Historically, RPGs have included random tables and die rolls to give the GM an air of impartiality. No, I didn't kill your character. The dice killed your character.
Stakes are different from "let's just see what happens," though. Stakes are not the same as Tasks. Stakes are "if I roll well, I get X; but if I roll poorly, you get Y." So stakes are about conflict, not just pushing buttons to see what happens.
Stakes are related to Hooks. A hook is something one or more players cares about. Hooks can be restated as goals. A hook such as "Ragnar respects his mentor" can be restated as "Ragnar has a goal of honoring and emulating his mentor." Stakes are the chance to achieve or prevent harm to the thing one or more characters care about.
If you've been reading Run a Game, you already know I believe there are two types of stakes: Threats and Opportunities.
- Threats happen when a conflict threatens to harm or frustrate the PCs' goals ("The villain is tarnishing Ragnar's mentor's reputation. What can he do to repair the damage and stop the villain?")
- Opportunities happen when the PCs can take a risk to advance their goals ("Ragnar's mentor learned a powerful battle stance by visiting the monk atop the mountain. Can Ragnar replicate his mentor's great journey?")
What You Need to Know about Stakes
Setting stakes is a GM "core competency." Getting good at setting stakes makes your games better. This is because the clearer the stakes, the stronger the tension. Here's the formula for fun in RPGs, once again:
Story -> Problems -> Tension -> Excitement -> Fun
Stakes apply at every level of your game. There are stakes for a single die roll to complete a task. There are stakes for a series of tasks to overcome a challenge. There are stakes for challenges to resolve a scene or encounter. There are stakes for resolving the component scenes to complete an adventure goal. There are stakes for completing adventure goals to achieve the campaign goal.
You should make the stakes clear for every die roll. That's right, describe the stakes for every single roll of the dice. This is what people mean when they refer to Vincent Baker's "say yes or roll the dice" credo. The only time a die roll matters is when failure matters. The only time failure matters is when something is at stake.
What stakes you use set the mood of your game. RPGs with lighter mood tend to focus on death and treasure stakes, with the former being rare and the latter being common. RPGs with darker mood tend to focus on mature, gritty stakes like innocent people being harmed, descent into madness and despair, and utter ruination. Compare dying in 5th edition D&D to finding your agent's Solace exsanguinated in Night's Black Agents. D&D death is a temporary annoyance, but the loss (and possible Turning) of a Solace is a moment of horror in Night's Black Agents.
Your players provide you with hooks to generate stakes from. Here's the thing... you can only generate stakes from what the players tell you they care about inside the game world. The players tend to tell you what they care about when they write a character background, or buy into your plot hooks. If there are no compelling hooks and the character backgrounds are thin on pathos, you're stuck threatening to kill the PCs all the time, and that's boring.
Stakes help you control the pace of the game. You adjust the pace of the game with three sliders:
- How often you introduce stakes (instead of exposition scenes)
- How often those stakes are threats instead of opportunities
- How high those stakes are - they should build to a climax
Stakes let you frame scenes around the real dramatic question of the conflict, rather than the shape of a dungeon room and stats of its inhabitants.
The more you use life-or-death stakes, the less meaning life-or-death stakes have!
Character death cannot serve as the stakes of every encounter. This is mainly because you really can't afford to.
Let's do the math. Say you're running a game for 4 hours a week for a year. That's 200 one-hour scenes. Let's say you want about two characters to die as major dramatic events during the campaign. And let's say you want dangerous life-or-death scenes to feel really scary - with about a 20% chance a character could die in that scene. We divide the number of expected character deaths by the number of hours played, then divide that by how scary we want our life-or-death scenes to be. That tells us we have to limit life-or-death scenes to 5% of all scenes in the campaign!
2 character deaths
÷ 200 hours
÷ 20% chance of character death
= 5% of scenes
Another way to look at that math is to reverse it to assess how lethal a game like D&D really is. As it turns out, it's not very lethal. Some even say D&D is too easy. Say 50% of scenes in your D&D campaign are going to use life or death stakes. That means, to get 2 character deaths over 200 hours of play, scenes have to have a mere 2% chance of character death. If you raise the threat level, you will have a geometric increase in character death. Going to 20% lethality means you'll have 20 character deaths, or about one every other session. Your players will rebel, unless this is an ensemble method survival horror game.
Character death should rarely be part of the stakes, except in D&D and Pathfinder, where (after a certain level) character death is just a temporary status effect (albeit a nasty one). You can make life-or-death stakes more fun by using house rules like the Death Omens rule from Reinhart at Chaos Engineering.
And again, if you're playing D&D or Pathfinder, even death stops being all that high stakes once the PCs can bring back the dead.
Consider this: Your 9th level Pathfinder rogue is dangling over a 200' drop onto jagged rocks, but so too is a 10,000gp ruby. You have one action. Do you save yourself or save the 10,000gp ruby that will surely be shattered and worthless if it falls on the rocks below?
Answer: The ruby, of course! It only costs 5,000 to bring you back from the dead, and 1,000gp (each) for the two Restoration spells to fix the negative levels you come back with. You can afford to come back to life and still have 3,000gp of profit compared to losing the ruby!
Obviously 99% of players would role-play their character, make the choice of self-preservation and avoiding horrible pain, and watch the ruby fall. But it's a perverse incentive. What if the situation were murkier? What if it was a room full of traps instead of a precipice? "Here, guys. Hold my Scroll of Raise Dead while I go into that room and try to get the ruby."
Even in a game that's all about fighting, you can go a long time without life or death stakes. Here's a big list of stakes in combat other than life or death. (Remember how much I love examples? That list is so good, I don't feel the need to give you examples. Just go read it!)
Insights from History
Our RPG hobby grew out of wargaming. In a traditional wargame, like Warhammer, players build armies and models, then battle to see who wins. The challenge is beating the other guy. In Gary Gygax's group, they started moving toward having several players play single heroes, with one player challenging the group with horrible monsters. The origin of our hobby had no stakes but death.
As the game evolved, the players decided that their heroes were adventurers seeking glory and treasure. The stakes grew from "life or death" to "treasure or no treasure." Players quickly became fond of their characters. Soon they started developing personalities, histories, and individual goals for them. Dave Arneson started to structure games around an ongoing storyline, with multiple plot threads, colorful antagonists, and a vibrant campaign world called Blackmoor. Soon, TSR published Greyhawk, and other campaign settings followed. The available stakes started to include "success or no success" at adventure goals, in addition to treasure and death.
As time went on, new RPGs were developed, and some emphasized the personal agendas and many games were written to "focus on story." Today, we have real "story games" where the players play with the story itself; but back in the 1990s, "focus on story" meant "stakes other than death and treasure." It was that innovative.
Compare Vampire: the Masquerade with Shadowrun with AD&D 2nd Edition. The three games came out at about the same time. AD&D 2nd Edition is still mostly "death or treasure." Shadowrun is a "death or nuyen" game, but it incorporates other character goals - enmity with various corporate and supernatural villains, allegiance with various social groups and political movements, advancing freedom and justice in a world gone terribly wrong, etc. Vampire: the Masquerade was quite the innovation, focusing first and foremost on characters' personal (usually political) agendas, with death-related stakes taking a back seat -- and almost no treasure-related stakes.
Today, even D&D has incorporated more sophisticated stakes into the game. Despite 5th edition's "back to our roots" trappings, it asks players to assign their characters an alignment, personality trait, ideal, bond, and flaw right on the front of the character sheet. Those are strong character hooks the DM can play with.
Ideals and bonds provide clear goal-related hooks; and personality traits and flaws provide the players with opportunities to have their characters make mistakes without the player being blamed for them.
Nothing Wrong with Death or Treasure
If you've been playing since the 70s or 80s, or if you haven't played much outside of D&D and Pathfinder, you may have limited experience with stakes outside of D&D's core "death or treasure" stakes. In my experience, even in plot-heavy D&D and Pathfinder games, the stakes of individual scenes are often death or treasure.
And that's probably OK for a particular style of play. The classic "power fantasy escapism" of classic D&D doesn't really need stakes beyond death or treasure. While 5th edition D&D was putting plot hooks on the front of the character sheet, it also introduced a spell called Revivify that brings back the dead. Fifth level clerics can cast it, and it costs just 100gp. 5th edition is trying to satisfy all of the types of players most of the time, including the people who want to play beer and pretzels D&D, roguelike D&D, or more hyper-lethal old school D&D. Those players are happy making characters with joke names, hacking and slashing, and generally having a good time of it without worrying too much about plot hooks. That's a blast, as long as everyone at the table buys into it.
Run the game for the group you have, or if you're pitching a game to your players, make it clear what kind of game it is when you describe the mood, conflict, and main action.