Ever hear someone say, “the d20 is too damned swingy?” I sure as hell have. It’s become a popular thing to piss and moan about it in online gaming circles. And so, weeks ago, I took a jab at the whiners in my True Game Mastery lesson about dice. I said the complaint was a load of horseshit.
And now I’m gonna explain why.
First, disclaimer: this is my once-monthly Random Bullshit column. In those, I take a break from providing good, well-reasoned advice and structured arguments to instead rant about whatever I want to rant about. I don’t promise anything useful, edifying, elevating, or even coherent. I’m just letting y’all in my head.
But the Swingy d20 Argument is a good thing to rant about. Sure, the complaint is crap — demonstrably and provably — but there’s an issue under it that’s worth talking about. A minor one; but an important one and a fixable one. A mostly fixable one.
Moreover, the Swingy d20 Argument also illustrates how unimportant dice and math and statistics and mechanics actually are vis a vis a game’s fluffier, more subjective gamefeel aspects. And how trying to fix a feeling problem with math fixes nothing.
And don’t worry if you have no idea what the Swingy d20 Argument is. I will explain it in my usual exhaustively over-detailed oh-God-make-it-stop way. And I’ll explain it… right now.
Part I: Dice Swing is Bullshit
It’s common for online gamers to bitch that the d20 is “too swingy” and much digital ink has been wastefully spilled on the subject. Especially lately. And when I say lately, I mean approximately eight years, nine months, and two weeks. Give or take.
Remember that oddly specific timeframe. It’s important later.
Look, this complaint ain’t a new one. Prominent game designers have been discussing the gaming bugbear that is the Swingy d20 for approximately 45 years, seven months… never mind. The timing ain’t important. Just know this argument’s been around for a while. But it’s always been a sort of academic, sideline thing. But in the last almost-decade, the complaint’s gone mainstream. Really mainstream.
What’s the complaint? Well, it’s a math thing. Specifically a probability and statistics thing which is the very best kind of math thing. But this is also the very worst kind of complaint to discuss with gaming nerds.
But let me lay it out for you…
Swinginess Defined
The problem is everyone’s favorite icosahedral plastic math rock: the twenty-sided die. When you test an in-game action by rolling a d20, every number on the die is equally likely to show up. There’s a five percent chance you’ll roll a 3, a five percent chance you’ll get a 10, a five percent chance you’ll get a 17, and so on. So, if you roll a hundred checks in a game session, you’d expect to roll five 1s, five 2s, five 3s, five 4s, and so on down the line to five 20s.
That ain’t really how probability works, but it’s close enough and easy to grok, so that’s how I’m explaining.
Because every result’s equally as likely as every other on a d20, gamers incorrectly call this a linear probability distribution. That ain’t correct either, but that’s the term gamers use so I’ll keep using it rather than point out that it would be more correct to call it a uniform probability distribution and to throw the word discrete in there for good measure. And I’m only including this paragraph so all the armchair mathematicians know I could get this shit more right than they can and back off any admitted inaccuracies I include for the sake of conceptual understanding.
So, the d20 yields a linear probability curve. If you graphed the odds of each outcome, it’d look like this…
But to understand how that makes the d20 swingy, we’ve got to talk about another way to roll dem bones.
Rolling Dice Normally
Some tabletop roleplaying games heretically eschew the ideal perfection of the greatest of the platonic solids and instead ask you to roll fistfuls of dice to determine your in-game outcomes. And right now, I’m speaking specifically of games that ask you to roll multiple dice, add the results together, and compare the total to some target number. That includes games like GURPS, Fate, Fantasy Age, and all those Powered by the Apocalypse games but it does not include games like Blades in the Dark, and Shadowrun because you don’t sum your dice rolls in those games. The probability rules are totally different.
In GURPS and Fantasy Age, for ee gee, you roll 3d6 and sum them whenever you try to do a thing. And the probability laws change completely when you do that. The lowest and highest results — 3 and 18 respectively — are each less than one-half percent likely to come up. Meanwhile, the middlemost results — 10 and 11 — are each more than 12 percent likely.
The reason for the weirdness is down to the fact that there’s only one combination of three numbers between 1 and 6 that’ll give you an 18. You need to roll all 6s. But to get a 10, you can roll 1, 3, and 6; or you can roll 2, 4, and 4; or you can roll 1, 4, and 5; or…
Look, I ain’t running through the 27 different combinations of d6 results that’ll get you a 10. Trust me that there are 27 ways. Okay?
The point is, if you rolled a hundred checks in a session of GURPS, you’d expect a full quarter of those rolls to yield a 10 or 11 and you might not see a single 3 or 18 at all.
Again, that ain’t technically correct but it’s conceptually correct so it’s good enough for us.
Graph the probabilities and you get a thing that looks like this…
And gamers call that a bell curve or normal distribution whereas actual math wonks wouldn’t even graph the results like that because, again, dice generate discrete results… blah blah blah math is hard.
The Swinginess is in the Shadows
The Swingy d20 Argument states that, because you’re just as likely to roll extreme results like 1s and 20s as you are to roll middlesome results like 10s and 11s whenever you roll a d20, the d20 is just too damned random to use for anything as vitally important as the outcomes in pretend elf games. Extreme results should be rare. Middling results should be common.
Most gamers can’t explain this shit coherently, though. Instead, they draw the graphs and then say, “See?! Don’t you see?!” And that’s because most gamers are just repeating crap they read online.
The thing is, randomness isn’t inherently bad. Or variability. Or swinginess. Let’s just call it swinginess. That’s what everyone calls it.
So what, really, is the problem?
Well, the folks who do write coherently on the subject — including DM David’s excellent analysis of the topic — outline the issue more clearly. And generally, the issue comes down to the likelihood of experts failing and the likelihood of incompetents succeeding.
Imagine, for instance, a very perceptive elf rogue with a very high Spot Hidden Things modifier. Using a d20, that character’s player is equally likely to roll a 3 as a 15 and thus they might get robbed of their success by a crappy die roll. If they were rolling 3d6, on the other hand, the odds of a 4 or 5 robbing them of their rightful success is far, far lower.
Likewise, when the obliviously dunderheaded barbarian looks over the elf’s shoulder with his crappy-ass Spot Hidden Things, he’s equally likely to roll a 17 as a 7 and he might thus end up with a totally underserved success by amazing luck. And if he succeeds where the elf failed, a wholly immersion-breaking crime against pretend-elf gamingdom occurs.
Yep, this is all down to noddle-armed wizards smashing down doors the barbarian only managed to loosen in a different skin.
There is however a slightly different — and technically more correct — way to look at the problem. Consider again that elf. Let’s say her modifier is so high she can only fail on the lowest possible roll on the die. In D&D, she’s got a +10 and she’s rolling against a DC 12. In Fantasy Age — a 3d6 system — she’s got that same +10, but she’s rolling against a TN 14.
In D&D, the elf will fail 5% of the time. In Fantasy Age, she’ll fail only once in two hundred tries.
You can make the same argument backward for the oblivious barbarian and his odds of success if he needs the highest possible roll.
The problem is both of those arguments are totally wrong.
The Math is Right, The Model is Wrong
The shit I said above about the laws of probability? The same shit everyone else is saying? That’s all technically and mathematically correct. Except for the parts that I admitted were inaccurate for the purposes of clear conciseness and the incorrect terminology I used because it’s the same terminology everyone else is using.
But there’s a big, important fact that everyone’s ignoring. The probabilities in the systems I’ve been talking about — D&D and Pathfinder and GURPS and Fantasy Age and Fate and all the PbtA games — aren’t governed by linear probability distributions or bell curves or any of that other shit.
The correct term for all these games is binomial probability distribution.
In D&D, actions don’t have twenty outcomes, they have two. They either work or they don’t. The same is true of GURPS. The same’s true of Fantasy Age and Apocalypse World and Fudge. Yes, you can indeed Fail, Succeed, or Crit in D&D, and in PbtA games, you can Fail, Sorta Succeed, and Really Succeed, and in Fantasy Age you can Fail, Succeed, and Succeed with a Stunt, but those odd extra outcomes are just success and failure with an extra little twist.
Consider this: in D&D, a 1st level character specced for archery who fires an arrow at a basic foe is 70 percent likely to hit assuming the average numbers in the core rules. An arrow fired by a starting archer in GURPS is 75 percent likely to find its mark. And a Dungeon World archer has an 83 percent chance to at least graze his target.
If you’re rolling dice and comparing the result to a target number to generate a mostly binary result, it doesn’t matter what you roll on the die at all. It doesn’t matter in D&D whether you rolled a 7 and hit or you rolled a 17 and hit. What matters is that you hit. And in GURPS, a 5 that succeeds — GURPS is a pain-in-the-ass roll under system — a 5 that succeeds is no better than a 10 that succeeds.
I could replace the die and math with percentile rolls and you’d never know the difference. And the odds, by the way, they ain’t a function of the dice you’re using. They’re just down to where the designers set the modifiers and Target Numbers. Just by changing the Target Number tables or the Attribute Modifiers, I could bring all three systems’ odds in line.
And if I hid the percentile dice from the players and made all the rolls myself in secret, they’d never know they were playing a different game.
Let me make this abundantly clear: if the actual number on the die doesn’t matter — only whether it’s higher or lower than some target — the di(c)e you roll is almost — but not quite entirely — irrelevant.
Why almost? Because of extreme circumstances. When a character tries something that is almost impossible to do — or to fail to do — swingier dice are more likely to produce surprising results. The eagle-eyed elf who can only fail on the worst result? She’s got a five percent chance of failure in D&D but only a half-percent chance to fail in GURPS.
But the flaw in that point is that extreme circumstances are pretty frigging rare. Tabletop RPGs aren’t designed to even present them. When did you last have someone roll a check in D&D they could only fail on a 1? Or could only succeed on a 20? Because I sure as hell can’t remember ever asking for a roll like that.
This is a spherical chicken in a vacuum issue. One that screaming internet math nerds bring up in online forums but not one that happens at actual tables.
And that is why I said the Swingy d20 Argument is a load of horseshit. It’s a technically mathematically correct analysis of dice probabilities that has literally nothing to do with tabletop roleplaying games.
Part II: But There is a Problem
As I said above, the d20’s swinginess has been a thing in the game design space since the 1970s. Hell, Gygax himself addressed it in sidebars in early D&D editions and Dragon magazine columns and Steve Jackson based his early fantasy game designs around it. Their analyses though were more nuanced, more even-handed, and showed a much better grasp of the underlying statistical principles. I am not claiming to be smarter than Gygax and Jackson and Cook. Though I’m also not saying I’m not.
What’s weird though is that this complaint’s gone mainstream in the last several years.
I say a lot that people are good at knowing when they’re unhappy and bad at knowing why. When you’ve suddenly got lots of people pissing and moaning about the same thing, there’s probably something upsetting people. And when people are upset about something, they start flailing around for explanations and solutions. And the Swingy d20 is the perfect example of a solution looking for a problem. Or a diagnosis looking for a symptom.
If you ain’t a math wonk and you read a bunch of crap about how the d20 is more likely to produce extreme results and if you’ve got this deep-seated dissatisfaction with dice-rolling outcomes in your favorite tabletop roleplaying game, you’re going to take the former as a plausible explanation for the latter. And much hay was made over the swingy d20 about the time D&D 5E was in open, public playtest thanks to Monte Cook splitting off from WotC over creative differences and publishing his own game, Numanuma, and then making a big deal about the swinginess issue in his design blog.
Me? I think there is something underlying the Swingy d20 folderol. I think people are dissatisfied with something. I know it ain’t about the probabilities though. That said, I want to be absolutely clear about just how big and important I think this shit actually is.
This Shit Ain’t Important
From here on, I’m assuming some number of gamers are unhappy and it’s got something to do with rolling dice. But how many people are unhappy? And how unhappy are they? I don’t know. No one does. But I’m betting it’s somewhere between “very few” and “almost no one.”
The two most popular tabletop roleplaying games in the Western world are powered by the d20 system. And I’m not dumb enough to claim — like many of my spiteful contemporaries — that D&D is only popular because it’s popular or because people are stupid or whatever. That’s an idiotic statement to make. People don’t invest heavily in things they don’t enjoy. Which means the problem is neither widespread nor game-breaking.
The problem here is that issues like this quickly turn into self-fulfilling prophecies. When you start telling people that the d20 is swingy and makes their characters fail too much, people start to notice their failures more. And people are already much more failure-focused than success-focused. They notice their failures.
In other words, if you read some rando blogger’s dumbass screed about the swingy d20 and then have a slightly unlucky game session, you too become convinced that the d20 is ruining everything.
So, the problem — whatever it is — isn’t a common, widespread, or game-breaking problem. It’s a little niggle. A minor annoyance. It bugs some, but not most.
But what is it?
Diagnosing the Problem
How do you diagnose a condition when you’ve got a male patient saying, “I have a mysterious and indescribable pain somewhere in the lower half of my body and I suspect it’s primary ovarian insufficiency?” Well, if you’re smart, you totally ignore the patient’s diagnosis on the basis that men don’t have ovaries. That idiotic and wholly incorrect diagnosis is the equivalent of gamers blaming their woes on wholly incorrect understandings of gaming probabilities.
Then you try to isolate the problem. You try to figure out where the pain actually is and when it happens and where it isn’t and when it doesn’t happen and all that crap.
Where’s the pain? In this case, the pain is very specific to the Dungeons & Dragons community. Even the overanalytical, power-gamey math wonks — I say with affection — of the Pathfinder community don’t complain about the swingy d20 much.
Moreover, while there’s always been chatter about the swingy d20 in the D&D space, it used to be an academic side issue that was occasionally spackled over with variant 2d10 or 3d6 rules that never caught on.
The point is, the pain’s pretty specific to D&D 5E gamers.
When does the pain happen? Well, I spent a torturous few days reading blogs and Reddit threads, watching YouTube videos, and poring through comment and forum feeds and I noticed some common points of contention. People talking about it talk very specifically about failing die rolls they don’t think they should have. Often, the complaint is phrased as experts who fail too often or fail routine tasks. Occasionally, there’s some additional whining about non-experts outperforming experts.
When does the pain not happen? Well, the pain doesn’t happen a lot. If it did, everyone would be quitting D&D 5E. Or, at the very least, the pissing and moaning and ranting and whining would be a lot louder. Loud enough, by the way, that the designer’s current surveys and playtests would show it. Remember, they’re doing extensive surveys and playtests right now. And given how they do what I will laughingly call their data analysis — which is an idiotic story for another time — if everyone hated the d20, we’d know it.
What does all of that suggest? A lot actually. But it’s all to do with…
Emergent Subjective Gamefeel Bullshit
Let’s get this straight: game mechanics don’t make games good or bad. It’s the dynamical interaction between the players and the mechanics — as well as the complex interactions between the game mechanics — that lead to a game feeling good, bad, fun, unsatisfying, whatever.
It’s never just one mechanic. Which is why you can’t fix unhappiness through math alone. And why you can’t even blame unhappiness on math.
The dissatisfaction with the d20 is absolutely not coming from some law of probability. It’s coming from a bunch of different game mechanics rubbing a certain small subset of players the wrong way. And that’s for several reasons.
Nobody Likes Failing
Let’s start simple and stupid-obvious: people hate failing. Failing makes people sad. People only yell about the d20’s swinginess when they fail die rolls. And of course, they’re unhappy and prone to yell then. They just failed a die roll.
Moreover, some people cope with failure better than others. Some folks are good sports and some are whiny crybabies. And this problem is limited to a small subset of gamers. Maybe it’s the gamers who just aren’t good sports?
But I can’t dismiss this wholly as people who can’t suck up their losses. First, because you aren’t any more or less likely to miss a roll in D&D 5E than in any other tabletop roleplaying game system. If the problem really was just about crybabies crying, it wouldn’t be more common in the D&D 5E community than anywhere else. And while I could argue — convincingly, I might add — that the D&D 5E community has a higher prevalence of whining crybaby bad sports, I still don’t think that’s the whole issue.
And that’s because most of the complaints point very specifically to experts failing when they ought not to.
The Curse of Bounded Accuracy
When the brain trust at WotC designed the greatest latest iteration of D&D, they tried very hard to fix many complaints about previous editions. One such complaint about the extremely popular 3rd edition of the game was that the numbers were just too damned high. Difficulty and Armor Classes often reached into the thirties and forties even at middling levels of play and skill and attack modifiers went deep into the double digits. It was too much.
So WotC’s designers introduced the concept of Bounded Accuracy. Their goal was to keep DCs and ACs from reaching much beyond 20 and keep d20 modifiers mostly in the single digits. And while I can prove, mathematically, that this choice contributes to the d20’s perceived swinginess, I don’t have to. I just have to tell you what people really didn’t like about 3E.
The number bloat had a bunch of nasty side effects, you see. If you wanted your character to succeed at anything beyond the lowest levels of play, you had to make that character an expert. You had to tweak your skills to keep your modifiers up. Playing a jack-of-all-trades was right out. Non-experts couldn’t succeed at most tasks. Meanwhile, the expert characters’ expertise completely invalidated the randomness in the game. Most routine tasks were trivial for experts. They couldn’t fail.
Bounded Accuracy was introduced to D&D 5E to counteract the so-called expertise problem. It decreased the distinction between experts — who couldn’t fail — and non-experts — who couldn’t succeed — to allow for more balanced challenges in which everyone could participate with more reasonable chances for success and failure.
I don’t think it’s crazy to suggest that Bounded Accuracy might be contributing to the perception that the d20 is too swingy by letting experts fail and non-experts succeed considering that’s exactly what the designers said it was supposed to do.
So why hasn’t the scrollbar hit the bottom of the page? Why are there still fifteen hundred words waiting below? Isn’t that the whole answer?
No, you dumbass! Didn’t I just tell you it’s never just one thing!?
The Skill-Based Skinsuit
Dungeons & Dragons is not a skill-based system. And it doesn’t claim to be. But it wears a skill-based skinsuit. And that really screws with people’s brains.
Understand that it takes more than just a skill-based core mechanic to make a game skill-based. There’s a design philosophy and a way of building challenges. Skill-based games feature teams of experts with disparate, esoteric skills, who usually split up to handle different parts of complex tasks or else tag each other in and out to deal with particular problems.
So, the Face gets a forged invitation to the charity gala. Meanwhile, the Hacker is in a hotel room across the street waiting for the Infiltrator to install a tap in an on-site junction box to give him access to the security system. Meanwhile, the Strike Team is waiting for the security cameras to shut off so they can take out the guards…
That’s skill-based gameplay in a nutshell.
Dungeons & Dragons is a game of heroic fantasy adventure. It isn’t designed for skill-based play. Off the bat, you’re told never to split the party. Each character’s supposed to participate equally in D&D’s three major modes of gameplay: combat, exploration, and social interaction. You’ve got to get your whole party through every obstacle and encounter. Everyone’s got to cross the river or climb the cliff. There are some esoteric skills, but they mostly provide options. Lockpicking ain’t the only way to get through a locked door, it’s just the quietest. And D&D is pretty loose about demanding you have the right skill for the job. Anyone can make a navigation check or forage for food, for example.
But D&D’s got a skill system and it seems to encourage you to build a party of specialized experts with all the skills covered. Moreover, whatever the rules say, challenges are presented in terms of skill checks. The presentation is so strong, Game Masters feel they must fit every action to a skill rather than call for an ability check and most players view their skill lists as the levers with which they interact with the world.
Modern gamers also expect skill-based gameplay. Our world is one of hyperspecialization. Everyone’s an expert in something. And our world is highly complex. If you’re not an expert at a task, you’re useless at it.
Put this crap alongside the Bounded Accuracy thing and you end up with a schizophrenic mess of a game. One that’s both about and not about hyperspecialized experts. And I think it’s this weird discontinuity that makes 5E especially prone to complaints about swinginess. It’s not just that D&D 5E was purpose-built to decrease the impact of expertise; it’s also that D&D 5E presents itself as a game about specialized experts.
But I also don’t think that’s everything…
Rolling Too Many Dice
Throughout the entire history of roleplaying games, there’s always been a problem with Game Masters — and players — using dice to resolve things they shouldn’t. Every core rulebook for every tabletop roleplaying game system ever designed includes at least a paragraph begging Game Masters not to overuse the dice. But I think the culture around D&D 5E is especially prone to overusing dice.
In the online D&D space, it’s deeply unpopular to rely too heavily on the Game Master’s authority and good judgment. And D&D 5E’s design conceits and presentation downplay the importance of Game Masterial Judgement. But without that judgment, the dice become the only way to resolve questions of success. Game Masters don’t feel right declaring an action a success — or an impossibility — and players balk at such declarations without a die roll preceding them. At most tables, it’s the players’ jobs — not the Game Master’s — to call for die rolls. “I’ll roll Persuasion,” says the player, “to see if the guard lets us in.” Then the player tosses the die. The GM’s input is not needed until the player announces their result.
In ages past, the Game Master’s judgment prevented failures at routine tasks and stopped incompetents from succeeding because no Game Master worth his screen would ask a player to roll to resolve a routine task or an impossibility.
And speaking of the Game Master’s judgment…
The Action Check Crap Shoot
While I do think D&D’s mismatched non-skill-based-but-still-skill-based presentation and design are the biggest factor leaving players pissed off at the poor d20, I don’t think it’s the biggest problem affecting people’s games. And I know there’s nothing most Game Masters can do about that mismatch anyway.
In the end, I think what’s leading to the dissatisfaction over the d20 mechanic has a lot to do with a sense of injustice or unfairness. After all, when people complain about this shit, they complain about failing rolls they shouldn’t. They complain about not succeeding when success is deserved. They’re complaining about fairness.
And I think most Dungeons & Dragons 5E Game Masters run totally unfair games.
Weren’t expecting that, were you? I’m just full of surprises.
But hear me out. This is important.
When your players come to a locked door or stubborn guard or a wilderness to navigate or whatever, the D&D system provides a simple means to handle that, right? A character with the right skill steps forward and rolls a skill check, and if they get a good result, the problem’s solved. Easy right? But let me ask you a question:
What can your players do to affect the outcome?
Module writers and homebrewer Game Masters tend to assign specific skills to specific challenges. Very occasionally, there’s a multiple choice list of skills of the “if the characters try to pick the lock, do this, and if they try to break down the door, do that” variety, but for the most part, Game Masters and module writers decide in advance which skills can overcome which obstacles.
The players’ odds are determined wholly by the game mechanics. The module writer or Game Master sets a DC based on objective criteria described in the core rulebooks. The character’s modifiers are determined before the game even starts. They’re determined during character generation or when leveling up. Some characters do have resources they can spend to enhance certain checks once in a while, but they’re rare and limited in most cases.
The point is, when the players come to a locked door or stubborn guard, they don’t have any say in how things play out. They must use Lockpicking or Persuasion, the DC is set, and their modifier was determined during character generation. Whether they realize it consciously or not, the players don’t feel like they’ve got any say in the outcome. The only thing a player can do to improve their odds of success is to build the best damned character they can build. Winning is about going in with the best modifier. And that’s all it’s about.
So when you go in with the right skill for the job and the highest possible modifier you could squeeze out and the dice hand you a failure, there’s literally nothing you could have done differently. Worse, there’s nothing you can do differently next time. Nothing to learn.
Months ago, I pointed out that the Advantage mechanic is a very nice, very streamlined mechanic that totally robbed every Game Master of the chance to hand their players a pile of circumstantial bonuses for their cunning plans and crazy capers. And when I sat for an interview on the Redcaps Podcast, I explained that situational modifiers are the Game Master’s primary means of empowering the players to affect the outcomes of their actions. Dungeons & Dragons 5E doesn’t give the Game Master any tools to reward the players’ approaches to the games’ challenges.
Gone are the day of players saying, “Wait, those metal bars in the corner… can I use one as a prybar?” Players don’t seek those sorts of situational advantages and Game Masters don’t reward them.
I think this is also why the d20’s swinginess isn’t a more widespread complaint. In theory, D&D 5E’s schizophrenic design and skill-system skinsuit should lead to an epidemic of swinginess complaints. But there are a lot of good Game Masters out there who encourage and reward clever play and empower their players and, as such, those players never get that initial feeling of unjust lack of control that leads them to read long-ass ranty screeds about dice probabilities.
And if you take nothing else from this long-ass, ranty screed, that’s the lesson to carry away.
Part III: Fixing the Issue
Over the past few months and especially in the days since I announced I was weighing in on this topic, there’s been chatter on my supporter Discord server about fixing D&D’s action resolution to mitigate the d20’s swinginess. Or its perceived and apparent swinginess. And since I mentioned the issue in passing months ago, I’ve gotten a lot of e-mails proposing solutions.
Here’s my advice: don’t waste your time on this shit.
Switching dice and rebalancing modifiers won’t help. And I think changing to a 3d6 or 2d10 system will have nasty side effects given the way modifiers interact with bell curves. See, modifiers in bell curve systems have a diminishing impact when the odds of success are already high or low. And that’s precisely when, from a game design perspective, you want modifiers to count the most. You want strong modifiers that let players even the odds when they’re at a disadvantage. Or that lets monsters even the odds when the players’ victories seem a foregone conclusion.
Moreover, the last thing you want to do is make heroic fantasy adventure more skill-based. Bell curves — such as they are — work best in skill-based systems about hyperspecialized experts and tag-in-tag-out Eigenchallenges.
Meanwhile, some folks have recognized that a lot can be solved if the Game Master just calls for fewer die rolls. Which is precisely why I advised y’all not to let players roll dice if their characters shouldn’t succeed or shouldn’t fail. That said, the mathematical systems some of you sent me for determining when dice shouldn’t be used are pretty wrong-headed. Just use your frigging judgment. No need to do any math at all.
And no, you absolutely shouldn’t let your players decide whether to roll or use a Passive Score when success is nearly guaranteed. Players should be able to affect the outcome with clever, in-world actions, but it’s the Game Master’s job to choose the rules used to resolve them.
Honestly, I think the best thing any Game Master or player can do if they feel the d20 is too swingy is to just forget they ever read anything about it. It’s such a small issue, it’s very prone to confirmation bias, and it’s a feature in D&D, not a bug. It’s how D&D is supposed to work.
But you Game Masters should think long and hard about what your players can do at your table to affect the outcomes of their rolls. What choices can they make in the moment – not during character generation — to change their odds? How do you empower them to tip the scales in their favor? How do you encourage them? How do you reward them when they do? And when a character fails a check, what can the player do differently next time? What can they learn?
That said, I know this ain’t the last word on the swingy d20. Switching dice systems and arguing about math is easy. Fixing the broken-ass way you run games is hard. No one likes hard solutions.
There’s a ‘Hipsters & Dragons’ article that suggested the default DC 10/15/20 range doesn’t make much sense for 5e, since you’ll almost never get a roll you’re likely to succeed at (+5 against 15 means your level 1 “expert” will fail half the time at a “Medium” task, okay for combat where you’ll get three tries but bad for other tasks where you get one shot). But I think the additional points here that GMs call for too many rolls, that players can’t affect beyond character generation, are important parts of the equation too. I think a hybrid approach, calling for fewer rolls, adjusting DCs and bonuses based on player ingenuity, and using DCs that are slightly more likely to lead to success, would do a lot to prevent players from feeling like the dice are screwing them (except when the dice are supposed to screw them).
If a task it’s so easy that a normal person would succeed in it most of the time, the GM shouldn’t ask for a roll. If that is true, even a “easy” task is still a task that is so open ended that even a expert would fail almost 50% of the time. If not, the GM wouldn’t had asked for a roll. Therefore, a DC 15 it’s perfectly fine for a medium difficult task, or, most of the tasks a GM would ask a roll for. If it’s easy, even a “normal” person has a 50% chance of fail (remember if it’s more than that, you didn’t neede to ask for a roll. And, in this situation, the expert it’s very likely to succeed (75%). For difficult tasks, it’s a task so hard that even a expert would fail most of the time (25% success rate) and a “normal” person would succeed only on a miracle 5%. In other words, if you consider that a easy,/medium/difficult task is the easy, a medium and a difficult of an already hard enough task, the 10/15/20 DC makes perfectly sense to me.
“What can your players do to affect the outcome” was a great reminder of what I want to work on, for my game
I’m afraid it’s becoming a lost art on both sides of the screen, and **I** think it might mostly be down to the way the game is presented. I just run games, I don’t do deep analytics, but the original AD&D DMG felt awesome and empowering, whereas the 5e DMG feels timid and vacuous. It seems more focused on teaching players about the mechanical possibilities of their characters than teaching dms to run games, and I wonder if that isn’t a cynically deliberate ploy to sell more DMGs to players….
I should have said “avatars” instead of “characters”, up there…
Do not assume malice where incompetence provides a sufficient answer.
As you might expect from a company that considers their primary tent pole magic the gathering, Wizards simply has a culture of design in which the pre-constructed part of the game is really important. The argument away from situation bonuses is basically the argument against allowing magic players to add new cards to their decks in the middle of tournaments.
Dice-dismay seems to be, in part, the result of both the primacy of identity/pride and not just the expectation, but the ardent desire for expertise.
By identity I mean the player’s conception, ideally unchallenged, of who their character is expressed by what they do. To any rigid identity, dice are an invalidation, rather than the creativity-inspiring chaos they were meant to be.
By desire for expertise I also mean the desire for differentiation; for every character to have “their thing”, and ideally for no one else to have it so that it’s more completely theirs. If you, the strong barbarian can’t break down the door, but the wizard did, that means you don’t have a thing. Your identity, your role in this role-playing-game, has been compromised.
It melts with the issue and solution proposed here: that some dm’s don’t provide enough ways to succeed beyond the sheet, i.e. the identity. So the identity is all they have. If the dice damage that, what’s left?
It’s rough because I can’t deny the allure of having and acting out of a strong identity. Being the expert and having your party rely on you to be that expert for them is powerful, emotionally validating stuff. Anyone would want that. That the game won’t, or maybe even shouldn’t, encourage that state of mind is thorny.
I think you’ve made a very good point about PCs’ identity, Hatelift. It really seems to be a goal for almost any player – to have their character as an irreplaceable part of the team. But I think that’s what Angry’s after in this article – *fantasy adventure* TTRPGs aren’t about teams of experts.
While I could argue about skill systems not being the only and/or the most important difference for fantasy-adventure vs. heist-investigation games, I’d like to build on your idea here.
I think that how the game’s presented to the players is very important here. If a player (or a GM, for that matter) is told that “every role in a party is equally important” or that “the adventurers need each other so that someone covers for everyone else’s weaknesses”, they’ll be focused on finding “their thing”. To be an expert.
But if the game was about adventurers – people who seek out a different, non-default life of raiding ruins and fighting monsters – they need each other mostly to have someone to watch their back. Even if those other aren’t expert healers, it’s good to have someone carry you out of the collapsing dungeon when your leg’s broken. And that mindset is, imho, the crucial part.
In one of my games, the party didn’t have much of the “roles” covered. But I didn’t call for as many rolls as rules suggested neither. And it happened exactly what I wrote above. The party fought some undead in an ancient tomb. 2 out of 3 PCs got seriously wounded. And the strongest character just hauled them from there, saving their lives – even though he was “the fighter” of the party and his wounds were attended by others most of the time. That played into camaraderie rather than into single PCs’ identities.
I always say to my players during the character generation: “Make whoever you want to play, it’ll be fine. It’s not like you’re gonna win every challenge either way.”
You could design a game where characters only have a few very narrow skills, or skills that are mostly irrelevant for the game’s main challenges (the fish-out-of-water scenario). But when they do manage to bring their skill into play, they auto-succeed. That way, it would be normal for the party to face challenges they are not particularly suited for, and they would be expected to use ingenuity to overcome them. Finding a way to leverage one’s expertise would be a challenge in itself, but it would be very rewarding and strongly cement a character’s identity.
It makes me wonder if presenting checks as ability checks instead of skill checks would help adjust the thinking, and presenting proficiency as one possible bonus option in the tutorial mode.
“This door is stuck. DC 12 STR would do it. Lockpicks won’t help, but if you’ve got Athletics or found a lever you could add your proficiency bonus.”
…that proficiency bonus for player expertise just came out but it seems elegant. I might start to use that and see if it works.
Ironically that is how D&D5E actually works. They state the Ability with a skill in brackets. Like DC15 DEX(Acrobatics). And really the DM should ask for a Dex check, and then say, if you have acrobatics you can add it.
The issue is that D&D5E married the skills and abilities so tightly that people think in terms of skills, and thus ask for skill checks.
Worlds Without Number for instance still uses skill, and you ask for skill checks, but which ability score is applied to the skill check can vary. The skills are also overall vaguer than in 5E (Which is where i think 5E really failed, why have three different “convince” skills?)
“By identity I mean the player’s conception, ideally unchallenged, of who their character is expressed by what they do.”
I think part of the problem is that D&D 5E wants your identity to be determined by your character sheet and not by what you do. Myriad options are provided to players and sold as “you can tweak everything to make your character exactly as you imagine”. And that’s stupid. When I was in high school the kids who gave themselves bad-ass nicknames were mercilessly ridiculed because you don’t get to be a bad-ass by calling yourself a bad-ass, you have to ACT like a bad-ass.
Your identity within a table-top role-playing game, whether you like it or not, is entirely determined by how you play the game at the table. Nothing you can do away from the table (including every choice on your character sheet) is part of the game or your identity within the game. If your identity can’t survive a failed die roll you don’t have an identity.
5E tries to sell the idea that you can make this unique character and bring it to the table to show all your friends, but in reality you choose an archetype (or kludgy, unfocused mess, depending on the source) and it develops into a character as you PLAY the GAME (including rolling dice and dealing with negative outcomes).
I think one way this problem, and the “should players be able to see their own search checks” problem, is in how the die roll is represented in a lot of groups as representing the quality of your attempt. So if you roll a one on your check to keep watch that’s actually something my character should be aware of, as the performance so terrible it’s not possible to try less on keeping watch is putting in earplugs and falling asleep.
So a bad roll isn’t just “something happened” it’s “seized by the mysterious spirit of incompetence, your character decides to just not try hard.”
Oh, this has completely articulated why I dislike the artificer’s flash of genius so much. It’s because it encouragers players to look at a problem in terms of “how many modifiers can I add to my roll” instead of “what actions can I take to improve my odds of success?” Maybe whenever flash of genius is used I should ask the player what form it takes or what particular idea they had. If I wanted to be really unpopular, that is…
This doesn’t seem right to me, but I do engineery design stuff so maybe I am hitting everything with my one hammer.
If friends & I read a new RPG (maybe… an ‘Angry’ product line RPG that might come out in the near future and I’m very excited about), and during the game we have the same ideas how hard tasks* are, that seems like good RPG design, like a Miata transmitting road conditions to the driver- the design is intuitive and becomes transparent during use.
Maybe the very cool ‘Angry Tension Pool’ is an example of good design.
* Tasks where the characters have full information, as opposed to ‘the floor there is icy, but you failed to notice that’
I agree that the argument for the “d20 swingy” issue holds little weight when measuring the same binary outcome. For example, a d100 (rolling 2d10s) and a d20 both have flat curves of 1% and 5%, respectively. One could argue for more resolution and smaller steps, preferring the d100 (2d10) over d20, but this has nothing to do with “swingy”.
However, when it comes to damage resolution, using a probability curve (rolling 2 dice and adding together) can lead to less “swingy” results than a single damage die. For instance, rolling 2d6 yields a 7 16% of the time, while rolling a 2 or 12 is only 2.7% of the time. On the other hand, a 1d12 has an equal chance of rolling a 1, 7, or 12, 8.3% each. In this case, someone could argue for more consistent and reliable damage, resulting in less “swingy” outcomes.
Similarly, if one wants less extreme results on a skill check, using 2 or more dice and adding them together would yield lower tail risk. However, this is not an argument for being less “swingy,” but rather for lowering extreme outcomes.
In conclusion, the “d20 swingy” issue is a myth, not because of math, but because of method. A typical d20 system is a flat curve and is easily calculated. It is no different from a d100 system, which is just smaller increments. However, if someone wants to build a game system that reduces the probability of extreme outcomes or create an average damage system, using 2 dice (or more) and a bell-curve model can meet those needs, but it doesn’t necessarily make anything “less swingy.”
Its about game design and method which determines whether to use a flat curve or bell curve, an apple vs. oranges argument.
In conclusion, while I find this article to be articulate and offering sound reasoning, I must admit to feeling frustrated that it was necessary to write it. As someone who was eagerly anticipating another excellent article on your game mastery, I cannot help but feel disappointed that some individuals will still fail to grasp the concept. Nonetheless, I remain grateful for your continued contributions to the gaming community and look forward to reading more of your work in the future.
What games examplify the “skill-based, do split the party” paradigm? In my experience, even in teams of experts, splitting up is very unfavorable because you risk encountering an unexpected challenge that requires another team member’s skills, and then as a non-expert you’re pretty much powerless. You need to do extensive recon to avoid running into that problem, something that only seems possible in heist games.
The only games I’ve played where splitting the party seems viable are more narrative games with emphasis on the characters’ motivations and emotional states rather than capacities.
That’s Call of Cthulhu for example. Usually, at least during the “investigation mode”, the party splits up to gather as many clues as possible as fast as possible. Someone goes to talk, gather rumours, others examine some books or artefacts while the rest break in somewhere, etc.
There’s wholly different flow to such games. There’s no “risk” to take in splitting up there, since the setting is usually more “civilised”.
Interesting, I don’t have the same experience. In my Call of Cthulhu games, the party mostly stayed together except for very simple tasks. There was less physical risk to splitting up, but we still felt the need to stick together in case the skillset of another member of the team was needed.
Angry hinted at a genre that exemplifies this. Pick your poison: Cyberpunk or Shadowrun (aka Cyberpunk But With Elves). Splitting the party is pretty typical in those games, often because the things that need doing aren’t all in the same place, or because one or two individuals are a lot less conspicuous than half a mob, or because there are ways to approach problems that don’t require a given character’s physical attendance, or because an alibi is easier when you can claim to have been in any one of three locations when the $#!+ went down. Even without sufficient recon, players can figure out roughly where their skills will work the best, and if they’re wrong, most characters are not so hopeless outside their niche that they’re screwed.
I was a new DM when we started our current three year campaign and I’ve had to do some heavy lifting to untrain and retrain my players on various things. This is one of them, and I’m having trouble fixing it. Like you said, it’s hard.
I think GMs LOVE the idea of players having influence on the outcome using decisions and actions more than their character creation, at least I do. I imagine you have some tips for breaking old habits and encouraging players to use their brains to think outside their character sheet.
Just to play devil’s advocate, I think you can determine when dice should be rolled using math.
Without any other context, when you roll a die and the GM narrates the outcome it is the same as flipping a coin. There is no way to prove that the odds of that outcome weren’t 50:50. So the only universal, probability based advice on the topic of dice rolling that I think is worth anything is this: roll dice if and only if you would be happy flipping a coin.
But there always is other context. If I tell someone to make a DC 20 Agility check they know that the odds are (probably) not 50/50.
Context changes your expectations about the odds (probably better than 50/50), the possible outcomes (if I fall from this height I’ll probably die), and the stories you tell yourself (I would have died if my bonus wasn’t so good) but it doesn’t change the fact that one thing happens and other things don’t. When the dust settles there’s either a 100% chance that you’re dead at the bottom of a cliff, or 100% chance that you aren’t. What matters is that the dice imply that other things COULD have happened, even though they didn’t. The actual odds matter very little.
At the end of the day rolling dice is way closer to flipping a coin than it is to the GM deciding the outcome with their judgement. If the context of the roll makes you uncomfortable flipping a coin, that’s your gut telling you that you don’t want to randomly determine the outcome because the context is telling you what SHOULD happen.
I think you somehow forgot your original argument with this reply. Your original argument was: The closer to a 50-50 odd of success/failure you have, the more you should rely on dice. The further away from the middle, the less likely you are to make the dice roll.
In this reply you somewhat muddle the waters with outcomes etc. (Partly because Sean seems to have not understood your argument)
I think the mathematical approach to “when should you roll dice” can be lifted from “passive skills” in 5E, if that’s your system of choice. If your passive in a skill is higher than the DC, no roll is required. (Of course 5E also breaks this system because of how high their “bounded” accuracy actually goes towards the higher levels.)
In Worlds Without Number the game straight up tells you no dice should be rolled if the DC is below 6. What the game actually tells you is to only have DCs of 7 and above. (It uses a 2d6 system, where most unskilled characters would have a -1 to their dice roll, making 6 the average). The game also states that you shouldn’t ask for rolls if it would make a character look incompetent in something they are supposed to be good at. But, as Angry states in the article: No matter how much a ruleset yells about dice rolling, players and GMs love to ignore it.
My argument was not as you describe. My argument is dice are coins. You flip a coin, you get what you wanted, or you don’t*. Dice don’t work in games because of probability, they work because of feelings. One of the big ones is feeling uncertain. Don’t use dice if you want certainty.
*(Different dice can bring feelings to the game that coins can’t, but it has nothing to do with probability distributions)
The whole concept of swinginess does seem like it’s only a perceived breach of verisimilitude for a small percentage of players to an already niche hobby. It’s interesting that no dice system I’ve used to determine results has ever felt wrong at the table. It’s only when reading a book, thinking about it, does it ever pull at my “this-is-not-realistic-o-meter”
Outside of being at the table though, such as rolling randomly for prep, meta-dice info is entertaining. An extra twist on a pair, or even more so for triplets, adds little sprinkles of some form of joy I don’t truly understand. I also use different color dice, so highs on one color might mean one thing, while highs on a different color mean something else. It turns rolling the dice into it’s own little mini-game of interpretation which I’ve come to enjoy. Most prep tables are d% with the interesting variations built into the table, so in any case still a non issue.
Also, for GURPS, it does have a 10pt over/under target critical success/fail chance, as well as degrees of success/failure, and is also a skill based system, so, conceptually, the bell curve seems to be useful. Although I’ve never run a GURPS game with d20, and come to think of it, I doubt it would feel broken if I did.
Interesting. I wonder if this problem only exists in specific game modes? I assume no one complains about the unfairness of their fighter missing an attack because everyone knows that even the best fighter can’t hit with every swing of their sword. The target number (AC in D&D) represents an opponent doing everything in their power to prevent that fighter from hitting.
Would it be reasonable to assume that people also don’t complain that their characters every deception in social situations should be believed? Or every attempt to pickpocket should be successful? If there is an adversary opposing the player, the player can rationalize their failure as the adversary merely being more successful.
But a wall you need to climb? Either you can climb the wall or you can’t, there isn’t any way to rationalize the failure away in the fiction. You either accept that you, the player, screwed up somehow… or you find a way to blame the dice.
Maybe we need a different approach to adjudicating actions that aren’t opposed like swimming or climbing or researching in a library.
This article does a great job of articulating why the mechanics of resolving social situations almost always suck. Rolling dice to determine the outcome of a conversation means that the actual words spoken were probably meaningless. In 5E there is no mechanical difference between a player forging some patents of nobility to convince a guard that they are a cousin of the king and should be allowed into the castle, and a player announcing “my character tells a really convincing lie to the guard to get them to let me in.” At best I could give the forger advantage on the check.
Don’t worry, I’ll cover this in an upcoming True Game Mastery Lesson. Because it is very important that both the dice rolled and the characters approach — not the specific words specifically said; but the strategy and information used — to matter equally. Which I’ve already touched on.
That said, the mechanics for resolving social situations in many roleplaying games actually do this pretty well. It’s just most people don’t use them correctly. Or even read them fully. The rules presented in the 5E DMG for running social encounters are actually pretty reasonably decently okay-ish.
I think an interesting wrinkle to this in specifically social situations is that we have an intuition that there’s actually a lot of different approaches that should work radically differently. “I’m going to threaten to sue the shopkeeper over his business practices to have his license suspended by the guild, I’m going to burn his shop down, I’m going to sue him in family court to have his daughter taken away.”, are three radically different approaches that should work completely differently all contained in the declaration “I threaten the shopkeeper”, and even saying something like “I threaten the shopkeeper’s shop” or “I threaten to sue the shop keeper” would tell you which of three approaches this PC is actually taking. And at some level of granularity you end up having to describe your approach so exhaustively you might as well just say it.
I don’t really agree. The “radical difference” might affect the outcome, he loves his daughter more than his shop, so the DC is easier.
Or it might come in the consequences of the action.
Burning down the shop could result in a city fire, or have you branded and arsonist. Going after his daughter might make the shop keeper hate you, and even try to pay people to hunt you down.
Sure they are all “Threats” but why should the action be mechanically different, when the outcomes can still differ.
It’s quite simple really, it’s the Cluedo approach: “My character wants to do [the thing] using [the skill] to affect [the target] by appealing to [the emotion]”
You don’t have to act it out, just state your intent and your approach to get what you want.
My point is that, as the DM, *because* the consequences can differ, I need to know which approach you take. And in the social sphere, you have to frame your approach with great specificity.
I guess I would at least request which lie they are telling. And will adjust the difficulty of the roll depending their proposition. For me it would be like telling : I kill the guard to enter. Without the “how”, the GM cannot adjudicate the action.
I would argue that social encounters where the players oratory skills matters isn’t a good approach either. That’s like having the strength character do push-ups whenever he’s trying to do an athletics check.
The point is your character has different skills than you. Not everyone is good at telling lies.
The difference in your example comes down to what proficiency the character has, forgery kit or deception. At the same time, I don’t know how it would really end up mechanically different in any case. Even if you just use a +2 bonus to a roll, the roll would still happen.
I think there is a little bit of overlap with the issue of mass-rolling / repeated attempts, which I think Angry has written on before.
In 5e, if a DM lets five party members all try the check at once, then for middling to high DCs it’s more likely that one of the four non-experts succeeds than it is that the expert will succeed. Outcome: gamers unhappy at the ‘swingy’ dice.
The 5e solution of course is to use the rules for group checks or help actions or whatever, but I suspect that in practise quite a few DMs don’t.
Same for rolling-until-success at high DCs when there’s no particular pressure. The obvious solution is to not call for any rolls and have the expert succeed after some time, but I suspect many DMs just call for roll after roll until somebody happens upon a high number (and again statistically, it’s probably not the expert who does). Again, outcome: gamers unhappy at the ‘swingy’ dice.
I feel like if the idea of a routine check was codified as a specific rule rather than a DM judgement call then players would understand it better, but I don’t think it should have to be that way
Reading about what you had to say about players calling their own rolls (not the GMs) may be a symptom of DND 5e’s stringently defined combat system. Ironically, I know you wrote a few articles about the “anime swoosh” and that the narrative options players have don’t break down during combat despite many GMs and players acting so.
I almost feel the opposite may be true, where despite claiming to have a “unified d20 system,” DND 5e seems to have two different “rule sets” – one with a set of pre-defined basic actions and class abilities for combat with an initiative system and action economy, and a more free-form mode based primarily on narration. (I feel there can be some debate over how valuable such a system is, like it’s great for grids-and-minis combat with some social intermissions, but not so for a group like mine that likes to focus on exploring and interacting with NPCs.) And then instead of treating these like two different modes, as I feel the rules point towards, players pull combat’s “here are the actions you can take” into the exploration/social pillars of the game, treating skills like the buttons instead of “basic attack” and “action surge.”
All this is to say, this gives me the interesting take that 5E would almost benefit more from leaning into an “anime swoosh” upon initiative being rolled to reestablish that out-of-combat situations are not meant to be treated with the same tools.
“But you Game Masters should think long and hard about what your players can do at your table to affect the outcomes of their rolls”
Nice take away. Hopefully after a few sessions with gentle prods towards problem solving the players start taking over some of that thinking, so long as GMs, pardon the pun, roll with it.
“situational modifiers are the Game Master’s primary means of empowering the players to affect the outcomes of their actions.”
Yes. 100% Underutilized at a lot of game tables. The weird thing is that many of the “celebrity DMs” out there make liberal use of situational modifiers. I wonder why that hasn’t lead to the practice gaining more mainstream acceptance.
I was recently reading through DnD 3.5, mostly inspired by Angry’s assertion that it’s the best edition of DnD.
One thing that I found odd was the discussion of modifiers, as it devotes a bit of space to the distinction between giving a +2 modifier to the player’s roll and reducing the DC by 2. When I read it, all I could think was ‘it’s the same thing, who cares?’ But I think I understand what they were trying to do now. The modifiers to rolls were for things the player did, and made it clear that this +2 was because of their clever plan (or this -2 because of their stupid plan).
Also because of the way that 3.5 constructed mechanical entities, lowering the DC on a check lowers the DC for everyone making the check, but the bonuses you accrue only go to you.
Which is interesting, because this actually implies that the players can do something with the first (personal bonus), to get the second (lower challenge for the group).
Like a player with a warhammer using it to hit their target in such a way to destroy part of their defense, like a tower shield, lowering the AC of the target for the entire rest of the fight and making it easier for the Group to hit them.
I get the feeling that most players (and maybe GMs) are actually not thinking about doing stuff like that nowadays, despite there being rules for destroying equipment and scenery…
TBH the rules for destroying held equipment like shields was just punishing enough that nobody I knew ever bothered with it, much like the rules for fighting from horseback or grappling.
If I may add my 2 cents, I think that it’s also often a matter of presentation, and that people are terribly uninspired when it comes to failure. There are hundreds of ways to fail and to succeed performing an action, especially if it’s properly declared and you know what the intention was. You can even do it retroactively, sometimes.
The example often used is of a hulking 18str barbarian failing to kick down the door and a tiny 8str wizard succeeding. But here’s the same situation in a way that doesn’t really trigger that “experts failed at their thing”.
Barbarian: I try to open the door by kicking it down.
DM: Roll strength, athletics.
B: 5, total of 12
DM: You fail.
Wizard: I also try to kick the door.
DM: Roll strength, athletics.
W: 17, total of 16
DM: You notice that the Barbarian’s powerful kick landed wrong, and one of the hinges got bent in a weird way, meaning the door is stuck. With the help of a piece of wood, you easily pop off the damaged hinge and the door fall down
Which highlights each character’s strength, it might even be a fun moment for the Barb.
The bigger mistake here is even asking the Wizard’s player to roll for the check when trying the same action, for two reasons:
1. If a task is so difficult that the one character best suited for the job can fail (and fails), letting anyone else try the same thing is most often a bad idea. Just announce the door to be too sturdy to be kicked down (regardless of the initially assigned DC) and invite them to try different approaches. Therefore, you’d be rewarding creativity and avoiding the dreaded “expert-failing-non-expert-succeeding” aspect.
2. If the task is such that one person can theoretically continue with their attempts until success is achieved (assuming that’s why you’re letting other characters “give it a go”), then just don’t request a die roll in the first place, especially if there is no risk or ticking clock involved. This just comes back to “roll less dice”.
Bonus: I’m being just nit picky here, forgive me, but what you described was a result of an investigation on the door by said Wizard, which was not the stated action and approach by said player.
That being said, I appreciate the creativity of coming up with a reason for the initial action to have failed. Game Masters often forget to include consequences to failures.
I think he means you narrate the wizard’s investigation as revealing the pre-requisite to succeed on the act of kicking in the door, and then the wizard using that pre-requisite technique to open the door with kicking.
So I clicked on the podcast and it’s the first time I’ve ever heard Angry’s voice and I have to ask: When did he find time away from managing Roman Reigns to write a blog?
I think the term “swinginess,” which I’m sure I have used too, is a carry over from weapon discussions. Especially the Great Sword and Axe. Where both weapons can do a maximum of 12 damage, but the Axe uses 1d12 and the sword 2d6. The Axe is more “swingy” because each damage result is equally likely to happen, the sword is more likely to give you 7s.
Here of course each result matters, and thus it’s up to the player to decide which style suits them best.
When it comes to skill checks I tend to keep a list of my players “passives” behind the screen. I will look at these to determine if they are likely to achieve a task or not when they make their character do it. That’s the baseline for when I ask for a roll or not.
Bounded accuracy i in 5E is bad, because they still kept the bonuses very high. At 1st level the difference between a good and a bad skill bonus is -1 to +5, and this difference will just keep growing. Once a primary ability is maxed out the lowest skill in the party is still possibly -1 and the highest can reach +11 and beyond. What good is it for a clever player to think in clever ways if he’s still going to be asked to roll that check which gave him a -1? Why is proficiency bonus +2 per tier and not just +1?
The way D&D is set up pushes players towards trying to get their best rolls.
You’d think a bounded system would also bound the bonuses available more.
Swinginess is still a misplaced argument for damage rolls. Imagine a game where the DM rolls all the damage secretly in a sound-proof box and the player’s handbook doesn’t list damage for different weapons. How would you ever know an axe and a sword use different damage dice?
The difference between a sword and an axe isn’t probability distrubtions, it’s feelings. Reading 2d6 on a page feels different than 1d12. Rolling two cubes feels different than rolling an icosahedron. Killing a dragon and saving the town feels exactly the same wether you did it with a 2d6 weapon or a with a 1d12 weapon. Decapitating a dragon with your mighty axe and shoving your greatsword down a dragon’s throat feel different.
You have made arguments from the point of view of outcomes. Sure. But, you also boil it down to single events. D&D isn’t single event games, it’s multiple events strung together. This is where your argument fails.
Why do I know it fails? Because I have played enough videogames where the dice rolls are “hidden” but the effects are still the same.
As you say it’s down to “feeling” but I can feel that axes are more swingy than swords. Why? Because I remember I felt that in Final Fantasy 12, where the game insisted Axes were the best weapons because their damage stat was highest. It’s like saying both a Sword and and Axe is 12 damage. The issue was axes had a damage space of 1 to their max damage, where as other weapons didn’t. I could see the low numbers come up more often with the axe than other weapons.
Hence even if you “don’t know” you can still tell that the axe is more swingy. When I roll my damage dice multiple times I can get a feel for how often I get the big damage, the low damage and the in between damage.
“Huh, I sure roll a lot more 1s with Axes, and I never have done it with Swords” I could then conclude that there’s a difference in how these dice are used.
Also: You can’t argue mechanics from a point of view of how the weapons “feel” to use. Because the weapons isn’t the point here, the mechanics are. I only mention Great axe and sword because those are two very similar weapons in D&D5E which aren’t the same. Where as using a Longsword or a Warhammer is exactly the same mechanically. It still creates different images in your head using one over the other.
Those multiple events are strung together by feelings, not math.
If the game shows you the numbers you can feel the difference, but if you don’t see the numbers you can’t. How could you count the frequency of “big damage” if the game doesn’t indicate how big your damage is? It can’t be the distribution that makes the difference because the distribution is the same either way. The difference is in your perception and how you feel.
You can ONLY productively discuss mechanics from the point of view of how things feel. The mechanics are never the point. People don’t play games for mechanics, they play games because they enjoy playing games.
People do not notice such trends accurately. In fact, trends in probabilities are perceived in tremendously biased ways. Confirmation Bias, the Gambler’s Fallacy, Hot Hand’s Fallacy, and many others exist precisely because people are bad at objectively identifying trends, their memories for even recent events are extremely faulty, and they are very good at imagining trends. So talking about how things feeling by referencing actually probability trends instead of psychology and fallacies is a waste of time. Your exposition is mathematically accurate and has zero to do with how actual people actually think.
Which is, again, the point.
THAT SAID, it is true that damage numbers are, in fact, the one place where the actual results matter. The outcomes aren’t binary. And yes, in that case, weapons that deal multiple dice worth of damage are more likely to produce AVERAGE results than either HIGH or LOW results. But how that’s perceived depends A LOT on the personality type of the player. It’s as common for most players to fixate on the fact that they aren’t rolling particularly HIGH results as it is for them to note they are much less likely to roll LOW results.
I’m a bit late to this, but I have to say: Very good article.
I used to be one of the people criticizing the “d20 swingyness”, for pretty much the reasons outlined in this article.
…and then I kept DMing 5e and sort of forgot about the problem.
Reading this article, I’ve got an inkling as to why: I started to really work on my DM skills, which included working through a bunch of Angry Articles and changing the way I adjucate PC actions and approach skill checks.
So I’m here to offer at least one data point that the suggested approach genuinely works: Stop worrying about d20 swingyness, make sure that your players have means to influence their odds of success and the frustration may very well disappear on it’s own.
Angry, somewhy this article is not showing up in Archive. You can find it in Random BS section, but not in Archive.
Now I’m curious if there are other secret articles that aren’t linked in the archive and won’t be accessible to general public.
As a PhD in statistics who has had to endure the metric tons of bullshit coming from players, DMs, and supposed designers, I have almost been moved to tears by your masterful rant. Thanks for doing the God of Stats’ work.