through arrow fire undaunted while the wizard has to really think about it. It doesn't break suspension of disbelief as much.

I think it stems from the 'not all tasks have the same distribution' which is kind of hard to model in a tabletop game.

[Psyren]

I'd be opposed to doing it retroactively. If you had already spied on them and knew the hand sign, then no one needs to make the check to know the hand sign at all. And that other check is going to be a completely different one with a completely different situation and considerations.

For the record I'm against doing it retroactively too. But, let's say I exposed the party to the hand sign by having their rogue possibly notice a thief deal going down in the tavern's back room, but I forgot to call for a roll then - personally I would handwave it and just say they saw/remember it, but another DM might want to test the scout's perceptiveness or recall. I don't think this approach is automatically wrong, even though I'd be more likely to go with no roll if the original fault was mine.

I try not to plan solutions at all. I plan situations and adjudicate on the fly. I let the players tell me how they approach things, and then draw on my knowledge of the situation.

But I also don't make plot-critical, single-point-of-failure situations as a general rule. Mostly because I don't have a plot to fail. There are just a bunch of things happening that they might be interested in. So my examples were in the moment, when they're in somewhere where knowing the symbolism used by the Cult of the Obscure will help, do they need to roll for that. Or should I just flat out tell them. And the results may be small or large. Depending on the situation.

But yes, generally, single points of failure are bad and should be avoided unless the PCs have really backed themselves into the corner and only have that one slim reed left. In which case, failure is on them.

[/quote]

This sounds like a sandbox - which is a valid playstyle, but personally I prefer a more crafted story (or at least, a crafted critical path.) This is not to say that the players can't come up with a solution I didn't consider along that path, but having a few prepared ones in mind helps me direct them if they get stuck.

Fundamentally, dice aren't about modeling effort or difficulty or anything like that. Dice are about taking the decision whether something succeeds or fails out of the players' hands (including the DM).

D&D would play just fine if everyone flipped coins. People just like that illusion of simulation.

It would, so long as you remembered to tailor what you flipped coins for to the PCs' changing (growing) competencies.

I think it stems from the 'not all tasks have the same distribution' which is kind of hard to model in a tabletop game.

It's easier than you think. Even if the chance of failure on the dice is not that different between actors (when it should be) - what ultimately happens on a failure for each actor is completely up to you.

[Telok]

I'm just going to say "no, that's not what I mean at all" and leave it at that.

Then there's a problem because that's what I'm hearing you say. If you just "leave it at that" then I can't understand your position and it all becomes garbage in - garbage out.

[Psyren]

Then there's a problem because that's what I'm hearing you say. If you just "leave it at that" then I can't understand your position and it all becomes garbage in - garbage out.

He's saying that the system works perfectly fine with a flat distribution (1d20) rather than a bell curve (3d6), because the latter is not an improvement; you shouldn't be calling for a roll in the first place if any of the outcomes of that roll is something you absolutely don't want to happen under any circumstance.

[EggKookoo]

It would, so long as you remembered to tailor what you flipped coins for to the PCs' changing (growing) competencies.

Right. To be clear, I'm not saying it would play the same. Just that it would play fine (at least how I would define "fine").

[KorvinStarmast]

The issue rather is when the DM is basing DCs on their own IRL capabilities, rather than considering what talent and training (and magic and luck) might look like for an adventurer in a D&D world.

Which puts you and I in violent agreement.

He's saying that the system works perfectly fine with a flat distribution (1d20) rather than a bell curve (3d6), because the latter is not an improvement; you shouldn't be calling for a roll in the first place if any of the outcomes of that roll is something you absolutely don't want to happen under any circumstance.

That's what I got out of it also.

[Segev]

Which is not one Segev is arguing because you are taking into account what the character could know/do whereas Segev is saying that if it's something he personally finds hard to do in real life then it's going to be hard for a character to do as well. So if Segev isn't a doctor and can't diagnose/treat an illness then the DC for that check is going to be 25+ because being a doctor is very hard and it doesn't matter that my character's background is in fact a doctor, if it's hard for Segev in real life it's hard for every character in the game.

No, that is not what I have said. I have said that my experience with rope climbing, both personal and observed, is that it is very difficult, and only some people can do it reliably. I would therefore never expect an untrained, untalented, 8-strength wizard to be able to climb a rope. DC 20 is the minimum that fits that.

Genre convention - or rather, expectations of the game rules - are that climbing something that isn't (particularly) hard to climb should just be automatic. Ropes are clearly things people use to climb, so they can't be "particularly hard to climb" in the terms intended by the game when the game goes out of its way to say that climbing normally doesn't require a check.

I am not a medical doctor; you're correct. I would not, however, expect that diagnosing a disease is ultra-hard; in my experience, people do it all the time without medical degrees, with sufficient success to treat it at the level that D&D typically needs it to be.

You do not accurately represent my position, in other words. Instead, you seem to wish to attack me, personally, and my alleged motives and/or intellectual bias, rather than addressing the case I have raised. Nobody has yet addressed it, other than to scoff and say, "You know better, and I refuse to play this game." That is not actually proving me wrong.

I have yet to see anybody provide for how the claims that:

• I should know that DC 20 is too high for climbing a rope
• I should use the criteria outlined in this thread that allegedly are implied by the non-closed loop in the book on calling "DC 10 is easy" etc. sufficient to make DCs, which I have outlined how that tells me DC 20 is probably right
• The fact that I know better because there is actually a guildeline in the book regarding how hard climbing things should be means there is no need for such guidelines because I should have known without it

...are a logically consistent, non-self-contradictory set of premises.

Please show me how rope climbing is objectively definitely something that should require no check, based on the criteria given for setting skill check DCs. The best you'll be able to do, I wager, is quote the flow chart Psyren provided, and assert that most of the time it's just plain uninteresting if the PCs fail to climb that rope. If I am right, then you've failed, because there are always other ways forward; problem-solving is part of the game. And if that's your criteria, then it also should be automatic to read those magic arcane runes even if you just have a party of big dumb fighters since it's uninteresting if you can't read those runes. (Hint: you shouldn't have those runes needing to be read to advance things, or you should have a means for the party to get them read by somebody who can provide the info they need. Problem-solving, alternate ways forward, etc.)

In the end, what nobody has been able to do is show that the case where we have a guideline that changed my determination for the DC of an ability check - and everyone seems to agree that my DC, which I'd be arguing for if we only had the criteria spelled out by people in this very thread for determining DCs as being the most sensible one, is egregiously high - doesn't demonstrate that guidelines for what the game's baseline expectations are about basic PC capabilities are good and even necessary. Instead, people have resorted to ad hominem attacks on my character and my intellectual honesty.

Never mind that I stand by my claim that, absent the genre convention that ropes just are climbable because that's a genre convention, your untrained 8 Str character should never be able to climb a rope, and your untrained Str 10 character should have a very hard time doing so.

[Psyren]

I have said that my experience with rope climbing, both personal and observed, is that it is very difficult, and only some people can do it reliably. I would therefore never expect an untrained, untalented, 8-strength wizard to be able to climb a rope. DC 20 is the minimum that fits that.

You have three fundamental errors here:

1) "My personal experience" - you are not a D&D adventurer, and no D&D stat is meant to map perfectly to the real world. Going by the PHB, someone with 8 Str has a carrying capacity of 120 lbs, and can push, drag, or lift twice that amount, even at Small size. Very few children can push something weighing 240lbs, but a much larger number can climb robes just fine, so concluding that 8 Str makes you incapable of climbing when a clearly much lower Str score can do so means your logic is flawed.

2) "Reliably" - Doing something "reliably" can be as simple as taking 20 (which takes 10x the time, not 20x as in 3.5.) Guaranteed success is pretty reliable. "Failing" a climb check can mean anything from you plummet from the rope to you slip slightly to you just stall your progress for a brief interval.

3) ...you're still using the climbing example when we've explained over and over that climbing anything in 5e doesn't need a check by default, the only time you should be calling for a check is if there's some kind of challenging circumstance with a meaningful consequence for failure.

[Segev]

You have three fundamental errors here:

1) "My personal experience" - you are not a D&D adventurer, and no D&D stat is meant to map perfectly to the real world. Going by the PHB, someone with 8 Str has a carrying capacity of 120 lbs, and can push, drag, or lift twice that amount, even at Small size. Very few children can push something weighing 240lbs, but a much larger number can climb robes just fine, so concluding that 8 Str makes you incapable of climbing when a clearly much lower Str score can do so means your logic is flawed.

Children also aren't trying to lift that much when climbing a rope. You are trying to make some VERY extensive arguments for why DC 20 is definitely too high, but you certainly aren't making a very strong argument for "DC don't bother just succeed."

Amazingly, though, with a guideline for how climbing is supposed to work, I happen to agree with you that DC 20 is egregious, since it should be "DC don't bother just succeed." I can only guess how many other things you would tell me I'm way off base for, and then act like it's my fault because I am somehow being unreasonable because I don't come up with what you consider reasonable DCs.

Because, again, I note that you're acting like DC 20 for rope climbing is, in fact, a problematic thing for me to rule if I am DMing. Where, before, people were telling me, "No, no, it's easy, just think about how hard you think it should be, and give a number and go! Don't worry about it!"

Since DC 20 is apparently so egregiously wrong that it's problematic (I am unsure if this is because, as we all agree, the guideline actually exists and tells us that that's way off, or if it's because you independently came up with a different DC based on your own analysis of how hard rope climbing is/should be and find that my ruling is so much different from yours that it's a problem that needs correcting), you're all telling me off for why I'm being unreasonable. I would have expected, based on the way it was presented to me, that you'd say, "Yeah, sure, DC 20 for your games. It doesn't matter, whatever you think is right." Of course, we have a guideline in the rules that tells us that I am, in fact, egregiously wrong.

Why this means that I must be being disingenuous for coming up with DC 20 if I use the guidance for skill DCs rather than the guidance for climbing is... actually rather telling: it disproves the assertions of the "we don't need guidelines" camp, and therefore I must be being unreasonable, rather than illustrating a legitimate problem.

Since, again, all three of the following cannot be true:

(1) Climbing is supposed to be something you can just do, according to the guidelines for climbing in the rules.
(2) The rules for assigning ability check DCs are as complete as they need to be and no guidelines are required, because you can assign whatever DCs seem reasonable/realistic/appropriate on the fly.
(3) Segev followed the guidance referenced in (2) and came up with DC 20 for climbing ropes.

And, since all three cannot be true, Segev clearly must be a liar no matter how much he demonstrates why DC 20 makes sense to him and would make sense to him if he were called upon to rule it in his game absent any guidance other than that presented in (2). He surely would not have come up with DC 20 based on that guidance, which is all we need to come up with perfectly reasonable DCs, and thus he would know without the guideline in (1) that it should be no check at all. It cannot possibly be that guidelines, such as the one in (1), would actually change the DCs that DMs assign to things; all DMs would definitely know what Segev is clearly only pretending not to know.

And, if I am wrong about the above paragraph being your belief about my position, please explain what you do believe about me and what I am asserting, because I am not seeing demonstration that I am wrong about how to come up with that DC according to the guidance provided, only "you know better" and "clearly, because some children can climb ropes, all adventurers can do so without checks, no matter what archetypal physique they have."

2) "Reliably" - Doing something "reliably" can be as simple as taking 20 (which takes 10x the time, not 20x as in 3.5.) Guaranteed success is pretty reliable. "Failing" a climb check can mean anything from you plummet from the rope to you slip slightly to you just stall your progress for a brief interval.

So... DC 20 is perfectly reasonable for "reliably" climbing ropes as long as you have a str of 10 or higher. ARe you now agreeing that that's a perfectly reasonable DC to assign?

3) ...you're still using the climbing example when we've explained over and over that climbing anything in 5e doesn't need a check by default, the only time you should be calling for a check is if there's some kind of challenging circumstance with a meaningful consequence for failure.

I'm using it because the only reason it doesn't need a check by default is that there is a guideline that says so. And since my entire point is that we need more guidelines for things governed by ability checks, the existence of a guideline that shows why my DC assignment would be WAY off base supports my claim.

Saying, "The guideline exists, therefore you can't use the example to demonstrate how a guideline has made the basis for the DC to perform this task to demonstrate how guidelines are useful," is a lot like saying, "Because a doctor treated the patient, you can't use this patient's recovery to support your claim that doctors are useful when patients are sick."

[Psyren]

Children also aren't trying to lift that much when climbing a rope.

And the 8 Str Wizard who can lift 240lbs isn't either? Stop basing your D&D DCs on what you can do IRL. (Or keep doing it, and accept that you'll never be aligned with people who aren't doing that.)

So... DC 20 is perfectly reasonable for "reliably" climbing ropes as long as you have a str of 10 or higher.

No, assigning a DC for something that shouldn't have one isn't reasonable. You keep skipping that step, and that's exactly the mistake your requested "guidelines" would encourage others to do.

I'm using it because the only reason it doesn't need a check by default is that there is a guideline that says so.

Yes, and it's the same guideline every other check gets. If there is no meaningful consequence for failure / they can simply try again until they succeed, don't ask for a roll. Climbing isn't special in this regard. Why is that so hard to grasp?

[Keltest]

(1) Climbing is supposed to be something you can just do, according to the guidelines for climbing in the rules.
(2) The rules for assigning ability check DCs are as complete as they need to be and no guidelines are required, because you can assign whatever DCs seem reasonable/realistic/appropriate on the fly.
(3) Segev followed the guidance referenced in (2) and came up with DC 20 for climbing ropes.

Well, either you failed to correctly follow the guidance in 2 or yes, they are all true. My personal money is on the latter.

The problem here Segev is that you are dramatically overestimating how hard climbing a rope is. Thats literally it. Its not that hard. Its not trivial for every human being, but people in reasonably good shape (ie not me) can do so with a minimum of difficulty, and thats without considering that there are ways to set up the rope to assist in climbing, like knotting it periodically for better support of the climber.

But I'm still convinced that you know this, which is why youre sticking with the rope example and the DC 20 amount, because if you changed it to, say, picking a lock instead, your argument would immediately become weaker, and you wouldnt have the "but the rules say I'm wrong" bit to fall back on whenever somebody calls you on having a poor basis for your DC of 20.

Because in order for your argument to make any sense, you have to have deliberately picked something egregious (since otherwise you just demonstrate it works correctly) and the only way you can do that is to call for a check in a situation that doesnt demand one in the first place, at which point youre just failing to use the system correctly.

[Telok]

He's saying that the system works perfectly fine with a flat distribution (1d20) rather than a bell curve (3d6), because the latter is not an improvement; you shouldn't be calling for a roll in the first place if any of the outcomes of that roll is something you absolutely don't want to happen under any circumstance.

And his text does not communicate that to me, especially nothing about 3d6s or bell curves. Ergo, problem.

And it works for his personal style of DMing that's based on years of experience, understanding statistics, and reading & dissecting the DMG over and over. Yet every DM & player I've personally met is, at best, apathetic about that very same system because they've had problems with it that he doesn't.

For example, it really doesn't mesh well with my campaign DMing style. I prep a lot of stuff far in advance, often whole dungeons before characters are ever generated. I may put in a dc 20 thingy somewhere because thats "very difficult" and (I think) appropriate for whatever it is. Then I can run the dungeon easily, spending my time & energy on the inhabitants, player actions, etc., without having to memorize or reference copies of the character sheets to know what their bonuses & details are, and without having to decide anything. But that style doesn't play well with 5e's checks, yet it's worked great for years across half a dozen systems that don't require me to stop and think if I need to prevent inapplicable results for every single potential noncombat die roll.

I just realized, my prep ends up looking like a lot of modules. Its personalized, organized, trimmed, and annotated with fallbacks & error traps, but it often kinda looks like a fixed up sandbox style module. And modules are what all the new DMs I've seen bounce off 5e were trying to use. I do recall a "if they'd intended you to not roll they wouldn't have put a dc on it" type response from one of those DMs. Failed it too, +7 vs dc 13 climb. Of course nobody else dared to try because the consequences of falling into a pit with oozes were potentially really "interesting". I wonder, modules (prep in general) are supposed to move work & effort out of game time & into the prep time, so you don't have to keep thinking of all the details.

[OldTrees1]

And his text does not communicate that to me, especially nothing about 3d6s or bell curves. Ergo, problem.

And it works for his personal style of DMing that's based on years of experience, understanding statistics, and reading & dissecting the DMG over and over. Yet every DM & player I've personally met is, at best, apathetic about that very same system because they've had problems with it that he doesn't.

Their post:

I guess...I just don't care about bell curves at all. Because that's not what I want from a resolution system. (snip elaboration)

Background context:
A few of us posters (EggKookoo was also involved) had been discussing how 5E has issues with the extremes of the RNG. We were highlighting our particular issues, preferences, and thought experiments. I had mentioned 1d10+5 at one point. PhoenixPhyre mentioned the lower extreme (rolling a 3 when you needed a 5) bothered them more than the higher extreme (rolling a 20 when a 19 would have failed). I considered their insights and wondered if I personally should explore a right leaning bell curve instead of the 1d10. That would cut out the low extreme while preserving the high extreme.

Summarizing their post:
PhoenixPhyre personally does not find bell curves fit their use case/pain point/playstyle/perspective. They elaborate on their perspective, but the main takeaway is that while I might still want to explore a right leaning bell curve, it would be a poor fit for PhoenixPhyre.

I hope that summary helps reveal their post was talking about bell curves and helps explain what that text did not communicate to you. PhoenixPhyre does not find bell curves ideal for them. Savvy?

[Goobahfish]

It seems to me that the issue stems mostly around the concept of 'can do/can't do' being resolved via RNG in an unsatisfying way. Because there is a 1-20 range (i.e. 19) in success/failure, it seems that for something to be easy for one character it has to be a coin flip for another (at worst) and likewise, something which is impossible for one character is a coin-flip for the pro. I'm assuming a -1 vs +8ish bonus.

The distribution here leads to several unsatisfying results. Realistically, a Monk should be climbing a rope with literally no effort while a wizard should be struggling to even try. The game doesn't reflect this as written. Establishing a DC for climbing a rope is non-trivial because it depends on who is climbing it as to what the intuition should be... again because of the above implementation (i.e. rules) issue. If you have different DCs for different players, you're basically implementing hidden circumstance bonuses from 3e under the covers. If you use your 'fighter' intuition, you get ... DC = 8-ish. If you use your Wizard intuition, you get DC 20-ish. This is a somewhat unavoidable problem given the rules.

It's easier than you think. Even if the chance of failure on the dice is not that different between actors (when it should be) - what ultimately happens on a failure for each actor is completely up to you.

That is a dangerous concept, certainly prone to breaking the suspension of disbelief in a game. Two players roll the same roll and the outcome is different? I can't imagine (barring special abilities) that such an action by a DM wouldn't be construed as favouritism/bias or just bad DMing.