D&D (2024) Should Bounded Accuracy apply to skill checks? Thoughts on an old Alexandrian article

[DEFCON 1]

Let's also remember the real point of all of this-- these checks are not the game unto and of itself. Their entire point is to vary the speed in which players acquire narrative advancement in the story they are participating in.

The entire game is a story that is happening at the table. That's why every single mechanic in the game has a narrative description layered on top of it-- so the players can imagine an interesting meaning for what these numbers represent that we are adding together all the time. A player rolling a twenty-sided die and then adding two numbers to it to get another number-- which is then compared to another number the DM has chosen-- is then given meaning by treating those numbers as something imaginable in the fiction. A d20 roll plus a number described as "the character's strength" plus a number described as "their skill in athletic actions" is acquired to beat a number the DM has decided in the fiction is "climbing over that wall to get to the other side". If the player can accomplish that... then the story can continue on from there. The players are now on "the other side of the wall" and they can make their next set of narrative decisions. And if the player can't accomplish that "getting their number higher than the DM's number"... then the story of what is "on the other side of the wall" has to stop for the time being and the player(s) have to imagine and come up with some other way to get past this "barrier" to see the story continue (if they even decide whether they want to do so.)

So it doesn't really matter what are the numeric totals the players can reach on random die rolls plus other numbers... nor does it matter what the numbers are that the DM just cooks up to decide whether or not the players can possibly succeed... because in both cases these are just dictating whether or not the players can "advance" in the story. These numbers and rolls are merely just narrative pacing issues.

At no point do any of us NEED to roll dice and add numbers to them to advance the narrative. We could all, if we really wanted to, just push the story forward whenever we wanted by just dictating that everything the players wish to do is accomplished. We don't NEED players to make "Perception checks" for example... the DM could just tell the players the pertinent information that their character "sees" and thus allows them to advance the story forward with the information they needed to make their next series of informed choices. But if that makes the story advance too fast... then the DM can throw in a "Perception check" to slow down what info they are going to pass on to the players, thus rendering their next decisions having to be made without "perfect information".

The game isn't actually "the acquisition and comparing of numbers". "The numbers" are merely a TOOL-- a tool that tells us what is happening in our imagination at any point in time.

 

[Clint_L]

Mostly I want to know what "whole bunch of weird, nit-picky rules about which specific types of attacks and which specific types of character can get critical hits in combat".

Overall the article seems typical of Justin's stuff, not without its useful insights, but also one-sided to the point of dishonesty when it talks about anything that might not unequivocally favour the view he's trying to push.

In general, the internet does not reward subtlety and nuance.

 

[Pedantic]

Let's also remember the real point of all of this-- these checks are not the game unto and of itself. Their entire point is to vary the speed in which players acquire narrative advancement in the story they are participating in.

The entire game is a story that is happening at the table. That's why every single mechanic in the game has a narrative description layered on top of it-- so the players can imagine an interesting meaning for what these numbers represent that we are adding together all the time. A player rolling a twenty-sided die and then adding two numbers to it to get another number-- which is then compared to another number the DM has chosen-- is then given meaning by treating those numbers as something imaginable in the fiction. A d20 roll plus a number described as "the character's strength" plus a number described as "their skill in athletic actions" is acquired to beat a number the DM has decided in the fiction is "climbing over that wall to get to the other side". If the player can accomplish that... then the story can continue on from there. The players are now on "the other side of the wall" and they can make their next set of narrative decisions. And if the player can't accomplish that "getting their number higher than the DM's number"... then the story of what is "on the other side of the wall" has to stop for the time being and the player(s) have to imagine and come up with some other way to get past this "barrier" to see the story continue (if they even decide whether they want to do so.)

So it doesn't really matter what are the numeric totals the players can reach on random die rolls plus other numbers... nor does it matter what the numbers are that the DM just cooks up to decide whether or not the players can possibly succeed... because in both cases these are just dictating whether or not the players can "advance" in the story. These numbers and rolls are merely just narrative pacing issues.

That's treading dangerously close to the 4e skill challenge synthesis. Ideally, those numbers aren't just a pacing mechanism, but a gameplay mechanism; they provide a set of abilities that players can leverage against the setting to get the outcomes they want. I want to be inside that castle, I am good at picking locks, let's find a lock that separates me from the inside of that castle.

To make player decisions interesting as a gameplay concern, and not simply as narrative color, it's important that their decisions matter with regard to their ultimate success. A player should be able to make a bad choice (or better, series of choices) that frustrates their goals, so that there is a reason to make better choices; ideally players are exploiting their advantages against the situation to the best of their ability.

Of course, I don't really think that's possible without clear DC rules in place for players to leverage, and we're short on even skill challenge levels of pacing guidelines.

 

[Pedantic]

If the point of the Rogue is to just have a character who automatically succeeds at certain skill checks, it occurs to me that they are overtuned in this regard. By the final Tier of play, Expertise could allow for a +17(!) bonus to a check (or higher, in theory- you could get a Dexterity above 20, or the ioun stone that increases your proficiency bonus by 1, which would become another +2), which, combined with guidance, help action, bardic inspiration, advantage, and so forth, would allow the Rogue to make checks beyond impossible.

You could be succeeding at DC 30 checks long before this point, of course. One of the frequent complaints about 5e is that the exploration and social pillars of play are basically vestigial, and a lot of these sorts of challenges are based on checks (in my experience). Can you open the door, climb the wall, avoid the trap, decipher arcane runes of power, survive on an outer plane, or convince the king to lend you his army?

The ability to just automatically succeed at these tasks potentially does more to harm the non-combat portion of play than things like goodberry.

It's funny to see skills as more damaging than "spells that let you skip the problem" a criticism I also felt was entirely off-base. You're right that at least has a resource management component, but like combat the real underlying resource should still be action economy. There's space for a complete deterministic skill game that is still a meaningful game, if the actions are specific enough.

Is the goal of the skill game really to ensure that players only have a 65%-80% chance to succeed with any action? That's the part of the equation that actually doesn't have a skill or decision making component. Dice are the least interesting part of the game.

 

[DEFCON 1]

That's treading dangerously close to the 4e skill challenge synthesis. Ideally, those numbers aren't just a pacing mechanism, but a gameplay mechanism; they provide a set of abilities that players can leverage against the setting to get the outcomes they want. I want to be inside that castle, I am good at picking locks, let's find a lock that separates me from the inside of that castle.

The gameplay mechanism is in the game merely to BE a pacing mechanism. The gameplay mechanism isn't the action of the game unto itself. If it was, we wouldn't need to go through all the trouble and effort of creating descriptions and names for all of these numbers, the point of the game would simply be "I roll a die, I add some numbers to it... if my number is high than your number, then I win." That's like playing any dice game. We don't bother thinking up what the dice "represent" when playing craps or liar's dice... they are just dice and the point of those games are to just roll them and try and win.

But we don't do that in D&D, we give those dice and numbers a meaning. Something they represent. And the times of using those dice and numbers aren't based on the gameplay of "Okay, it's now my turn to roll" and just going back and forth between the players and DM in some sort of schedule... the gameplay mechanism is in service to something else... that being everyone's imagination of what their "characters" are "doing" within the "story".

We don't roll dice just for the sake of rolling dice because that's what "the game is"... we only roll dice when our collective imaginations have decided that our "story" needs a roll of the dice to generate a new story beat.

 

[UngeheuerLich]

I think his reasoning is fundamently flawed. When he says: at level 1, you are struggling fighting against dire wolfs, casting dancing lights and opening the tavern door lock, but at level 20 you are soloing smaug, casting metor swarms but still having troubles opening the tavern door lock, he mixes up several things.

The wizard and the rogue still have trouble, standing solo against smaug. The fighter and the rogue still only cast dancing light. And the fighter and the wizard have trouble opening the door.

The rogue however with reliable talent and expertise will open every lock with ease. With the exact same ability that he rants against in the paragraph above.

Expertise and reliable talent are what makes the rogue stand out. They never have trouble opening complex locks easily after level 7. As the fighter has no problems standing tall against big threatening enemies.

 

[Parmandur]

I was reading an article by Justin Alexander roasting the 5e skill system, and arguing (among other things) that Expertise is bad because it breaks Bounded Accuracy, and Reliable Talent makes it worse. And with this, I disagree.

I think that Bounded Accuracy is excellent for combat's standard rolls: attack vs AC, and saving throw vs DC. That's when you need numbers that challenge the whole party: some characters may have a better chance than others, sure, but the d20 roll doesn't become irrelevant because this one is guaranteed to succeed and that one is doomed to fail.

That dude doesn't understand what Bounded Accuracy entails, or the range of numbers involved. Bounded accuracy is about numbers uo to 30: Expertise and Reliable Talent do not make DC 30 rolls irrelevant, but rather makes them achievable but in doubt.

Working as designed.

 

[tetrasodium]

I strongly believe characters should be allowed to be highly competent in their fields of expertise, at least once they reach tier 2. I want players to look at a challenge that's in their character's wheelhouse and go "Don't worry, I got this." If you're a rogue with Thievery expertise and proficiency in Thief's Tools, there shouldn't be many locks that resist you. If you're a bard with expertise in Performance and proficiency in a number of musical instruments, you should be able to put on a really good show. A cleric with expertise in Religion should know obscure points of theology. In games where violence is a less common occurrence than in D&D, I could also see a dedicated fighter-type being able to take out pretty much anyone, but in D&D fighting is too much a part of the core loop for that to be reasonable.

"So how do you challenge PCs?" Well, for one thing you should realize that if you have an expert thief, locks shouldn't pose a challenge to them. Locks are a challenge to other people, but that player has made the choice to be awesome at opening locks, and they should be allowed to do that. Instead, you challenge them in other ways. Perhaps the lock isn't easily accessible, and the rogue has to either talk their way into the place where it is, or physically infiltrate the location. Or maybe the door is not just locked but also guarded physically, so the rogue or some other party member has to lure the guards away. Or maybe there are patrols, so you have a limited time to do whatever you were planning to do in the locked room.

This is Completely mistaken in how it tries to paper over a badly designed skill system with a secondary clause in that same skill system as if they are the same. It's ignoring the fact that 5e already addresses that sort of nonrolled skillcheck by telling GMs not to call for a roll when the outcome is not in doubt. Topping that off by ensuring the outcome can't be in doubt when a roll is actually called undercuts the purpose of even having a skill system that also sometimes relies on rolls & DCs

A skill system designed around conflicting logic like "Don't ask for a roll if the results are not in doubt and the results can't be in doubt when you do ask for one" is automatically a failure in how it presents a catch22 damned if you do damned if you don't design embodied by wargames where the only winning move is not to play [with either option]. GM's need far more function in a skill system than that.

 

[nevin]

The issue I had with this is that in order to make skill checks interesting, you need to make them challenging to some degree. This isn't "ho ho ho, you failed the check, the adventure is over" thinking- failure can be interesting in it's own right, if characters are allowed to try again after dealing with a setback, or "failing forward".

It's more that, let's say we have two characters. One is an archer Fighter who favors Dexterity and has a Street Urchin background. The other is a Rogue who took expertise in Thieves' Tools.

Right off the bat, all things being equal, the Rogue has a +2 bonus over the Fighter. At this point, it's no big deal, because you would expect a Rogue to be better at picking locks than a Fighter.

By level 11, however, the Rogue now has a +4 bonus over the Fighter, and (assuming Dex 20), simply cannot roll worse than a 23. If the DM selects skill check DC's that the Fighter can reliably hit (say, DC 19-24), then the Rogue almost always succeeds at these tests, making the Fighter redundant save for those times the Rogue isn't able to make the check for some reason, and almost taking failure off the table, making these sorts of skill checks uninteresting and maybe even pointless!

If the DM instead raises the bar so that the Rogue occasionally fails, now we're talking about DC's of 28-30, the maximum DC the system allows for.

The Fighter might be able to hit these on a natural 20, or, in the case of DC 30, finds it impossible to do so. Which means that they might as well not even have proficiency in Thieves' Tools. Further, in those cases where the Rogue cannot make the check, they are almost certainly going to fail, meaning that the skill checks are again almost pointless and not interesting at all.

Yes, there are other ways you can go about this- forcing doors open, finding keys, hidden entrances, use of magic, but if these are viable alternatives, again, the Fighter's proficiency basically doesn't matter!

Expertise becomes back breaking when it means that either characters with it succeed all the time, or you need Expertise to be considered good at something.

Considering that 5e is a system that lets any character attempt most rolls (Thieves' Tools being the only standout), this is really bothersome.

Someone might say "well, the Fighter can Help the Rogue" to gain advantage, but that just exacerbates the situation- you either have it so one character always succeeds or that no one else but one character has a chance to.

Not exactly a great skill system, IMO.

so a bad skill system means the guy that didn't take the skill fails. Seems like a normal skill system. To me a skill system that anyone can succeed on isn't really a skill system and you might as will do away with the skills system and just DM hand waive it to move things on.

 

[Clint_L]

The issue I had with this is that in order to make skill checks interesting, you need to make them challenging to some degree. This isn't "ho ho ho, you failed the check, the adventure is over" thinking- failure can be interesting in it's own right, if characters are allowed to try again after dealing with a setback, or "failing forward".

It's more that, let's say we have two characters. One is an archer Fighter who favors Dexterity and has a Street Urchin background. The other is a Rogue who took expertise in Thieves' Tools.

Right off the bat, all things being equal, the Rogue has a +2 bonus over the Fighter. At this point, it's no big deal, because you would expect a Rogue to be better at picking locks than a Fighter.

By level 11, however, the Rogue now has a +4 bonus over the Fighter, and (assuming Dex 20), simply cannot roll worse than a 23. If the DM selects skill check DC's that the Fighter can reliably hit (say, DC 19-24), then the Rogue almost always succeeds at these tests, making the Fighter redundant save for those times the Rogue isn't able to make the check for some reason, and almost taking failure off the table, making these sorts of skill checks uninteresting and maybe even pointless!

If the DM instead raises the bar so that the Rogue occasionally fails, now we're talking about DC's of 28-30, the maximum DC the system allows for.

The Fighter might be able to hit these on a natural 20, or, in the case of DC 30, finds it impossible to do so. Which means that they might as well not even have proficiency in Thieves' Tools. Further, in those cases where the Rogue cannot make the check, they are almost certainly going to fail, meaning that the skill checks are again almost pointless and not interesting at all.

Yes, there are other ways you can go about this- forcing doors open, finding keys, hidden entrances, use of magic, but if these are viable alternatives, again, the Fighter's proficiency basically doesn't matter!

Expertise becomes back breaking when it means that either characters with it succeed all the time, or you need Expertise to be considered good at something.

Considering that 5e is a system that lets any character attempt most rolls (Thieves' Tools being the only standout), this is really bothersome.

Someone might say "well, the Fighter can Help the Rogue" to gain advantage, but that just exacerbates the situation- you either have it so one character always succeeds or that no one else but one character has a chance to.

Not exactly a great skill system, IMO.

So...this seems like good design, working exactly as intended. At high levels, characters are facing elite challenges, and the hardest of those should only be possible for the character who is truly specialized at the task. Even if the fighter takes skill in picking locks, they will never be as good as a high level rogue, though taking special measures, such as multi-classing or spending a feat on skill expert, can get them close. Just as the rogue will never be as good at bashing down a door, as you mention, or going toe-to-toe with a giant.

And your criticism that "no one else but one character has a chance" is patently untrue. On an extreme case - let's say the super secure vault of a high level NPC - the DM intends for it to only be openable by characters optimizing their toolset. It's not supposed to be easy. At the same time, my current party has three characters who are all quite good at opening locks, in different ways. Our artificer would give any rogue a run for their money at breaking into places through subtle means, while our barbarian is very good at just going through them.

That's basically every team story ever: the reason you have a team is because different folks are good at different things.

The point of skill checks is to help tell a good story, which means that there are challenges to give the story stakes. If the check is so easy that there is no reasonable chance of failure, why even bother with a skill check? Just let the story move ahead to the good stuff.