Common issues with DC based skill systems

[Zuras]

In the Pathfinder discussion thread, there is a vigorous discussion about the Pathfinder 2e skill system, particularly issues with arbitrarily set difficulty levels.

I’m interested in everyone’s general experience with DC based systems, but rather than clutter that discussion further, I decided a separate thread was in order.

What problems do you regularly encounter with DC based skill/resolution systems?

I’ve seen many people complain that they can make RPGs devolve into a game of mother-may-I, and while I clearly see how this could happen, it’s not something I’ve ever encountered myself. Even in systems where you’re literally making everything up as you go along, like FATE, the table always reached a consensus that a given task difficulty was fair and the aspects a player applies to it made narrative sense.

The problems I *have* encountered, and have to periodically guard myself against as a GM have actually been:

a) forgetting that requiring a series of multiple successes dramatically worsens the odds for any task, and

b) Instinctively scaling up challenges so there is still some chance of failure, without any in-universe logic. I have to guard myself against making every lock in town suddenly masterwork difficulty after the rogue gets Reliable Talent.

c) Relatedly, accidentally gating all non-specialist characters out of participating by setting the DCs to be a reasonable challenge for the party expert (especially in games with constant scaling/no bounded accuracy). This is fine for tasks like developing a cure for the plague virus, but terrible for many social and exploration tasks where other PCs should reasonably be able to participate.

[False God]

a) forgetting that requiring a series of multiple successes dramatically worsens the odds for any task, and

See I personally don't see this as problematic, and I often run skill challenges as "three successes before 3 failures" and it helps substantially with your secondary problem. It also can break a scene down more incrementally, it can be useful in occupying a character's time while the enemy acts making choices more important and scenes more dynamic. You can add levels of "the next roll is easier because you succeeded" or "its harder because you failed" without having to say "Oh you failed at the lock now you're done." on a single check.

b) Instinctively scaling up challenges so there is still some chance of failure, without any in-universe logic. I have to guard myself against making every lock in town suddenly masterwork difficulty after the rogue gets Reliable Talent.

D&D and d20 systems teach DMs to have rolls for everything. Frankly, if the DC is lower than a characters skill points, that character doesn't need to make a check. Challenges can also be "scaled up" with the multi-roll method without actually increasing the DC, and IMO, multi-stage checks often result in better challenges.

c) Relatedly, accidentally gating all non-specialist characters out of participating by setting the DCs to be a reasonable challenge for the party expert (especially in games with constant scaling/no bounded accuracy). This is fine for tasks like developing a cure for the plague virus, but terrible for many social and exploration tasks where other PCs should reasonably be able to participate.

Again, I find it far more useful to gate than to not, even when it may seem unreasonable to do so. The Party has, in MTG terms "card advantage" and I prefer to gate than to simply let them throw dice at the challenge until someone breaks through. I've had far more unhappy players in situations where their specialist character was overwhelmed by someone else's lucky dice and they're denied the ability to do the thing they basically built a character for.

If the non-specialized party members want to help, they can help the specialist. Secondary roles "lab assistant", "backup dancer", "useful idiot" and more are all grand ways to assist specialists during gated events. There are also other tasks available during most skill checks, everyone should be able to do something they are good at, even if they aren't participating in what they feel is the "main event".

----
I think one big failing of the d20 system is that it encourages "umbrella" checks, "Roll Nature to detect nature stuff." when you could both "gate" and allow group participation by allowing a multitude of checks, "Nature", "Geography", "Local", "Appraise", professions like hunting or woodworking, could all be used to allow everyone to pick up different elements around them, rather than either gating half the people out or brute-forcing it with a bunch of rolls against a single skill.

The nature of the "one roll, pass or fail" system d20 games encourage results in a lot of meaningless mundane rolls, and very few important but highly difficult rolls. Often resulting in players feeling "bogged down" with little unimportant minutia, and then feeling stonewalled when it comes to elements that actually matter.

[kyoryu]

lve into a game of mother-may-I, and while I clearly see how this could happen, it’s not something I’ve ever encountered myself. Even in systems where you’re literally making everything up as you go along, like FATE, the table always reached a consensus that a given task difficulty was fair and the aspects a player applies to it made narrative sense.

Practically speaking, this matches my experience. I also play a lot of Fate, so that's unsurprising :)

I'm not sure what the actual problems with this are. There's a lot of "values" tossed about, but I haven't seen a lot of "this actually has made the game worse" kinds of things.

a) forgetting that requiring a series of multiple successes dramatically worsens the odds for any task, and

b) Instinctively scaling up challenges so there is still some chance of failure, without any in-universe logic. I have to guard myself against making every lock in town suddenly masterwork difficulty after the rogue gets Reliable Talent.

c) Relatedly, accidentally gating all non-specialist characters out of participating by setting the DCs to be a reasonable challenge for the party expert (especially in games with constant scaling/no bounded accuracy). This is fine for tasks like developing a cure for the plague virus, but terrible for many social and exploration tasks where other PCs should reasonably be able to participate.

Yeah. I've seen all of these actual problems show up, frequently.

[NichG]

In the Pathfinder discussion thread, there is a vigorous discussion about the Pathfinder 2e skill system particularly issues with arbitrarily set difficulty levels.

I’m interested in everyone’s general experience with DC based systems, but rather than clutter that discussion further, I decided a separate thread was in order.

What problems do you regularly encounter with DC based skill/resolution systems?

Yes to all the problems you mentioned, but in addition:

- Whether or not retry is possible/whether its possible to make retry trivial becomes an overly important thing the bigger the variance on the dice gets with respect to modifiers/DCs. So you have to almost have two different designs - a set of DCs appropriate for things that are usually one-off checks, and a separate set of DCs (usually a quarter or half dice size higher) for things that are usually going to be able to be retried at leisure. When those 'usually' assumptions get broken, it ends up being very impactful, maybe overly so compared to what you're trying to have skill represent.

- More subtly, DC based skill systems push one into a binary success/failure mindset, which tends to have a much smaller dynamic range with regards to differences in character skill. If 'failure' is the only stakes then you either have e.g. professionals frequently totally failing to do certain things that should be rote for them OR basically skill advancement has to be such that the variance of the dice can easily be made negligible, in which case once you've hit that level of skill there's sharply reduced returns for becoming even more skillful, and before you hit that level its almost like 'don't bother trying'.

So something where the roll determines the cost to succeed or the extent of success or even just numerical parameters of the outcome can be a lot smoother with regards to how it feels to advance/be skilled/be a novice/etc, while still letting characters be consistent and predictable in what they can achieve when it makes sense. E.g. rather than 'pick the distance you want to jump, roll to see if you can jump it', it'd be more like 'roll to see the distance you can jump for free, and spend 1 stamina point per +1ft you need to add to that' or 'you do/do not have the skill to jump it; if you do, then roll to see if you lose an action regaining your balance after landing'

[PhoenixPhyre]

The big issues I see are

1. Rolling where there aren't interesting consequences for both/all results. Consequences that move the story along and change the scene. Retrying a check shouldn't really ever be a thing, at least not straight up. If it deserves a roll, the outcome of that roll should matter.

2. Along with that, lumping too big of outcomes on single points of failure. Interesting consequences should be incremental. In a D&D-like, you wouldn't tie the outcome of an entire combat to a single attack roll. Or generally shouldn't. Each small check moves the combat onward; missing has consequences because it lets the other person do bad things longer. But missing once doesn't mean you failed. And it's not an arbitrary count of sucessess/failures. "Skill challenges" should have a similar structure--a sequence of things to interact with, possibly in parallel. Something everyone can interact with, the success or failure of each of which influence the final outcome more than just modulating check difficulty. Embrace the branches.

3. The multi-headed hydra party. Where everyone has their specialty and is incompetent outside of it, so the DM lets them swap in the specialist for every check but otherwise act like one creature. Everyone faces skill challenges; everyone should participate. He who asked the question deals with the check. In social situations, the NPCs may not always respond to the person with the highest social skills. Yes, that means that dumping those skills has negative consequences and a well-rounded person is generally better than a narrow specialist. Adventurers are generalists, not specialists.

4. "Me too" checks, where everyone acts like one success is enough for the whole party (a variant of #3). In those cases where there's no reason why they wouldn't talk, either do a group check (50+% have to succeed to succeed) or just give them the information based on the best person's outcome.

[kyoryu]

See I personally don't see this as problematic, and I often run skill challenges as "three successes before 3 failures" and it helps substantially with your secondary problem. It also can break a scene down more incrementally, it can be useful in occupying a character's time while the enemy acts making choices more important and scenes more dynamic. You can add levels of "the next roll is easier because you succeeded" or "its harder because you failed" without having to say "Oh you failed at the lock now you're done." on a single check.

I think the issue is more like "multiple rolls in a roll into you succeed/fail scenario". Best of 3, x successes before y failures, etc. are good ways of handling multiple rolls. "You have to sneak past 10 guards, they each get their own roll" probably isn't.

I think one big failing of the d20 system is that it encourages "umbrella" checks, "Roll Nature to detect nature stuff." when you could both "gate" and allow group participation by allowing a multitude of checks, "Nature", "Geography", "Local", "Appraise", professions like hunting or woodworking, could all be used to allow everyone to pick up different elements around them, rather than either gating half the people out or brute-forcing it with a bunch of rolls against a single skill.

Agreed. In general I prefer the approach of "tell me what you're actually doing and I'll tell you what to roll."

- Whether or not retry is possible/whether its possible to make retry trivial becomes an overly important thing the bigger the variance on the dice gets with respect to modifiers/DCs. So you have to almost have two different designs - a set of DCs appropriate for things that are usually one-off checks, and a separate set of DCs (usually a quarter or half dice size higher) for things that are usually going to be able to be retried at leisure. When those 'usually' assumptions get broken, it ends up being very impactful, maybe overly so compared to what you're trying to have skill represent.

My preference is, generally, to set the stakes of the roll such that failure means "you've hit the failure/decision point". This doesn't work well in systems that clearly define what "failure" means.

- More subtly, DC based skill systems push one into a binary success/failure mindset, which tends to have a much smaller dynamic range with regards to differences in character skill. If 'failure' is the only stakes then you either have e.g. professionals frequently totally failing to do certain things that should be rote for them OR basically skill advancement has to be such that the variance of the dice can easily be made negligible, in which case once you've hit that level of skill there's sharply reduced returns for becoming even more skillful, and before you hit that level its almost like 'don't bother trying'.

Yeah. I prefer to frame things as "does this go well, or poorly?" That creates a lot more dynamic potential and gets rid of a lot of those problems. If you're very competent, success at the primary task may be presumed.... things going well or poorly may really be about secondary effects.

The big issues I see are

1. Rolling where there aren't interesting consequences for both/all results. Consequences that move the story along and change the scene. Retrying a check shouldn't really ever be a thing, at least not straight up. If it deserves a roll, the outcome of that roll should matter.

Agreed. A lot of times that's because the system has other failure states built into it with incremental costs, and what you're really doing is collapsing that.

Like, in D&D, searching for a secret door takes time. In a dungeon, that means there's a chance of a random encounter. The real "failure" cost isn't not finding the door - it's getting a random encounter. Collapsing that into one roll just means you make "you're still searching when the monster shows up" the consequence of failure. (Or, more nuanced, "you're still searching when you hear something approach from afar, do you keep at it or do you head out?")

2. Along with that, lumping too big of outcomes on single points of failure. Interesting consequences should be incremental. In a D&D-like, you wouldn't tie the outcome of an entire combat to a single attack roll. Or generally shouldn't. Each small check moves the combat onward; missing has consequences because it lets the other person do bad things longer. But missing once doesn't mean you failed. And it's not an arbitrary count of sucessess/failures. "Skill challenges" should have a similar structure--a sequence of things to interact with, possibly in parallel. Something everyone can interact with, the success or failure of each of which influence the final outcome more than just modulating check difficulty. Embrace the branches.

I'm okay either way, depending on how much you want to focus on it. I'd actually like a one-roll combat option. But, either way, there needs to be a way to say "this is critical, let's spend time on it" or "this isn't so critical, let's not", both of which have a wider range of results than the binary.

3. The multi-headed hydra party. Where everyone has their specialty and is incompetent outside of it, so the DM lets them swap in the specialist for every check but otherwise act like one creature. Everyone faces skill challenges; everyone should participate. He who asked the question deals with the check. In social situations, the NPCs may not always respond to the person with the highest social skills. Yes, that means that dumping those skills has negative consequences and a well-rounded person is generally better than a narrow specialist. Adventurers are generalists, not specialists.

Well, I generally do think that the specialists should do what they specialize in. That makes sense. It may make sense that that doesn't always work - the king may want to talk to the barbarian for some reason.

4. "Me too" checks, where everyone acts like one success is enough for the whole party (a variant of #3). In those cases where there's no reason why they wouldn't talk, either do a group check (50+% have to succeed to succeed) or just give them the information based on the best person's outcome.

Group checks should be that. "They failed, let me try" doesn't work for me. This gets back to having actual stakes for failure. (In that case, a second character might try, but you still have the consequences of the first failure to deal with).

Things like stealth checks, I tend to go with "the slowest and loudest". IOW, base the check on the worst person, and let the higher skilled people assist as appropriate.

[PhoenixPhyre]

Well, I generally do think that the specialists should do what they specialize in. That makes sense. It may make sense that that doesn't always work - the king may want to talk to the barbarian for some reason.

In the absence of other factors, yes. The specialist is the one most likely to try things that will likely cause them to make the kind of rolls they're specialized in. But that doesn't mean that only the arcane specialist ever asks "what do these runes mean" or that the arcane specialist can jump in when the barbarian asks that question and take their place.

If time and circumstance permit, the barbarian may call over the arcane specialist and have them try to figure it out. Or they may not. But if they don't...the barbarian makes any rolls necessary.

[Satinavian]

I’m interested in everyone’s general experience with DC based systems, but rather than clutter that discussion further, I decided a separate thread was in order.

Potential problems :
1) The DC tables might not properly work out when you use the values typical everymen or professionals in that job have and calculate the odds.

2) The DC tables might be to vague or not even exist

3) There might be weird skaling going on.

4) Rndomness for the sake of excitement, no matter whether it makes sense.

[AceOfFools]

The big problems I’ve encountered with DC-skill systems are:

Arbitrary difficulty: There was an official DnD released adventures for 4e that had the solid, reinforced door of the mayor’s office ve easier to knock down than the cracked, door of the run-down, falling apart slum you found 2 levels later. It makes it really hard to invest in a world as a place that exists as anything other than an abstraction.

Excessive rules: on the flip side, there are formulas for the difficulty one needs to know or reference for basically every skill in the d20 system. If you want to have rules to set DC based on the narrative reality of the game, well, someone needs to remember or keep handy a cheat sheet for all of those rules, which no one likes.

Bad balance: I ran a few d20 modern games in a very DnD style: X-combat encounters per day, clever encounter spaced that monsters—yes, monsters—could interact with. It quickly became clear that this was not how the developers intended the game to be run, because the Smart Hero class was devoted entirely to “get high numbers on specific skill checks,” and they were a liability in any environment where rolling those skills wasn’t the main way to interact with the world. Which then reached a problem with:

Serial play: If you need to hit high numbers to do a thing: hackers to clear security cameras, faces to negiotate truces, etc, AND you need to specialize build resources such that only highly specialized characters can hit those numbers, it’s very easy to find your players waiting, doing nothing while the character who can do the thing does the thing, and thus only doing things serially. This is hardly unique to DC skill systems, but it is a possible consequence of them.

Conceptual lock out: One complaint I have with 5e’s bounded accuracy is that it’s really hard to get more than 3~5 points better at a skill than someone who doesn’t have this skill as core to their character, but does need the base stat for build reasons (e.g. a fighter whose not a terrible fighter cannot be that much better a lawyer than a wizard, assuming you use an int based skill for law). This is not a hypothetical example, it annoyed me in actual play.

You might notice that not ever DC-based game has all of these issues in equal amounts. There are ways to mitigate them, but these generally introduce other trade offs, because game design is like that. Nothing works for everyone.

[Tanarii]

1) generating some kind of success or failure state when the attempted activity should be automatic (success or failure).

2) something you should be able to attempt is denied because there is a skill for it and you don't have it / aren't trained to a sufficient level in it.

These were particularly visible in the old days when Thieves skills were a (incredibly low) fixed percentage.
DMs regularly called for the checks for any activity that sounded like it involved them, regardless of the details of the attempted action reasonably mandating success.
DMs saying that non-thieves couldn't attempt actions at all such as sneaking or hiding, because Move Silently and Hide in Shadows existed.

Various systems attempt to solve this in various ways, usually by some variant of complex rules to only rolling "when it matters".

(Minor 3) when the probability of success/failure is a weird range once you determine a roll should happen. This is especially common in % or d20 systems.

[Zuras]

1) generating some kind of success or failure state when the attempted activity should be automatic (success or failure).

2) something you should be able to attempt is denied because there is a skill for it and you don't have it / aren't trained to a sufficient level in it.

These were particularly visible in the old days when Thieves skills were a (incredibly low) fixed percentage.
DMs regularly called for the checks for any activity that sounded like it involved them, regardless of the details of the attempted action reasonably mandating success.
DMs saying that non-thieves couldn't attempt actions at all such as sneaking or hiding, because Move Silently and Hide in Shadows existed.

Various systems attempt to solve this in various ways, usually by some variant of complex rules to only rolling "when it matters".

(Minor 3) when the probability of success/failure is a weird range once you determine a roll should happen. This is especially common in % or d20 systems.

The awfulness of the Thief skills table is one of the reasons I’ve never been able to stomach any of the faithful AD&D retroclones.

[kyoryu]

If time and circumstance permit, the barbarian may call over the arcane specialist and have them try to figure it out. Or they may not. But if they don't...the barbarian makes any rolls necessary.

I think a lot of rolls are made very abstractly. I think that's a mistake - they should be tied directly to the situation and what is actually happening in the game world.

Usually that will let the specialist specialize. Sometimes it won't.

Conceptual lock out: One complaint I have with 5e’s bounded accuracy is that it’s really hard to get more than 3~5 points better at a skill than someone who doesn’t have this skill as core to their character, but does need the base stat for build reasons (e.g. a fighter whose not a terrible fighter cannot be that much better a lawyer than a wizard, assuming you use an int based skill for law). This is not a hypothetical example, it annoyed me in actual play.

This is really mostly a skill+stat thing, which is the default in a lot of systems. If you don't presume that, it creates a lot less lockout.

1) generating some kind of success or failure state when the attempted activity should be automatic (success or failure).

2) something you should be able to attempt is denied because there is a skill for it and you don't have it / aren't trained to a sufficient level in it.

(Minor 3) when the probability of success/failure is a weird range once you determine a roll should happen. This is especially common in % or d20 systems.

That's a lot of why I like to treat "can I do something?" as a separate mechanic from "how hard is it?", as compared to the "it's only too hard if your roll can't succeed". While that's conceptually tidy, I find it leads to a lot of issues.

Also, skill layout, levels of skill acquisition, default rules, etc. all factor into that.

[False God]

This is really mostly a skill+stat thing, which is the default in a lot of systems. If you don't presume that, it creates a lot less lockout.

This was a change I made to any of my d20-based home games after playing WoD. When I call for a roll, I'll tell you which skill and which stat (if any, maybe multiple choice) is applicable. It also dramatically reduces the need for high DCs.

[kyoryu]

This was a change I made to any of my d20-based home games after playing WoD. When I call for a roll, I'll tell you which skill and which stat (if any, maybe multiple choice) is applicable. It also dramatically reduces the need for high DCs.

I play a lot of Fate, and Fate just has skills. A lot of people want to use so-called "two column" Fate (which would combine skill + approach or stat), and I think that is a net negative, for a lot of the reasons that were pointed out.

[SimonMoon6]

c) Relatedly, accidentally gating all non-specialist characters out of participating by setting the DCs to be a reasonable challenge for the party expert (especially in games with constant scaling/no bounded accuracy). This is fine for tasks like developing a cure for the plague virus, but terrible for many social and exploration tasks where other PCs should reasonably be able to participate.

I think that's a feature not a bug.

If I'm playing a specialist at trap-finding, someone who has put a ton of skill points into trap-finding, someone who has taken a Skill Focus feat (or something similar) to be better at trap-finding, someone who has entered a prestige class to be even better at trap-finding, someone who has bought magic items to supplement their ability to find traps... and then Dumdum the Barbarian finds a trap before I do, then I think something is seriously wrong with the universe.

Granted, D&D 3.x fixes that problem by saying, "Oh, well, only rogues (and maybe other specialists) can find *magic* traps". But that's just another form of gate-keeping.

Why should I specialize at something if *everybody* has a 25% chance of success? With enough PCs, that's a virtual guarantee of success without any specialization.

Ideally, what should happen is that there will be some skill challenges with low DCs and some skill challenges with high DCs all in the same adventure together. So, sure Dumdum the Barbarian might have a chance to sneak past the drunken patrons of a bar to exit out the backdoor unseen (DC: 5), but only Sneaky McSneaksALot will be able to sneak past the watchful eyes of the half-Shadow Dragon Beholder to get to the Ultimate Treasure (DC: 75).

Or, Dumdum the Barbarian can climb over a decaying wooden fence (DC: 5), but only Climby McClimbsALot can climb up the sheer glass cliff covered in slippery oil (DC: 100).

So, yeah, *everyone* can sneak, but only one person can sneak when it's really important. Everyone can climb, but only one person can climb when it's important.

[KaussH]

1) generating some kind of success or failure state when the attempted activity should be automatic (success or failure).

2) something you should be able to attempt is denied because there is a skill for it and you don't have it / aren't trained to a sufficient level in it.

These were particularly visible in the old days when Thieves skills were a (incredibly low) fixed percentage.
DMs regularly called for the checks for any activity that sounded like it involved them, regardless of the details of the attempted action reasonably mandating success.
DMs saying that non-thieves couldn't attempt actions at all such as sneaking or hiding, because Move Silently and Hide in Shadows existed.

Various systems attempt to solve this in various ways, usually by some variant of complex rules to only rolling "when it matters".

(Minor 3) when the probability of success/failure is a weird range once you determine a roll should happen. This is especially common in % or d20 systems.

The Thieves skills were an easy fix. I always treated them like powers in the old days. So anyone could Hide, but be detected. but if you made your Hide in shadows then you couldnt just be detected, abilities needed to be used. If you had climb walls, you didnt roll for ropes and roof running, ect. It did not fix the detect traps and that, but thats a whole diff matter.
Now if you failed your Super power, then you were just as well hidden as any normal person, normal rules applied.

Now keep in mind this was for 1/2nd ed.

[PhoenixPhyre]

So, yeah, *everyone* can sneak, but only one person can sneak when it's really important. Everyone can climb, but only one person can climb when it's important.

Yeah, I don't like this. Because it means that only one person can participate at a time. It'd be like saying "everyone can fight, but in an important fight only the fighter can hit the enemy."

Everyone should be able to participate in everything. The "specialist" should be able to make the most progress and may be able to carry things alone for easy stuff, but any significant task should require multiple people working together. In that instance, the specialist's advantage is that he can cover some of the load for people who fail, meaning you might only need a couple of successful people instead of the whole team.

Of course, this means that conceptualizing "ability checks" as just one success to rule them all, with failures ignored needs to go away. Good. The way you keep from piling on is you...don't let people pile on. The person who takes the action does the thing. And then the situation changes before you ask the next person what they do. Every success, every "failure" has consequences. So succeeding 75% of the time instead of 25% of the time means you get the good consequences 75% of the time, not 25% of the time.

It also helps to drop the "one and done" approach. A situation that calls for ability checks and actually matters should require multiple checks, each of which moves the situation towards one of many end states. But none of which are individually decisive.

[stoutstien]

Yeah, I don't like this. Because it means that only one person can participate at a time. It'd be like saying "everyone can fight, but in an important fight only the fighter can hit the enemy."

Everyone should be able to participate in everything. The "specialist" should be able to make the most progress and may be able to carry things alone for easy stuff, but any significant task should require multiple people working together. In that instance, the specialist's advantage is that he can cover some of the load for people who fail, meaning you might only need a couple of successful people instead of the whole team.

Of course, this means that conceptualizing "ability checks" as just one success to rule them all, with failures ignored needs to go away. Good. The way you keep from piling on is you...don't let people pile on. The person who takes the action does the thing. And then the situation changes before you ask the next person what they do. Every success, every "failure" has consequences. So succeeding 75% of the time instead of 25% of the time means you get the good consequences 75% of the time, not 25% of the time.

It also helps to drop the "one and done" approach. A situation that calls for ability checks and actually matters should require multiple checks, each of which moves the situation towards one of many end states. But none of which are individually decisive.

Yea. The idea of the "skill guy" became so ingrained doesn't help either.

You either need to bring the range of results into a smaller range or ditch DCs/thresholds all together.

[Quertus]

I mean, there's a lot of bad implementations of DC-based skill systems, I'll definitely agree with that. And I guess it's good for a potential developer to look at the pitfalls, to see what "doing it wrong" looks like. But I've not really seen any problems inherent to the concept, with the possible exception of the difficulty in... hmmm... writing good "yes, but" content for such systems. The variable results based on DC hit for Gather Information checks in some modules are an example of trying and partially succeeding - partially succeeding because a) everyone who hits the same DC gets the same results; b) there's no clear in-character communication of "can I retry? should I retry?"; c) there's much more clear metagame ability to know to retry these rolls.

So, it's a baby-step in the right direction, but still pretty bad compared to what it should look like. And limited to very few rolls.

4. "Me too" checks, where everyone acts like one success is enough for the whole party (a variant of #3). In those cases where there's no reason why they wouldn't talk, either do a group check (50+% have to succeed to succeed) or just give them the information based on the best person's outcome.

Um... what answer other than the bolded bit are you suggesting some people would do? I'm not seeing what people would do wrong here.

So, sure Dumdum the Barbarian might have a chance to sneak past the drunken patrons of a bar to exit out the backdoor unseen (DC: 5), but only Sneaky McSneaksALot will be able to sneak past the watchful eyes of the half-Shadow Dragon Beholder to get to the Ultimate Treasure (DC: 75).

Or, Dumdum the Barbarian can climb over a decaying wooden fence (DC: 5), but only Climby McClimbsALot can climb up the sheer glass cliff covered in slippery oil (DC: 100).

So, yeah, *everyone* can sneak, but only one person can sneak when it's really important. Everyone can climb, but only one person can climb when it's important.

So, first, I don't like the correlation of "difficult" with "important". I very much think that we should divorce those two, and admit that we're discussing when rolls are difficult, not necessarily when they are important.

That said, when rolls are important, that's when you want your skilled guy making the rolls. Sure, let Dumdum the Barbarian attempt to draw the patient's blood, but when they go in for surgery, leave not just the heart transplant, but also the selection of which blood type blood to administer to Dr. McHealsALot.

Yeah, I don't like this. Because it means that only one person can participate at a time. It'd be like saying "everyone can fight, but in an important fight only the fighter can hit the enemy."

Everyone should be able to participate in everything.

Sorry, but no. Or, rather, not everyone should participate in everything in the same way. If your relative goes in for brain surgery, are you going to be making Medicine rolls to assist the surgery? I certainly hope not (unless you secretly have qualifications I'm not aware of). If somebody is building a rocket for NASA, I could write the code, but should I be working on the physical product, making Craft:Rocket rolls? Absolutely not!

IME, letting morons touch the important parts of the code just slows me down. And letting me touch, well, most anything that is someone else's high-skill specialty other than programming, probably isn't going to move their project forward in a productive direction, either. Not that I haven't helped plenty of projects without having the core skills, simply by dent of taking the stress off other low-skill tasks.

I guess the question is, how much to you want your game to be Simulationist, mapping out value of rolls and Gant charts, vs how much you want your game to be Gamist, and allow efforts of various participants to matter in a more simplified, abstract way.

[King of Nowhere]

what's the alternative to dc based systems? I can't even imagine on that wouldn't have the same issues.
I mean, a dc may become a dm-may-I as he sets the difficulty check, but if you do not have to roll some dice and add some modifier depending on your skill, then it really is a dm-may-I. And if you go for a full narrative approach, well, you can do that with a dc system too. many times I had envisioned a task with a high dc only for the players to find some roundabout easier way to go at it that didn't require any rolling.

The problems I *have* encountered, and have to periodically guard myself against as a GM have actually been:

a) forgetting that requiring a series of multiple successes dramatically worsens the odds for any task, and

b) Instinctively scaling up challenges so there is still some chance of failure, without any in-universe logic. I have to guard myself against making every lock in town suddenly masterwork difficulty after the rogue gets Reliable Talent.

c) Relatedly, accidentally gating all non-specialist characters out of participating by setting the DCs to be a reasonable challenge for the party expert (especially in games with constant scaling/no bounded accuracy). This is fine for tasks like developing a cure for the plague virus, but terrible for many social and exploration tasks where other PCs should reasonably be able to participate.

Those are not much problems as mistakes made by a dm. just like railroading too much, or inserting a powerful dmpc to overshadow the players, those are dark temptations that would push a dm to abuse his power - made all the more insidious by the fact that those things are ok when done sensibly and in small amounts.
Basically, they are not a problem with the system, they are mistakes that can be made.
As for a), it can make sense in context. staying hidden for a long time increases your likelyhood of getting spotted, so it's sensible to ask for periodic checking. and the main point of using multiple guards is the increased chance that one of them will get lucky and watch the right place at the right time.
As for b), a good question to ask is "how difficult should this be for those characters?" at high level, it will generally mean there's no need to roll anything.
as for c), regarding specifically the social tasks, one has to remember that the npcs are not out there to be obstacles. if the players offer a good deal, there's no need to roll anything.

[NichG]

what's the alternative to dc based systems? I can't even imagine on that wouldn't have the same issues.
I mean, a dc may become a dm-may-I as he sets the difficulty check, but if you do not have to roll some dice and add some modifier depending on your skill, then it really is a dm-may-I. And if you go for a full narrative approach, well, you can do that with a dc system too. many times I had envisioned a task with a high dc only for the players to find some roundabout easier way to go at it that didn't require any rolling.

The next time I make a fantasy heartbreaker system I'll probably do something like this:

- Skills take base values 0 to 7. Certain activities have a minimal skill rank to even attempt them. There are no direct opposed rolls.
- Total dice pool combines an Attribute with Skill
- When you want to roll to activate a use of your skill, pay 3 Stamina and roll your total pool of d10s looking for 8+
- You can spend a success to succeed at a thing within your capabilities.
- If your action is opposed you must spend an extra success per source of opposition.
- If you are rushing, otherwise interfered with, you must spend an extra success.
- You can spend a success to avoid a side effect of your course of action, increase a numerical consequence by 1 increment (20% of baseline ability), reduce the Stamina (or other) cost of the action by 1, or reduce the time needed by half (multiplicative)
- Characters can buy one extra success for 3 Stamina per 2 points of base skill rank they have.

[Hrugner]

A good, non arbitrary replacement skill system could be a static value with no dice and each point below the difficulty requires some additional assistance or equipment to achieve. So you'd have the inexperienced climber require a guide and full climbing kit plus magical assistance taking about an hour to setup and execute the climb, and the climbing specialist skates up the wall in 2 minutes. Social skills would require you to work the room to build support, and a letter of introduction, before tackling the hard target, picking a lock could require absolute silence, enough tools to disassemble the door completely, and make a bunch of noise, that sort of thing. Then skill growth consists of buying off tools or assistants, and DC is represented by a tool list rather than a number. It would also give you a good chance to add utility to spells and items by saying what tools they replace.

Doing it this way would also make skill solutions more reliable, putting them on more even footing with destructive solutions.

[Zuras]

Those are not much problems as mistakes made by a dm. just like railroading too much, or inserting a powerful dmpc to overshadow the players, those are dark temptations that would push a dm to abuse his power - made all the more insidious by the fact that those things are ok when done sensibly and in small amounts.
Basically, they are not a problem with the system, they are mistakes that can be made.
As for a), it can make sense in context. staying hidden for a long time increases your likelyhood of getting spotted, so it's sensible to ask for periodic checking. and the main point of using multiple guards is the increased chance that one of them will get lucky and watch the right place at the right time.
As for b), a good question to ask is "how difficult should this be for those characters?" at high level, it will generally mean there's no need to roll anything.
as for c), regarding specifically the social tasks, one has to remember that the npcs are not out there to be obstacles. if the players offer a good deal, there's no need to roll anything.

They’re problems with the system in the sense that some systems make it harder to make mistakes. Compare D&D 5e to something like Neoclassical Geek Revival, which forces the automatic success issue quite a bit by effectively making taking 10 the default resolution for all checks and providing options for both bell curve and flat roll resolution. Many systems also have concrete systems for repeated checks, like tracking effort in Dungeon World or Suspicion in NGR.

D&D 5e’s skill system is perfectly usable, it would just be nice to have some additional features, kind of like the rear-view backing camera on a car. You don’t have to have it but it’s quite handy.

[SimonMoon6]

Everyone should be able to participate in everything.

How does the fighter participate in casting spells?

Everyone can participate in the adventure, but everyone should be contributing different things. The fighter doesn't have to have a magical ability to heal his friends; the cleric can do that. Likewise, the fighter doesn't have to be able to climb a wall or open a chest; the rogue can do that. The fighter doesn't have to have a way to teleport the party into the dungeon; the wizard can do that.

Everyone should have different things they're good at or else you're back to 4th edition where everyone can do the same exact things but given different names.

And nobody wants that.

[JNAProductions]

How does the fighter participate in casting spells?

Everyone can participate in the adventure, but everyone should be contributing different things. The fighter doesn't have to have a magical ability to heal his friends; the cleric can do that. Likewise, the fighter doesn't have to be able to climb a wall or open a chest; the rogue can do that. The fighter doesn't have to have a way to teleport the party into the dungeon; the wizard can do that.

Everyone should have different things they're good at or else you're back to 4th edition where everyone can do the same exact things but given different names.

And nobody wants that.

Why is “casting spells” an activity in and of itself, rather than a method via which some PCs do their thing?

[Telok]

The biggest issue with the systems is generally they're either under explained, without sufficient examples and guidelines, or they're... I call it 'monotone', where they have no flexibility built in but the same dice setup is used for absolutely everything. Or both, they cam be both at the same time too.

You can adapt of course. If you're good you can use pliers as a hammer, wrench, and screwdriver. But it's not good at it and you have to work a lot harder. Likewise a dicing system designed for a combat engine which teaches people to roll straight success/fail for every discrete action turns out to be bad for a "schmoozing at a party" challenge. Like you can use the pliers as a screwdriver it'll mostly function, but it more effort and work and clunky mistakes.

Personally I like "30 over 13" type stuff if I'm forced to use a single die like the d20 games. You need 30 points over target number 13 to get to the "solution". The invested character that can throw a 45 result is flat out better than the no-skill rolling d20+2. Although really, the only question you ever ask the die is if a character can accomplish something within the time allowed.

[Pex]

A lack of guidelines on the difficulty of a task leads to inconsistent play. It's not enough to say give DC X if it's Easy DC Y if it's Hard because one DM can say a task is Easy DC X while another says it's Hard DC Y. It is the game designers' job to provide examples of difficulty. Of course they can't account for everything possible everywhere. They aren't supposed to. Examples are enough for a comparison. More robust systems provide aid for adjusting for circumstances whether to add or subtract from the DC, provide a + or - to a roll, or in 5E's case Advantage/Disadvantage. This is a matter of implementation, not the concept itself of having defined examples.

Another issue I have is changing the difficulty based on who is doing it. A task should have its own difficulty based on itself. One character having a better chance to succeed should be based on character build choice. If a wall is DC X to climb then it is DC X to climb for everyone everywhere. It is DC X for Alex. It does not become DC Y (Y > X) because suddenly Bob is there and is higher level. If you need a DC Y for whatever reason then it should be a different type of wall worthy of being DC Y, and by math Alex will have a harder time climbing it as is proper. Training in a skill (skill points, proficiency, whatever skill build system used) should make things easier for you by giving you a higher modifier to your roll than someone who is not so trained, i.e. did not dedicate the skill build method into it. The task is still Dc X. The one trained will already have an easier time of it. You don't need to change the DC to Y just because a lesser trained character tries to do it. It is a matter of rules implementation to say only characters trained can do a certain skill use. If Alex is not trained while Bob is, then only Bob may roll for it. The rules should explicitly say this if that's what is desired. Without that explicit mention Alex gets to roll at the same DC as Bob. Bob just has a better chance of succeeding because of his higher modifier. If the DM wants to house rule that mention into a game that doesn't have it, he should say so in Session 0 before character creation.

[King of Nowhere]

A good, non arbitrary replacement skill system could be a static value with no dice and each point below the difficulty requires some additional assistance or equipment to achieve. So you'd have the inexperienced climber require a guide and full climbing kit plus magical assistance taking about an hour to setup and execute the climb, and the climbing specialist skates up the wall in 2 minutes. Social skills would require you to work the room to build support, and a letter of introduction, before tackling the hard target, picking a lock could require absolute silence, enough tools to disassemble the door completely, and make a bunch of noise, that sort of thing. Then skill growth consists of buying off tools or assistants, and DC is represented by a tool list rather than a number. It would also give you a good chance to add utility to spells and items by saying what tools they replace.

Doing it this way would also make skill solutions more reliable, putting them on more even footing with destructive solutions.

this is already covered by a dc-based system. become good enough at a certain skill, and you have automatic success. apply measures to make the task easier (like the climbing equipment example), and you can make it an automatic success for anyone. if the task is hard enough, it can't be done.
a dc-based system includes a chance to succeed or fail when a task is just hard enough that you could maybe succeed.
it also includes a way to handle opposite use of skills (like stealth against perception) which your proposal would lack.

The next time I make a fantasy heartbreaker system I'll probably do something like this:

- Skills take base values 0 to 7. Certain activities have a minimal skill rank to even attempt them. There are no direct opposed rolls.
- Total dice pool combines an Attribute with Skill
- When you want to roll to activate a use of your skill, pay 3 Stamina and roll your total pool of d10s looking for 8+
- You can spend a success to succeed at a thing within your capabilities.
- If your action is opposed you must spend an extra success per source of opposition.
- If you are rushing, otherwise interfered with, you must spend an extra success.
- You can spend a success to avoid a side effect of your course of action, increase a numerical consequence by 1 increment (20% of baseline ability), reduce the Stamina (or other) cost of the action by 1, or reduce the time needed by half (multiplicative)
- Characters can buy one extra success for 3 Stamina per 2 points of base skill rank they have.

i have no idea what exactly stamina would be in your system, but while it makes sense for a jump or tumble, why would a knowledge or spot check require stamina? and by "spending a success to succeed at a thing" you basically remove the different difficultes: sure, you need 5 ranks to even try this difficult task, but once you do have those 5 ranks you only need one success, so it's almost trivial. also, there's no good mechanic for an opposed check.

But finally, let's just accept that your skill level gives you a number of dice to roll, and you need a certain number of success on each one; that's a more convoluted system, but it's virtually identical to setting a dc. being more skilled increases your odds by giving you more dice to roll instead of giving a bonus to your single roll.
i would say it does increase variety - if you have a +10 to a skill, you are never going to roll below 11, while if you have 10 skill dice you can still roll 0 success - which increases the odds of the super expert failing spectacularly at something very easy. can be good for laughs, but not desirable.

I'm not picking on you two specifically, it's just that the DC system - used correctly - is simple and effective, and I've never seen a system or a proposal that would not have worse problems.

tl;dr the DC based skill system is the worst possible skill system, with the exception of all others.

D&D 5e’s skill system is perfectly usable, it would just be nice to have some additional features, kind of like the rear-view backing camera on a car. You don’t have to have it but it’s quite handy.

d&d 5e is a particularly poor application of the skill system. at least based on my experience with baldurs gate 3, which is my main source of knowledge on how 5e works.
you don't gain skill ranks. you can never be a specialist. I can be a senior professor, but ultimately I have nothing more than skill proficiency, and a college freshman is going to have my same modifier. which in turn makes the handling of how difficult a task should be a complete mess-up.

[JNAProductions]

Something to note is that 5E and other D&D systems’ skill systems are for ADVENTURING.
They are NOT meant to simulate teaching a college course, or activities like that.

[PhoenixPhyre]

Something to note is that 5E and other D&D systems’ skill systems are for ADVENTURING.
They are NOT meant to simulate teaching a college course, or activities like that.

Yeah. The model is that adventurers are generalists with mild area of focus. And all the ability check proficiency areas are adventure focused. NPCs, in the course of their regular lives (and PCs as well, outside of adventuring) don't use ability checks. Commoners don't need farming proficiency to farm, they know about their daily tasks and environment without any checks. A cleric knows his faith without needing proficiency or even a roll. Etc.