So in the DMG 2024 they repeat the same DCs, but they say that "nearly impossilbe" needs a 19 or 20 to succeed with a high enough proficency score, but this seems like it fails to account how much more likely expertise is to get, with Rangers and Wizards now getting expertise (and Druids and Clerics getting a little expertise). Starting at level 15 most of the expert characters should be able to succeed on a 15 at nearly impossible, skill, which if they get advantage comes closer to rolling a 10 or higher for something labeled "nearly impossible." I'm not totally against it, I jsut wonder if the description of the nearly impossible is innacurate and might make DM's surprised when it actually is easier than stated at high levels.
Well - what's 'impossible' to the man on the street, and a 15th level hero should be two rather different things.
Rollback Post to RevisionRollBack
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
No, it remains nearly impossible == the PCs are just that good.
Dang nabbit.
Rollback Post to RevisionRollBack
Only a DM since 1980 (3000+ Sessions) / PhD, MS, MA / Mixed, Bi, Trans, Woman / No longer welcome in the US, apparently
Wyrlde: Adventures in the Seven Cities .-=] Lore Book | Patreon | Wyrlde YT [=-. An original Setting for 5e, a whole solar system of adventure. Ongoing updates, exclusies, more. Not Talking About It / Dubbed The Oracle in the Cult of Mythology Nerds
As long as you recognize that typical games stop by level 10, DC 30 is nearly impossible for most builds and difficult for almost all builds. In a game that goes to level 20, there's a good chance the PCs have the tools to succeed at DC 40 checks.
I think the point of characters at higher levels is that doing the impossible is "normal" for them in general. I mean, these are characters fighting multiple dragons at a time and winning, dropping meteors down from the heavens with a flick of a wrist and cheating death with a few command words. Its not weird that they can do the impossible with skills as well.
I agree with what everyone else has said; doing the impossible is what higher level player characters are supposed to do. Taking a Rogue with thieves tools as the obvious example how likely is it that with years of experience and practice that all of a sudden they'll struggle to pick a fairly standard lock? They should be opening as quickly and easily as I would using the key
1. Move all of the DCs in the table up by 10. Yes, 10. That is how much is needed if you want them to actually be a challenge at level 15. 2. Scalable DCs. Pathfinder does this and has DCs set by level banding. It works very well to provide challenge at higher level play.
Secret Option 3: Keep the DCs and stop relying on locked chests and hidden doors. This sounds harsh, but by level 10, the types of areas that your player characters are likely to be exploring might legitimately have a chest that is trapped, magically warded, and locked. Though the other thing is that by level 10 there really isn't going to be much in a chest that is going to be exciting for a player character.
I think a more accurate description of difficulty is along the lines of
Low Difficulty: a low difficulty check is one where, if you were to require the entire party to make a check, it would be reasonable to expect half of them to succeed. Low difficulty is 12 + level/4.
Medium Difficulty: a medium difficulty check is one where it's reasonable to expect an average party contains someone who can make the check. Medium difficulty is 15 + level/2.
High Difficulty: a high difficulty check is one that will be a challenge for characters who have dedicated significant effort to being a master of a skill. High difficulty is 20 + level.
(these numbers are not the product of extensive analysis, so YMMV, but remember that there are a lot of skills; just because a party can make DC 30 stealth checks doesn't mean they can make DC 20 religion checks).
I hear what people say about "what's impossible to normal people isn't to the heroes" but I just wish the DMG said THAT, not "Low-level characters have no chance to accomplish a DC 30 task, while a level 20 character with proficiency and a relevant ability score of 20 still needs a 19 or 20 on the die roll to succeed at a task of this difficulty."
A 7th level wizard with expertise in Arcana who has a Int of 18 and has help from the Warlock and guidance from the Druid has .4 probability of success of accomplishing a nearly impossible Arcana task.
A 7th level rogue with expertise in stealth who has a Dex of 18 and the wizard has cast Pass without trace has a .55 probability of success of accomplishing a nearly impossible Stealth task.
That reality is not reflected in the advice to DMs given in the DMG.
I hear what people say about "what's impossible to normal people isn't to the heroes" but I just wish the DMG said THAT, not "Low-level characters have no chance to accomplish a DC 30 task, while a level 20 character with proficiency and a relevant ability score of 20 still needs a 19 or 20 on the die roll to succeed at a task of this difficulty."
A 7th level wizard with expertise in Arcana who has a Int of 18 and has help from the Warlock and guidance from the Druid has .4 probability of success of accomplishing a nearly impossible Arcana task.
A 7th level rogue with expertise in stealth who has a Dex of 18 and the wizard has cast Pass without trace has a .55 probability of success of accomplishing a nearly impossible Stealth task.
That reality is not reflected in the advice to DMs given in the DMG.
Part of the reason this is the way it is, is that D&D 5e 2024 is still built on an ever-increasing escalation of old-school mechanical design that continues to be used in a modern game in a way that it was never intended to be used.
Strictly speaking the ability scores of a typical 1e character were between 8-14. An 18 in Dex was exceedingly unlikely. In modern D&D, you are all but guaranteed to have at least one stat at 18 regardless of the method used. Furthermore, an 18 in 1e meant you would have at best a +3 modifier in modern-day 18 = +4 and its possible to get that up to 20 for a +5. More importantly, the ability score system was designed to be a roll-under system, not a DC-based system.
So basically modern D&D has taken a design and used it in a way for which it was not intended. Fast forward 50 years later with several iterations of modernized power escalation and you end up with a bit of a mess.
Fortunately old school application still fixes the problem.
When you make characters, roll 3d6 down the chain. Bring the power curve down, problem is solved. Your players will not complain how easy it is, but how hard and that is really where it should be. This solves a lot of other problems like CR math and encounters balanced, power-game longevity problems and pretty much everything else people complain about when it comes to 5e. It really is that simple.
Part of the reason this is the way it is, is that D&D 5e 2024 is still built on an ever-increasing escalation of old-school mechanical design that continues to be used in a modern game in a way that it was never intended to be used.
Not really. The core problem is that they came up with this 'bounded accuracy'* concept... and then failed to actually make it work within the game system. The reality is, if you give PCs new toys as they gain experience, and some of those toys work on skill checks... skill checks that are hard at level 1 are no longer hard at higher level. That's just the way the math works. In addition, they threw out the tools that 3e and 4e used to (somewhat) control this escalation -- for example, in 4e, guidance, bardic inspiration, flash of genius, etc, would not stack, because they would all be providing the same bonus type and bonuses of the same type don't stack.
*note that bounded accuracy is based on a fundamental misunderstanding. 3e and 4e didn't scale DC based on the level of the party, they scaled DC based on the level of the challenge.
I hear what people say about "what's impossible to normal people isn't to the heroes" but I just wish the DMG said THAT, not "Low-level characters have no chance to accomplish a DC 30 task, while a level 20 character with proficiency and a relevant ability score of 20 still needs a 19 or 20 on the die roll to succeed at a task of this difficulty."
A 7th level wizard with expertise in Arcana who has a Int of 18 and has help from the Warlock and guidance from the Druid has .4 probability of success of accomplishing a nearly impossible Arcana task.
A 7th level rogue with expertise in stealth who has a Dex of 18 and the wizard has cast Pass without trace has a .55 probability of success of accomplishing a nearly impossible Stealth task.
That reality is not reflected in the advice to DMs given in the DMG.
Isn't this kind of what the game is all about though? The team coming together to solve problems that any one of them couldn't solve alone? I mean what point is there to Pass Without a Trace if it doesn't help to make the impossible possible? That's like the definition of magic. You want 2-3 characters to expend resources and still fail as much as they would without those resources?
I do think it's kind of lame that they pretend the game doesn't scale at all and one set of DCs will hold throughout all levels, but I don't think those DCs should be designed for things to still be unlikely to succeed after allies pool their features and resources. The game is designed to reward exactly that kind of thing.
Part of the reason this is the way it is, is that D&D 5e 2024 is still built on an ever-increasing escalation of old-school mechanical design that continues to be used in a modern game in a way that it was never intended to be used.
Not really. The core problem is that they came up with this 'bounded accuracy'* concept... and then failed to actually make it work within the game system. The reality is, if you give PCs new toys as they gain experience, and some of those toys work on skill checks... skill checks that are hard at level 1 are no longer hard at higher level. That's just the way the math works. In addition, they threw out the tools that 3e and 4e used to (somewhat) control this escalation -- for example, in 4e, guidance, bardic inspiration, flash of genius, etc, would not stack, because they would all be providing the same bonus type and bonuses of the same type don't stack.
*note that bounded accuracy is based on a fundamental misunderstanding. 3e and 4e didn't scale DC based on the level of the party, they scaled DC based on the level of the challenge.
Fully agree, the bound accuracy architecture is fine, but the implementation is flawed. Still I would argue that the base problem is that most characters get a default +6 on checks, +4 for prime attribute which is always going to be at 18 and +2 proficiency bonus. That alone is already way too many bonuses given that the basic DC's are 10-15. Even a hard task at DC 20 means a 1st level character with no other benefits is succeeding on a 14. It all escalates from there.
With lower ability scores, the math in all but the most extreme cases is going to be hard to break. Im not saying its perfect, but you get the side benefit of getting other balancing elements with lower ability scores.
The assumption in modern gaming culture is that if you are X class your Y attribute should be 18 and your skill choices are all based on that attribute. This is just standard character creation at this point, min-maxing is pretty much normalized at this point. The simplest fix for that is simply break that part of the game.
I've done this in my games and it just works. I make no other changes, I just make people roll 3d6 down the chain and that fixes the majority of the problems caused by mathematics.
It's not popular at first and I will grant you, players accustomed to having scores between 14-18 will complain, but in the end it just makes for a better game overall and if everyone is in the same boat, its all relative. You just have to get players to accept that 14 is the new 18.
I hear what people say about "what's impossible to normal people isn't to the heroes" but I just wish the DMG said THAT, not "Low-level characters have no chance to accomplish a DC 30 task, while a level 20 character with proficiency and a relevant ability score of 20 still needs a 19 or 20 on the die roll to succeed at a task of this difficulty."
A 7th level wizard with expertise in Arcana who has a Int of 18 and has help from the Warlock and guidance from the Druid has .4 probability of success of accomplishing a nearly impossible Arcana task.
A 7th level rogue with expertise in stealth who has a Dex of 18 and the wizard has cast Pass without trace has a .55 probability of success of accomplishing a nearly impossible Stealth task.
That reality is not reflected in the advice to DMs given in the DMG.
Isn't this kind of what the game is all about though? The team coming together to solve problems that any one of them couldn't solve alone? I mean what point is there to Pass Without a Trace if it doesn't help to make the impossible possible? That's like the definition of magic. You want 2-3 characters to expend resources and still fail as much as they would without those resources?
I do think it's kind of lame that they pretend the game doesn't scale at all and one set of DCs will hold throughout all levels, but I don't think those DCs should be designed for things to still be unlikely to succeed after allies pool their features and resources. The game is designed to reward exactly that kind of thing.
I personally think that scaling DC's to level is the worst kind of design. It basically means leveling up is meaningless and when my group played PF2e, that is exactly how it felt. We got more powerful, but the difficulty scaled to our level so in the end, we might as well just stay level 1 and not bother with the complexity that comes with having lots of feats and abilities. It just makes the math irrelevant.
I think teamwork, using abilities and clever combos should be part of the game, but I also think the math needs to be more tuned up and again, the easiest way to do this is just to lower the "plus" scale. 3d6, I stand by it, its the answer to this problem. With 3d6 the players can power game, optimize and min-max to their hearts content and the game remains balanced to a far great degree.
The general rule is that the players should be complaining that they aren't powerful enough... always. It should always be a struggle, there should always be a high chance of failure and through clever tactics, combos and efficiency they can put the odds in their favor but never gain guarantees.
At least that is my view on it. Easy does not make for a good game, hard is always better.
I personally think that scaling DC's to level is the worst kind of design.
It has to be clear that it's based on the level of the challenge, not the level of the characters. Giving the PCs harder skill challenges is no different than how at level 1 you might set them against a a handful of goblins, whereas at level 10 you might use the same number of hill giants.
Think of it more like how at level 1 you are trying to break down the wooden door of a tavern, and at level 20 you're trying to break down the spell-reinforced, adamantine door of the archlich's inner sanctum. It's not just arbitrarily higher numbers, it's a reflection of the greater challenges a higher level party takes on. The numbers are just numbers, it's up to the GM to put those numbers into a context that makes sense within the game world.
In a scaling game with groups of characters whose skills and features can vary widely, it's dumb to even have the terms "easy," "difficult," or "impossible" when talking about DCs. As Penta said above, it should be thought about as what challenges the party at their current state. And it's up to the DM to justify those increasing numbers though escalating opposition.
And obviously this doesn't mean everything they face has a 50-60% chance of success. This isn't Skyrim. It's fine to throw in easier or harder stuff where it makes sense.
You ... kinda try to get a 12+ on the die - always, for all levels - to be sure you succeed. I mean, your mileage may vary, but that should be kinda the right sorta spot. Thereabouts.
And that's fine. A little silly, but also fine. Because otherwise we get a situation where players (you know, hypothetically) would go 'man, these goblins were really challenging at level 1, but now that we're all level 20, I'm just not feeling it anymore.'
Rollback Post to RevisionRollBack
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
One strength of the DMG is that they do a good job explaining what challenges are appropriate at different tiers of play.
It would be nice to see examples of this tied to skills. The skills system is really just tacked on, and while combat they have so many ways to give you minor boosts and tactial shifts, with skills it's like "what if three of the classes just blasted out of the water."
I get that skills are tough because they have to represent so many different things, but they got started with the DC chart for Travel Terrain. I was thinking they could definitely layer on some templates that you could layer on to those base travel scores to show that at tier 1 it's tricky to cross a desert, but at tier 3 crossing a desert is just described in narration, but when they try and cross the desert of the Mummy king infused with necromantic magic it totally shifts the DCs, and will require spells and some members holding expertise to navigate the desert.
Basically what I'm saying is a future Dungeon Master focused supplement could explore what skill checks look like at different tiers of play, and spend some time remembering that Bards, Rangers, and Rogues can hit some Very Hard and Impossible scores, and how to create tasks that allow everyone to participate in a skill exploration section while one character just has bonkers good skills.
One way I've thought of for resolving challenges is to treat them with the rarity system. e.g.
Common quality lock: DC 15
Uncommon quality lock: DC 20
Rare quality lock: DC 25
Very Rare quality lock: DC 30
Legendary quality lock: DC 35+
Artifact quality lock: plot device difficulty.
The other option is to emulate 4th edition and give something like a lock or a trap a challenge rating (so a CR 2 lock is expected to slow down the party about as much as a CR 2 guard), but it's harder to assign good difficulties to that.
To echo those above, there are elements of intentional design as well as flawed implementation. My understanding on the intention is:
DC 10 = medium DC 15 = hard DC 20 = very hard DC 25 = near impossible DC 30 = impossible
As players level up the types of challenges they face change. So a level 15 party would absolutely breeze through a quest designed for a level 2 party including the skill checks - the lock on a trunk in the basement of a tavern infested with rats should be on par with the combat challenge of those rats and a level 15 party shouldn't even need to roll to unlock it. However, a level 15 party shouldn't be going around killing rats in the basement of taverns anymore. A level 15 party should be breaking into the vault of a Pit Fiend which should have a much more challenging lock than that trunk in the basement of a tavern. So the DCs a level 15 party should be facing for skill checks should be in the 20-30 range but not because you are just scaling up the difficulty, but because they are taking on more difficult problems.
That said, these DCs were designed for a game governed by bounded accuracy where almost all buffs were simple Adv/Disadvantage so didn't stack, and there were few flat bonuses that could stack, which is not at all how the game actually plays anymore.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
So in the DMG 2024 they repeat the same DCs, but they say that "nearly impossilbe" needs a 19 or 20 to succeed with a high enough proficency score, but this seems like it fails to account how much more likely expertise is to get, with Rangers and Wizards now getting expertise (and Druids and Clerics getting a little expertise). Starting at level 15 most of the expert characters should be able to succeed on a 15 at nearly impossible, skill, which if they get advantage comes closer to rolling a 10 or higher for something labeled "nearly impossible." I'm not totally against it, I jsut wonder if the description of the nearly impossible is innacurate and might make DM's surprised when it actually is easier than stated at high levels.
Well - what's 'impossible' to the man on the street, and a 15th level hero should be two rather different things.
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
No, it remains nearly impossible == the PCs are just that good.
Dang nabbit.
Only a DM since 1980 (3000+ Sessions) / PhD, MS, MA / Mixed, Bi, Trans, Woman / No longer welcome in the US, apparently
Wyrlde: Adventures in the Seven Cities
.-=] Lore Book | Patreon | Wyrlde YT [=-.
An original Setting for 5e, a whole solar system of adventure. Ongoing updates, exclusies, more.
Not Talking About It / Dubbed The Oracle in the Cult of Mythology Nerds
As long as you recognize that typical games stop by level 10, DC 30 is nearly impossible for most builds and difficult for almost all builds. In a game that goes to level 20, there's a good chance the PCs have the tools to succeed at DC 40 checks.
I think the point of characters at higher levels is that doing the impossible is "normal" for them in general. I mean, these are characters fighting multiple dragons at a time and winning, dropping meteors down from the heavens with a flick of a wrist and cheating death with a few command words. Its not weird that they can do the impossible with skills as well.
I agree with what everyone else has said; doing the impossible is what higher level player characters are supposed to do. Taking a Rogue with thieves tools as the obvious example how likely is it that with years of experience and practice that all of a sudden they'll struggle to pick a fairly standard lock? They should be opening as quickly and easily as I would using the key
You have two options if it's unsatisfactory.
1. Move all of the DCs in the table up by 10. Yes, 10. That is how much is needed if you want them to actually be a challenge at level 15.
2. Scalable DCs. Pathfinder does this and has DCs set by level banding. It works very well to provide challenge at higher level play.
Secret Option 3: Keep the DCs and stop relying on locked chests and hidden doors. This sounds harsh, but by level 10, the types of areas that your player characters are likely to be exploring might legitimately have a chest that is trapped, magically warded, and locked. Though the other thing is that by level 10 there really isn't going to be much in a chest that is going to be exciting for a player character.
DM session planning template - My version of maps for 'Lost Mine of Phandelver' - Send your party to The Circus - Other DM Resources - Maps, Tokens, Quests
Actor, Writer, Director & Teacher by day - GM/DM in my off hours.
I think a more accurate description of difficulty is along the lines of
(these numbers are not the product of extensive analysis, so YMMV, but remember that there are a lot of skills; just because a party can make DC 30 stealth checks doesn't mean they can make DC 20 religion checks).
I hear what people say about "what's impossible to normal people isn't to the heroes" but I just wish the DMG said THAT, not "Low-level characters have no chance to accomplish a DC 30 task, while a level 20 character with proficiency and a relevant ability score of 20 still needs a 19 or 20 on the die roll to succeed at a task of this difficulty."
A 7th level wizard with expertise in Arcana who has a Int of 18 and has help from the Warlock and guidance from the Druid has .4 probability of success of accomplishing a nearly impossible Arcana task.
A 7th level rogue with expertise in stealth who has a Dex of 18 and the wizard has cast Pass without trace has a .55 probability of success of accomplishing a nearly impossible Stealth task.
That reality is not reflected in the advice to DMs given in the DMG.
Part of the reason this is the way it is, is that D&D 5e 2024 is still built on an ever-increasing escalation of old-school mechanical design that continues to be used in a modern game in a way that it was never intended to be used.
Strictly speaking the ability scores of a typical 1e character were between 8-14. An 18 in Dex was exceedingly unlikely. In modern D&D, you are all but guaranteed to have at least one stat at 18 regardless of the method used. Furthermore, an 18 in 1e meant you would have at best a +3 modifier in modern-day 18 = +4 and its possible to get that up to 20 for a +5. More importantly, the ability score system was designed to be a roll-under system, not a DC-based system.
So basically modern D&D has taken a design and used it in a way for which it was not intended. Fast forward 50 years later with several iterations of modernized power escalation and you end up with a bit of a mess.
Fortunately old school application still fixes the problem.
When you make characters, roll 3d6 down the chain. Bring the power curve down, problem is solved. Your players will not complain how easy it is, but how hard and that is really where it should be. This solves a lot of other problems like CR math and encounters balanced, power-game longevity problems and pretty much everything else people complain about when it comes to 5e. It really is that simple.
Not really. The core problem is that they came up with this 'bounded accuracy'* concept... and then failed to actually make it work within the game system. The reality is, if you give PCs new toys as they gain experience, and some of those toys work on skill checks... skill checks that are hard at level 1 are no longer hard at higher level. That's just the way the math works. In addition, they threw out the tools that 3e and 4e used to (somewhat) control this escalation -- for example, in 4e, guidance, bardic inspiration, flash of genius, etc, would not stack, because they would all be providing the same bonus type and bonuses of the same type don't stack.
*note that bounded accuracy is based on a fundamental misunderstanding. 3e and 4e didn't scale DC based on the level of the party, they scaled DC based on the level of the challenge.
Isn't this kind of what the game is all about though? The team coming together to solve problems that any one of them couldn't solve alone? I mean what point is there to Pass Without a Trace if it doesn't help to make the impossible possible? That's like the definition of magic. You want 2-3 characters to expend resources and still fail as much as they would without those resources?
I do think it's kind of lame that they pretend the game doesn't scale at all and one set of DCs will hold throughout all levels, but I don't think those DCs should be designed for things to still be unlikely to succeed after allies pool their features and resources. The game is designed to reward exactly that kind of thing.
My homebrew subclasses (full list here)
(Artificer) Swordmage | Glasswright | (Barbarian) Path of the Savage Embrace
(Bard) College of Dance | (Fighter) Warlord | Cannoneer
(Monk) Way of the Elements | (Ranger) Blade Dancer
(Rogue) DaggerMaster | Inquisitor | (Sorcerer) Riftwalker | Spellfist
(Warlock) The Swarm
Fully agree, the bound accuracy architecture is fine, but the implementation is flawed. Still I would argue that the base problem is that most characters get a default +6 on checks, +4 for prime attribute which is always going to be at 18 and +2 proficiency bonus. That alone is already way too many bonuses given that the basic DC's are 10-15. Even a hard task at DC 20 means a 1st level character with no other benefits is succeeding on a 14. It all escalates from there.
With lower ability scores, the math in all but the most extreme cases is going to be hard to break. Im not saying its perfect, but you get the side benefit of getting other balancing elements with lower ability scores.
The assumption in modern gaming culture is that if you are X class your Y attribute should be 18 and your skill choices are all based on that attribute. This is just standard character creation at this point, min-maxing is pretty much normalized at this point. The simplest fix for that is simply break that part of the game.
I've done this in my games and it just works. I make no other changes, I just make people roll 3d6 down the chain and that fixes the majority of the problems caused by mathematics.
It's not popular at first and I will grant you, players accustomed to having scores between 14-18 will complain, but in the end it just makes for a better game overall and if everyone is in the same boat, its all relative. You just have to get players to accept that 14 is the new 18.
I personally think that scaling DC's to level is the worst kind of design. It basically means leveling up is meaningless and when my group played PF2e, that is exactly how it felt. We got more powerful, but the difficulty scaled to our level so in the end, we might as well just stay level 1 and not bother with the complexity that comes with having lots of feats and abilities. It just makes the math irrelevant.
I think teamwork, using abilities and clever combos should be part of the game, but I also think the math needs to be more tuned up and again, the easiest way to do this is just to lower the "plus" scale. 3d6, I stand by it, its the answer to this problem. With 3d6 the players can power game, optimize and min-max to their hearts content and the game remains balanced to a far great degree.
The general rule is that the players should be complaining that they aren't powerful enough... always. It should always be a struggle, there should always be a high chance of failure and through clever tactics, combos and efficiency they can put the odds in their favor but never gain guarantees.
At least that is my view on it. Easy does not make for a good game, hard is always better.
It has to be clear that it's based on the level of the challenge, not the level of the characters. Giving the PCs harder skill challenges is no different than how at level 1 you might set them against a a handful of goblins, whereas at level 10 you might use the same number of hill giants.
Think of it more like how at level 1 you are trying to break down the wooden door of a tavern, and at level 20 you're trying to break down the spell-reinforced, adamantine door of the archlich's inner sanctum. It's not just arbitrarily higher numbers, it's a reflection of the greater challenges a higher level party takes on. The numbers are just numbers, it's up to the GM to put those numbers into a context that makes sense within the game world.
In a scaling game with groups of characters whose skills and features can vary widely, it's dumb to even have the terms "easy," "difficult," or "impossible" when talking about DCs. As Penta said above, it should be thought about as what challenges the party at their current state. And it's up to the DM to justify those increasing numbers though escalating opposition.
And obviously this doesn't mean everything they face has a 50-60% chance of success. This isn't Skyrim. It's fine to throw in easier or harder stuff where it makes sense.
My homebrew subclasses (full list here)
(Artificer) Swordmage | Glasswright | (Barbarian) Path of the Savage Embrace
(Bard) College of Dance | (Fighter) Warlord | Cannoneer
(Monk) Way of the Elements | (Ranger) Blade Dancer
(Rogue) DaggerMaster | Inquisitor | (Sorcerer) Riftwalker | Spellfist
(Warlock) The Swarm
You ... kinda try to get a 12+ on the die - always, for all levels - to be sure you succeed. I mean, your mileage may vary, but that should be kinda the right sorta spot. Thereabouts.
And that's fine. A little silly, but also fine. Because otherwise we get a situation where players (you know, hypothetically) would go 'man, these goblins were really challenging at level 1, but now that we're all level 20, I'm just not feeling it anymore.'
Blanket disclaimer: I only ever state opinion. But I can sound terribly dogmatic - so if you feel I'm trying to tell you what to think, I'm really not, I swear. I'm telling you what I think, that's all.
One strength of the DMG is that they do a good job explaining what challenges are appropriate at different tiers of play.
It would be nice to see examples of this tied to skills. The skills system is really just tacked on, and while combat they have so many ways to give you minor boosts and tactial shifts, with skills it's like "what if three of the classes just blasted out of the water."
I get that skills are tough because they have to represent so many different things, but they got started with the DC chart for Travel Terrain. I was thinking they could definitely layer on some templates that you could layer on to those base travel scores to show that at tier 1 it's tricky to cross a desert, but at tier 3 crossing a desert is just described in narration, but when they try and cross the desert of the Mummy king infused with necromantic magic it totally shifts the DCs, and will require spells and some members holding expertise to navigate the desert.
Basically what I'm saying is a future Dungeon Master focused supplement could explore what skill checks look like at different tiers of play, and spend some time remembering that Bards, Rangers, and Rogues can hit some Very Hard and Impossible scores, and how to create tasks that allow everyone to participate in a skill exploration section while one character just has bonkers good skills.
One way I've thought of for resolving challenges is to treat them with the rarity system. e.g.
The other option is to emulate 4th edition and give something like a lock or a trap a challenge rating (so a CR 2 lock is expected to slow down the party about as much as a CR 2 guard), but it's harder to assign good difficulties to that.
To echo those above, there are elements of intentional design as well as flawed implementation. My understanding on the intention is:
DC 10 = medium
DC 15 = hard
DC 20 = very hard
DC 25 = near impossible
DC 30 = impossible
As players level up the types of challenges they face change. So a level 15 party would absolutely breeze through a quest designed for a level 2 party including the skill checks - the lock on a trunk in the basement of a tavern infested with rats should be on par with the combat challenge of those rats and a level 15 party shouldn't even need to roll to unlock it. However, a level 15 party shouldn't be going around killing rats in the basement of taverns anymore. A level 15 party should be breaking into the vault of a Pit Fiend which should have a much more challenging lock than that trunk in the basement of a tavern. So the DCs a level 15 party should be facing for skill checks should be in the 20-30 range but not because you are just scaling up the difficulty, but because they are taking on more difficult problems.
That said, these DCs were designed for a game governed by bounded accuracy where almost all buffs were simple Adv/Disadvantage so didn't stack, and there were few flat bonuses that could stack, which is not at all how the game actually plays anymore.