Setting the Stakes in D&D 5E

Stakes

What would you like to see in a 6th edition of Dungeons and Dragons? 

The last time this question made the rounds on Twitter.com I realised I had a very definite and specific answer. I'd be willing to pay full price for a 6th edition of D&D that was identical to the 5th edition in every way except for this: the PHB contains an additional paragraph, under the section on the core mechanic, about setting stakes. This is a crucial, crucial, crucial part of the resolution mechanic in the vast majority of TRPGs which D&D simply doesn't acknowledge. Here is an archived post on this topic from the days of The Forge (a phenomenon with which indie game designers seem to have had a long and passionate affair followed by a messy breakup, but which Wizards of the Coast seem blissfully unaware of). 

Why is setting the stakes so important, and how exactly should we go about it in D&D 5E? In fact, first of all, why should we go about it? Am I being unfair to WotC here by assuming the omission is rooted in hubris and incuriosity? Maybe they don't mention setting stakes because it's not actually an important thing in their game. But I think it is. I want to back up a bit and explain why, by addressing a common complaint I hear about 5E, especially from fans of earlier editions, which I think is actually a symptom of the lack of attention to stakes. 

WotC, I am available.

In Defence of Bounded Accuracy 

Here's the complaint: 

The numbers in 5E are too low. There isn't enough difference between a player character who's good at something and one who isn't. If the party want to examine some ancient magic runes, the wizard might roll adding their Intelligence and Proficiency Bonus for a total of +7, and the fighter might roll at +0, and although it's much more likely that the wizard succeeds and the fighter fails, it is still too common, over the course of a campaign, that the reverse happens. That +7 isn't enough to overcome the big stretch of randomness created by the d20, so we end up with frequent instances of the learned wizard being stumped while the big dumb fighter somehow knows all about it.

The complaint is about bounded accuracy, a design principle WotC settled on early in the development of 5E. Modifiers and target numbers stay within a certain range regardless of player character level, instead of being engaged in an arms race that results in bonuses of +40 against target numbers of 50 at high levels. The purpose of this is to keep things manageable - so the DM doesn't need to constantly scale up the target numbers as the PCs grow in power just to keep things interesting. After all, if the number I need to hit increases in lockstep with my bonus, then the numbers cancel each other out and I might as well not have bothered. Bounded accuracy is the solution to this problem. 

But, without the concept of setting stakes, it creates the problem above: someone with the right ability and proficiency in the right skill isn't really that much better off than someone without either. 

Or are they? Well, it depends on something rather larger and more abstract that lies behind the nuts and bolts of modifiers and target numbers, and even design principles. It depends on design philosophy, on how we conceive of the TRPG medium itself. What do these two approaches to the maths of the game suggest about the underlying philosophies of their respective editions? I'm going to compare 3.5 to 5E, skipping 4E because I'm not familiar with it (sorry - I'm aware it's reversed its fortunes and is now everyone's favourite edition - I'm a bad nerd). 

Design Philosophy Showdown 

When I finally got around to playing D&D for the first time, as a first-year uni student in a happier time, I was struck by something as I looked through the 3.5 Player's Handbook. The book's aesthetic had a pronounced strain of... let's say naturalism. Much of the character art is sketches with the guiding lines still on, details of hands and eyes, characters not in action poses but laid out before us for examination, Vitruvian Man style. The world of D&D is a strange land and the PHB is our guide to understanding its flora and fauna. A manual of Natural Philosophy for the curious layperson. By today's standards it pushes this conceit into rather dodgy territory around the point where we get sketches of the different playable races' skulls. 

Dungeons and Dragons and Anatomy Lessons

But it matches the design philosophy of 3.5 perfectly: take a look at the rules themselves and you'll see pages and pages of tables listing the target numbers for skill checks of various types. Climbing a wall - with many handholds - with few handholds - when it's slippery - and so on, for every skill. 

This is an RPG that regards itself as a physics engine: deterministic, mechanical, comprehensive.  The game's rules crunch through data about an imaginary world, throw in a random element and output "success" or "failure", with an unspoken assumption that the meaning of those words should be obvious in a given situation. 

5E, by comparison, regards itself as a tool kit. By default you play the game by simply describing what's happening. The dice and mechanics only come out at moments of uncertainty, and they have nothing to say about their own meaning in those moments. 5E expects us to figure out what "success" and "failure" mean for ourselves, based on the context of the fiction surrounding the roll. The game expects us to have already filled in all of that fiction through play, organically, instead of as an itemized list of factors. It expects the GM to pause for a moment before the dice are rolled, and consider "what does it mean if this character fails this roll? What does it mean if they succeed?" 

And the character themselves is part of this equation. Not just their ability and proficiency, which are already a numerical factor in the roll. That's just the tool we use to weight the die, it doesn't tell us what we need to know about who they are. I mean the fiction invested in that character. 

In our arcane runes example above, the wizard probably gets more information on a failed roll than the fighter does on a success. Perhaps the wizard is able to make an educated guess as to the significance of the runes without knowing quite what they say, while the fighter happens to recognise the runes for "cloud" and "porridge" while having no idea what to make of that. 

Here's another example. The characters are in a tavern and the bard decides to play some music for the patrons, hoping to make a good impression so the townspeople will be more helpful in future. They roll a natural 1. The next day they return to the tavern and the fighter decides to sing a song. They roll a natural 20. Does this mean the bard, trained from a young age in performance and music to a level of skill that is literally magical, happened on this occasion to break the e-string on her lute and fall off the stage? And the fighter, who has never performed to an audience before, somehow sang to make the heavens weep with joy? No, obviously not. It means the bard gave a technically flawless but uninspired performance and the townspeople were mostly indifferent, and the next night the fighter gave a raucous, so-bad-it's-good rendition of a sea shanty and everyone joined in because they were drunk. 

This is stakes-setting in 5E. Two important caveats: 

First, no one can be blamed for not getting this instinctively. As I've mentioned, the text of the game does a very poor job of explaining it. It may not present itself as a scientific manual the way 3E does, but it does seem to take for granted that the stakes of a given roll should be obvious. Don't be fooled. Play the game for a while and you'll find the design of the game very much expects you to take a moment to figure out the stakes each time the die is rolled. 

Second, all of this applies to ability checks. Basically none of it applies to attack rolls and saving throws. In combat, you're playing an entirely different game, one in which the stakes are set for you by an overdetermining system of turns, actions, hit points and special abilities. To see what it's like to carry the toolkit approach into the heat of battle, you need to look beyond D&D...

When You Delve...

Not abstract (done well, in this case by the wonderful Dyson Logos)

Maps and Not Maps

Here's a phrase that gnaws at my brain during the day and seeps into my dreams at night like water from a leaky ceiling: "abstract dungeoneering".

It's the name of an article by one Angry GM, and it's been approached in various ways by others, including Jason Cordova of The Gauntlet RPG community. Abstract wilderness exploration is its close cousin. The idea is that some adventure spaces are too big and mostly empty to be worth mapping. We draw a conventional representational map of such an area and when we get to the table find ourselves at a loss as to how to actually run it without it devolving into an extremely long, extremely dull conversation about t-junctions and empty rooms. Absent a better way of running such an adventure, we just avoid these kind of spaces.

For a while I've been using point maps for this, a technique I was made aware of by Chris Kutalik. You still have a map, it just looks like a flow chart. The nodes on the chart are the "rooms", that is the points of interest, the places where player characters make choices. The lines connecting the nodes are corridors, tunnels, roads, game trails or just the quickest route overland. This is helpful. It gives you a good, clear distinction between the parts where the players interact with the world meaningfully and the parts where you just give them a one-sentence description confirming that they do, in fact, walk.

But abstraction is a powerful drug, and soon enough this wasn't enough for me. I still had to draw maps. I still had to think about how the locations are connected. Most of all, with any kind of map at all I find myself trying to reconcile the structuring of a game session with the representation of an imaginary space. And do we really need to represent that space at all? No one is actually going to navigate it; what they're going to do is have a conversation with us about navigating it. Which bits of that conversation are important, really? Can we take the next step and do away with maps entirely?

Half abstract

The Move

So my take on abstract dungeoneering, presented in case you could make use of something like this, is this special move for Dungeon World and similar games - but it should be easy enough to adapt to D&D 5E and whatever else.

Delve
When you lead the party in exploring the dungeon or the wilderness, roll+INT or +WIS. On a 10+, mark 1 progress, or mark 2 progress and encounter a danger. On a 7-9, mark 1 progress and encounter a danger. On a 6-, encounter a danger. 

The GM's side is a matter of deploying locations and dangers from a list or random table. Here are some important terms:

progress: This is how far the players have penetrated into the dungeon or wilderness, or how much of it they have explored. It's an abstraction. The players keep a tally.

location: This is a place the players may stumble upon. They come to a new location every time they make the move, regardless of the result (0 progress doesn't mean they went nowhere, it means they didn't get any closer to their goal). Which location? Just choose, or roll on a random table. A location is like an entry on a key that comes with a conventional map. Tick it off once it's discovered.

special location: These are the locations that lie deeper in the dungeon or wilderness. It might be the specific place the party are looking for. It might be the exit, or the stairs to the next level. In a wilderness area it might be the entrance to a dungeon. (Use these "transitional" special locations to link together different sections of a dungeon). Each special location has a progress number associated with it. When the players reach this amount of progress, they discover this special location instead of a normal one.

danger: Obstacles, hazards, bad weather, and of course wandering monsters. These are what make the exploration difficult. For obstacles and bad weather you can dock progress if it makes sense: they can figure out a way past a caved in tunnel or unexpected canyon or river, or they can just backtrack and lose 1 progress (you maybe you shade in a section of a clock as well). Also, you can put the danger wherever it seems most interesting: they might get ambushed in a corridor between two locations, or the danger might be lurking at the location to complicate things when they get there.

Super mega abstract 9000

I've tested this exactly once, in a one-to-one game with a single player, but it seemed to go really well. A few thoughts:

The fact that the players find a location regardless of the result of the move is really important: this move is a method for navigating between locations, it doesn't replace exploration and it's not here to turn the adventure into a dice-rolling mini-game. Also, don't fall into the trap of thinking of "locations" as rewards and "dangers" as punishments for rolling badly. Dangers are obstacles and nuisances, and locations might well have some treasure or other boon stashed away, but it's all just loosely organised Content (tm). In fact, in general don't feel too beholden to the system. Its job is to help you structure an adventure without having to reconcile that structure with an actual representation of an imaginary place. If your players have come through three locations and two dangers and there's half an hour left before everyone has to go back to reality, just arrive them at the special location they're looking for regardless of what the move and the dice say. This is just like skipping over a section of a representational map to speed things up, except it feels less like cheating because you weren't thinking of your dungeon as an actual place to begin with. Your prep is a set of tools with which to run the game. The map is not the territory and all that.

This won't be for everyone. I'm sure there are many GMs and players who derive great pleasure from representational mapping, playing out navigation, and the sense that the game space is a fixed space. Certainly there's a style of play (at least one) where this kind of trick doesn't belong. But if it belongs in your game, I hope you get some use or inspiration from this. Happy delving.

Doom Clocks, abstracted info-gathering and a Clash of Traditions

The Doom Clock

In his PbtA-based, Ghostbusters-meets-D&D dungeon-crawler game Breakers, John Harper includes a particular conceit that grabbed my imagination and wouldn't let go, in the way certain ideas do. It's called the Cloud of Woe, and it's an unkillable entity of pure evil that coalesces and ruins the player characters' day if they spend too long faffing around. Mechanically, it's a countdown clock: every time they rest, or if they fail certain rolls, or stand around arguing, the GM shades in a segment of the clock. Once the clock is full, the Cloud appears and it's left to the GM how definitive the game-over is, but at the very least the PCs are going to have to flee.

Cloud of woah

The Cloud of Woe seems to have two very obvious implications: 1) someone urgently needs to design a TTRPG based on the popular John Wick movie franchise in which the players play criminals performing some kind of heist, but the titular boogeyman-assassin has been hired to stop them, and when the clock fills up the GM narrates the bloody deaths of all the characters in a storm of bullets, ju jitsu and improbably immaculate hair.

And 2), and more the focus of this post, this mechanic is easily generalisable to all dungeon crawls. I'm not suggesting it's revolutionary to include some kind of timer in your game to apply pressure and stop the players treating the adventure like a holiday. There are lots of ways of doing this, and countdown clocks in particular are a tried and tested trick, especially in the world of PbtA. But there's something so beautifully whole and definite about the Cloud of Woe. It's inevitable in a way the slew of countdown clocks you might use to run your sprawling, narrative-led Apocalypse or Dungeon World campaigns aren't. Those games treat their clocks as contingent, as descriptive as they are prescriptive of the fiction. Those clocks can be stopped or turned back or completely destroyed. This is because they are designed for keeping track of the stakes in an ongoing narrative that could spin out in any direction, not for imposing a structure on the game. In a game specifically designed around dungeon crawls (or other self-contained adventures) you want something that imposes a hard limit, a fail state.

So for a while I've been thinking about applying the Cloud of Woe idea to Harper's earlier experiment with the same kind of game, World of Dungeons. This is an ultra stripped down version of Dungeon World, and it's where the narrative-focused PbtA tradition overlaps with the dungeon-focused minimalism of the OSR. (I'm treating it as my jumping off point for pursuing my curiosity about the OSR, which could be very neatly expressed by a jealous girlfriend meme).

The Cold War's gift to tabletop roleplaying game design

WoDu leaves things a little more open than Breakers: the rules are clearly designed for simple, constrained dungeon crawling, but it doesn't give you the specific premise. So to project the Cloud of Woe back onto WoDu, we can call it the "Doom Clock". Each dungeon has a Doom. This is what happens when the time runs out. The dungeon is flooded. The ceiling collapses. The ancient god whose body the party is traversing awakens. Whatever it is, it's incontrovertible. It might well be the end of the party; it's definitely the end of the adventure.

Trying this out, I went with an army of metal-and-fire automata that stand motionless in the central chamber until the presence of intruders ignites their ancient animating magic and they awaken to mercilessly destroy our hapless level 1 treasure-hunters. Right now we're halfway through this two-session game and the clock is halfway full. So far so good.

Abstract Information Gathering and a Clash of Traditions

But it's here that the Doom Clock comes into tension with another (also rather Harperian) idea that I'm using: abstracted information gathering. That is, in place of the old school technique of the "rumour table", I decided to just have the players make a roll at the beginning of the session to see how much knowledge they managed to scrape together about the dungeon before setting out. In standard PbtA "custom move" speak it looks like this:

When you attempt to gather information on the trials ahead, roll +Intelligence (to study tomes of obscure lore) or roll +Charisma (to ask the townspeople what rumours they have heard). On a 10+, hold 3. On a 7-9, hold 2. On a 6-, hold 1. You can spend 1 hold at any time during the adventure to ask the GM for a relevant piece of information about the dungeon or its history.

It's pretty bland but it's hopefully functional: the PCs get some abstract "information points" that they can spend to learn something helpful when the need arises. No need to prep a rumour table and no chance that you spend precious table time playing out the acquisition of knowledge that might never become relevant depending on how the adventure shakes out. Lovely.

But when the players came across the army of constructs and opted to spend one of their points to find out what it was, I honoured the mechanic by essentially telling them that, yes, this is the Doom Clock for this dungeon: "you read about the sorceress's army of golems, which animate to defend the dungeon when they sense intruders. It might take them a little while because the magic is so old, but you think if you spend too long down here you're definitely gonna have to reckon with these guys".

Isn't this a bit of a copout? This kind of abstraction is a powerful way of skipping to the exciting part, but here it seems to come into conflict with OSR principles that I was hoping to square with the PbtA framework. Specifically, the OSR wants players to solve problems through their own ingenuity, not with a dice roll and stat modifier. When it comes to information gathering, that seems to mean they should concern themselves with the nitty gritty - and thus I, as a GM, should concern myself with it too. The non-abstract approach is to actually seed clues to the nature of the Doom Clock throughout, and let the players try to discover them. This is more work, and there's far more potential for it to go wrong or be done badly, but it means that when the players discover what it is they're up against so they can prepare for the exact nature of the cataclysm that awaits them if they get complacent, they've earned that discovery as players.

And that's pretty close to the heart of the disagreement between these two philosophies. PbtA gifts things to its players because it wants them to embody genre-specific archetypes, and that means they get to do cool stuff. Dungeon World's Fighter never loses her sword, and adventurers come prepared. But the OSR says if you want your character to be cool then you need to make them cool: make sure you don't lose your sword, and make sure you find the information you need. Where do we draw the battle lines in squaring these approaches? Are they even remotely compatible at all? Maybe I'll gain some more insight once I start my foray into the OSR proper.

In the mean time, if you'd like to join me in playing a whole bunch of different TTRPGs for five or fewer sessions at a time, get in touch on Twitter (@AlexBro97829019) and I'll send you a link to the Discord server.