My players have been doing a lot of climbing lately. In my D&D games, I mean. They’re not rock climbers, they’re gamers. In my regular, weekly D&D games, the player-characters have been doing a lot of vertical movement of late. And it’s not because they all just played Breath of Wild or one of the ten-thousand crappy imitators. It’s because of me.

I know that’s a weird thing to notice. And it’s a weird way to start an article. But this whole article exists solely because, as I watched one of my two groups clamber up the inside of a broken tower a few weeks ago, I thought to myself, “you know, there’s a lot more cliffs and shafts and broken towers in the Angryverse than there used to be.”

Full disclosure — and this is an important disclosure — I don’t run much D&D 5E these days. Not regularly. Not for fun. My for-fun home games use the good ole third-point-fifth edition. The World’s Greatest Edition of the World’s Greatest Game. And I’m running it vanilla. Pure, core 3.5, as the gods intended.

And that’s why there are so many things to climb in my world all of the sudden. Because three-five has a fumble mechanic. Sometimes.

Crits and Fumbles, Fumbles and Crits

Today’s topic du jour is crits and fumbles. Actually, it’s fumbles. F$&% crits. Crits don’t matter. While everyone’s pissing and moaning about DBox One changing the crit rules, I’m still waiting for fumbles. But they’ll never happen because D&D is only allowed to have things that make players happy. Which is on par with letting toddlers make their own meal plans.

Remember ice cream for breakfast is fun and adult-onset diabetes ain’t a toddler problem.

So today, I’m making the case that a fumble mechanic would do a lot more for the game than dickering around with crits. Hell, I’m going to prove that a properly implemented fumble mechanic would make it way easier for even crappy GMs to build non-combat challenges. And that fumbles are way more valuable than crits.

Let’s Get Definitional

First thing first. I’ve got to make sure you’re on the same page as I am. I know you think you know what crits and fumbles are — or critical hits and critical failures — but I can pretty much guarantee your definition is narrow and stupid.

Crits — critical hits — were invented by the Germans in the early 1800s and then refined at the University of Minnesota in 1974. I s$&% you not. That’s absolutely true. After being refined by a U of M professor, crits were included in the Empire of the Petal Throne TTRPG. But I’m not telling you anything here you can’t find out on TVTropes.org.

Real talk: usually, when people hear the phrase crit, they think extra-powerful attack. When a character in a game makes an attack, there’s a random chance that attack will do extra damage or inflict a condition or whatever. Crits are a gaming staple these days. They come in lots of flavors and have lots of names. The mechanics for determining when crits happen varies between genres and media. In shooters, a crit occurs when a player successfully lands a hit on a small, vulnerable portion of a bigger target. A headshot. And in stat-based, mathy games, the random number generator running the show sometimes declares an attack a crit.

Actually, I’m being much broader in my definition than most TTRPG players. When roleplaying gamer hears crit, they think roll a natural 20 and deal double damage. That definition’s too f$&%ing narrow. And even the definition above is too narrow. Double damage attacks are but one way to implement critical hits and critical hits aren’t the only crits there are. Non-attack actions can also have crit mechanics.

So, being as deliberately definitive as possible…

A Crit Mechanic is a game mechanic whereby there is some — at least partially luck-based — chance an action might succeed extra good.

Ask a gamer to define fumble — a critical miss — and they’ll similarly say something like, “that’s where you roll a natural one on an attack and then the GM describes how you impossibly decapitated yourself with your own dagger.” And that’s pretty funny considering how rare fumble mechanics are in TTRPGs. Dungeons & Dragons has never had an explicit, core fumble mechanic, but everyone knows if you roll a natural one, something terrible happens.

Anyway, to be deliberately definitive here…

A Fumble Mechanic is a game mechanic whereby there is some — at least partially luck-based — chance an action might fail extra bad.

See how the auto hit and auto miss on twenties and ones thing doesn’t count?

And to answer the comment I know is coming, no, crits and fumbles are not the same as degrees of success and failure. On paper, they lead to similar results in some cases, but they are not the same thing and don’t affect the game the same way.

The point is, to qualify as a Crit or Fumble Mechanic under Angrican Scripture, the mechanic must be at least partially based on pure, random chance — players can manipulate the odds a little, but not too much and they certainly can’t eliminate the possibilities — and the mechanic must result in enhanced — or dehanced — outcomes.

Warning: Hot Take Incoming

Are we clear on what I mean when I say Crit and Fumble? Good. Now, let me make some crazy assertions.

First, Crits aren’t worth jack s$&%.

Second, if you insist on putting Crits in your game’s combat engine, it’s way more critical that the baddies can dish out Crits than that the players can.

Third, TTRPGs without Fumble Mechanics are way harder to run properly than TTRPGs with Fumble Mechanics.

Fourth, Crits belong in combat and Fumbles belong outside combat.

Therefore, the best TTRPG is one in which the monsters can roll Crits on attack rolls and the PCs can Fumble non-attack actions.

Now, I can prove every statement above, but I don’t want to. Because I’m only interested in the third statement. That Fumbles — Non-Attack Fumbles — make TTRPGs easier to run and more engaging to play.

Fumbles in Disguise

Remember when I said D&D has never had explicit, core Fumble Mechanics even though everyone’s been arguing about Fumbles in D&D since the days when arguments were conducted by mailing actual letters to printed magazines and waiting a month to make the next point. Well, I lied. D&D hasn’t ever included an official, explicit, core rule like “on a natural one, you stab yourself,” but under the proper Angrican Church definition, D&D 3.5 includes not one, but two Fumble Mechanics.

First, if you roll a natural one on a saving throw against a magical attack, there’s a chance that an item you’re wearing or carrying might get busted. There’s a priority table and some random rolls and any at-risk item gets a save of its own… it’s fun, but it’s irrelevant. It ain’t what this article’s about.

Second, there are several skills in the Player’s Handbook that explicitly say, “if you fail a check with this skill by five or more, something especially terrible happens.” If you fail by a narrow margin — four or less — your action just doesn’t work. But if you fail by more than that, life sucks.

For example, Alan’s rogue is crossing a narrow plank stretched between two buildings during a rooftop chase. He’s booking it and needs a DC 20 Balance check to scamper across the makeshift bridge. If he rolls an 18, his balance falters. He’s got to check his speed, giving his foes a moment to close the distance. If he rolls a 13, though, he gets to find out what his knees look like from underneath. He falls.

It’s as simple as that. And there are a bunch of skills like that: a marginal fail represents a lack of success or progress in the attempt. A bigger failure means a fall off the tightrope or from the wall or a trap exploding in your face.

I’m going to call that the Fail by Five Rule. And it’s great.

The Unsung Beauty of Fail by Five

As I said in the Long, Rambling Introduction™, I run a lot of D&D 3.5 these days. As a result, I’m rediscovering a lot of s$&% I’d forgotten about it. And because I’m older and because I’ve spent fifteen years now critically analyzing the GMing arts and the design of TTRPGs, I’m seeing things very differently.

When I was a bouncing baby 25-year-old and D&D 3E was the new hotness, my mere fifteen years of experience running TTRPGs wasn’t enough to give me the design chops I’ve got today. I’m pretty sure I barely noticed the Fail by Five rule. I used it. Most of the time. I have dim memories of that. And I have dim memories of my players not really liking it. But it didn’t register as anything special.

Now, though, I recognize how much of a game changer — pun f$%&ing intended — it actually is.

I’m including a lot more non-combat obstacles in my game. And they’re a lot more engaging. But they’re also way less stressful to design. Moreover, I don’t really worry about — or care about — whether a die roll is necessary or not.

Consider a simple climbing challenge and the 3.5 rules that govern it. With a successful Climb check, a character can move one-quarter of their normal land speed up, down, or across a vertical surface for one round. The DC’s based on the thing they’re trying to climb. Steep slopes and knotted ropes are DC 0 — which a weak but encumbered PC can totally fail, by the way. A rough, natural rockface with lots of hand- and foot-holds is DC 15. And if a character tries to hustle — they can move up to half their land speed — they take a -5 penalty on the check. Easy, right?

Without any other stressors or ticking time bombs, I can get a short but engaging scene from a 15-foot cliff in the players’ way. An unimpeded character can scale it with two Climb checks. Impeded characters need three. And if a check fails by five or more, the character falls from whatever height they’re at. If the character just started climbing, the PC falls on his or her a$& and only their pride gets hurt. If they’re near the top though, that’s a sweet 1d6 of falling damage. And damage doesn’t just get hit-diced away during a short rest.

I know that sounds like just a bunch of die-rolling — roll three dice to get to the top — but that’s the danger of being too reductionist and analytical and not actually playing s$&% at the table before you judge it. Because it plays pretty fun.

What usually happens is that the party elects a trailblazing climber. Someone whose job it is to get to the top, cutting handholds along the way or securing a rope or whatever, so everyone else can climb easier. Being the best Climber’s a function of Strength and skill training and armor and encumbrance. And there’s always a question about whether it’s worth everyone removing their armor and gear and hoisting them up by rope to make the climb easier.

Anyway, the party settles to watch the trailblazer make their ascent. I — the GM — describe each result appropriately and track the distance covered according to the rules. Failures represent near slips, falls, pauses, or whatever. And everyone’s kind of nervous. Especially at low levels because falling damage can hurt.

Once the trailblazer hits the top, they use the Use Rope skill to secure a rope — unless they cut handholds with a hatchet or used pitons — and everyone else follows, rolling against drastically reduced DCs thanks to the trailblazer’s efforts. There’s usually a comical moment when the poor, noodle-armed wizard gets caught halfway up the rope, unable to progress because he can’t pull his own body weight up and his feet keep slipping off the wall.

It’s a fun scene. Tense and dramatic. And that’s just with the basic rules and a note that says “15-foot natural rockface, Climb DC 15”. Raise the cliff, add jagged rocks beneath it to double the falling damage, make it an overhang so there’s no wall to brace against, take away the stuff ropes can be secured to, and you get some pretty tense challenges with minimal prep work. And the same is true for lots of physical obstacles in D&D 3.5. And it all takes very little preparatory effort.

Now, I’m not laying all this s$&% at the feet of Fail by Five. These obstacles are easy to build because D&D 3.5 has lots of tools. But Fail by Five is the one tool on which all the others hinge. Because without it, physical obstacles are too much work to build and run.

F$&% the Five Simple Rules

Let’s change gears so I can show you just how fundamentally the Fail by Five Rule changes D&D’s core action resolution mechanic.

A long, long time ago, I changed the way everyone ran TTRPGs and established myself as the Game Masteriest of Game Masters when I published a list of guidelines for determining whether or not an action warranted a die roll. There’s a long story about how and why I wrote that article, but it ain’t relevant. What is relevant is that if I’d never switched over from D&D 3.5 to D&D 4E, I’d never have written it. And if D&D 5E had veered closer to D&D 3.5, it’d be obsolete.

In other words, under D&D 3.5, no one needed to know how to Adjudicate Actions Like a Motherf$&%ing Boss.

See, built into that article — and its precursor Five Simple Rules for Dating My Teenage Skill System — was this idea was that, with nothing at stake, players could repeat action checks over and over until success and those repeated rolls wouldn’t carry any sense of rising tension.

In other words, if there’s nothing at stake, don’t call for a die roll; and don’t use multiple dice to resolve actions.

But the Fail by Five Rule actually solves the problems inherent in no-stakes die-rolling. And it does it far more elegantly than my 10,000 words of claptrap ever could. And it did so without eliminating the underlying Success/Fail binary in D&D’s core mechanic. The Fail by Five Rule adds stakes to any roll it governs. And it usually governs the sorts of rolls that come up in tense scenes.

Moreover, the Fail by Five rule usually applies in situations that actually benefit from incremental progress resulting from repeated action checks. Which it’s perfectly suited for. Whenever you’re in one of those multiple die rolls to incremental progress situations, marginal failures ratchet up the tension even though technically nothing’s changed.

Remember that noodle-armed wizard dangling from a rope rolling over and over to drag himself up? Mathematically, on paper, that sounds dull as hell. It’s just making the same roll again until you win or lose for reals. But in the human brain, it’s a nightmare of ratcheting tension. Mix it with natural human Loss Aversion and the Gambler’s Fallacy that afflicts every human brain — even the smart ones that think they’re above such things — and you have this situation where the player — and his friends — are thinking, “this can’t last, if he doesn’t succeed now, he’s dead for sure.” And it keeps getting worse with every marginal failure.

Better still, that s$&% creates decision points that only exist in human brains. Again, mathematically, every check is independent and nothing’s changing. But as the tension turns to panic, the player starts thinking, “okay, I’m never going to make it up, I’d better start climbing down” or the character just starts screaming for rescue.

It’s all very human. The Fail by Five Rule is a perfect example of the MDA approach to game design. The math and the mechanics don’t matter, it’s the play dynamic that makes it work. It’s what happens when flawed and panicky human brains deal with that s$&% that matters.

The thing is, in D&D 3.5, I don’t have to adjudicate actions like a mother$&%ing boss. And I don’t have to build complex mechanics to model extended obstacles and progressive challenges. I just follow the Fail by Five rule, add some creative twists like overhangs and spikey rocks and swarming bats, and let the dumb human brains at my table do the rest.

Aided, of course, by my awesome narration.

Honestly, if you mix the Fail by Five Rule with the Take 10 and Take 20 Rules, you don’t need to sweat how motherf$%&ing boss-like your adjudication actually is.

But what about all the checks that don’t fall under the Fail by Five Rule?

From One Extreme to the Other

In D&D 3.5, only certain checks have Fail by Five outcomes. Only certain skills are Fail by Fivers. And while I do extend the rule to apply to a few skills the core rules don’t — like Appraise and Knowledge checks wherein a marginal failure means I tell you nothing and a Fail by Five means I lie to you and which I call my Dunning Kreuger Rule — while I extend the Fail by Five rule to a few extra actions myself, I don’t apply it to every skill and action. There are lots of situations where it just doesn’t fit. And forcing it’s no good.

Forcing it’s also unnecessary.

I know lots of folks decry the binary nature of D&D checks. Checks that result either in success or in nothing changing seem boring. And so, the obvious fix is to add nasty failure states and ugly consequences to every check. And you might be inclined to think I’m in that boat, based on my treatise on action adjudication. But I’m not.

There is nothing inherently wrong — or dull — about the idea that some checks lead to, “no result; try again.” In fact, those outcomes are good. The key to good pacing is variety. No-stakes and low-stakes situations are necessary to balance out the high-stakes and crazy-stakes situations. After the party’s crept along the narrow ledge over the bubbling lava pit and everyone’s done changing their pants, it’s fine if the rogue’s Open Lock attempt ends with, “well, crap, this treasure chest is being stubborn… I’ll have to try again, hold on.”

I know I crapped all over no-stakes die-rolling and unnecessary die-rolls when I wrote Adjudicate Actions… and Five Simple Rules…. I practically screamed that any GM who demands an unnecessary, no-stakes die roll should be boiled in his own guacamole and buried with a tortilla chip lodged in his heart. It was an extreme case, but I needed an extreme case just to get some of you f$&%ers to cool it on letting players roll dice just to get out of bed or eat a meal.

And it was a hell of a lot easier to say, “no low-stakes, unnecessary die rolls ever” than to come up with a formula for the precise number of allowable unnecessary die rolls in a given session.

Look, it’s fun to roll dice. And players feel like they’ve got control whenever you let them touch the dice. Rolling dice is tense even when the situation has no inherent tension. Because human brains are dumb. Rolling dice doesn’t take that long, either. Yes, you can bog down your game with a half-dozen unnecessary rolls in a scene that doesn’t need them, but one or two extra rolls here and there? That’s actually more fun, not less fun.

The point is, I put forth an extreme argument because D&D 4E had nothing, mechanically, to help you manage tension and die rolling. And D&D 5E doesn’t either. It gives you a bunch of rules for rolling dice — rules that carry no inherent weight or tension — and then expects you to figure out when to use them and when not to. D&D 3.5, by contrast, went to great pains to add tension — or the illusion of tension — to the die rolls that needed it most.

And, by the way, D&D 3.5’s Take 10 and Take 20 rules also provided a solid mechanical framework for determining when not to roll dice. And no, the Passive Check rules in D&D 5E don’t do that. Not at all.

If your game system curbs the worst excesses of dice misuse, you — the GM — don’t need good judgment. Sure, good judgment always helps and good GMs will always bring their judgment to bear, but there’s a lot less pressure on the GM to get it right every time if the game system’s doing the heavy lifting for you.

Don’t Take This Too Far

By now, I’ve probably got you convinced that Non-Attack Action Fumble Mechanics are the best thing ever and you should totally cram them into every game you run. And I’m sorry for that. It’s the danger of being so brilliant and charismatic. I can’t help but oversell s$&%.

You do not need Non-Attack Action Fumble Mechanics on every die roll. You need balance. And that’s how I want to close.

Non-Attack Action Fumbles work best as a threat, but not a constant threat. You don’t need Fumble Mechanics on every action. And, really, it shouldn’t be down to the GM to decide when they’re needed and when they’re not.

D&D 3.5 actually strikes a good balance. It connects the Fumble Mechanic to certain skills. The ones where it makes the most sense to have both non-success outcomes and outright failure outcomes. And also the ones that usually involve incremental progress and drawn-out actions.

Some game systems with Fumble Mechanics — or Complication Mechanics — demand the GM apply them to every die roll. I’m looking at you, Genesys and FFG Star Wars. And that ain’t good. When you force GMs to invent degrees of failure outcomes for actions that make more sense as succeed-or-no-change or some excessive form of super-failure or an action that either works or it doesn’t, either the GM shuts down or they end up spinning off nonsensical ridiculous results. Results that don’t follow directly from the action itself. Or results that result in external changes to the world.

Yes, it’s funny as hell when some NPC believes a lie so strongly, it turns the game into a hilarious sitcom situation, but that’s not what it means to fail at Deception. That’s a GM spinning bulls$%& because they’re trying to find an extra faily-fail because a player rolled a one and the GM came up with nothing that makes actual sense.

In the end, RPGs only work if the outcome follows directly and predictably from the cause and if the GM is never forced to think too hard to explain what a die roll actually means.