What is an encounter? An encounter is a series of actions (resolved with rules and rulings) strung together around overcoming the opposition to a specific goal. It’s one scene of the story. It begins with a question and it ends with an answer. If there’s no significant opposition to the goal, there’s no obstacle to overcome, no actions to take to overcome it, and the player can just say “I do a thing” and you say “the thing is done” and it’s over. On the other hand, if there’s no goal, it’s not worth spending time on. Players may fight monsters for no other reason except that they’re there, but they will get bored with that about as quickly as they will get bored with being forced to describe their unopposed hike from one town to another in excruciating detail. This is why Monte Cook’s shortest possible adventure has both an orc and a pie. You need at least one complete encounter for an adventure, so the world’s smallest adventure is exactly one encounter with only the absolutely necessary elements, a goal (the pie) and opposition (the orc). As a corollary to this, whenever players face opposition to their goals, they should have an encounter.
The question that begins an encounter might be something like “will the party get past these marauding orcs alive?” Or it might be something like “which members of the Lord’s Alliance, if any, are willing to commit forces to combatting this month’s apocalypse?” The former is a wilderness encounter, the latter is a council meeting, but both of them begin with a question. Likewise, both of them end as soon as possible, as soon as the question is answered. In screenwriting they call this principle “in late, out early,” and it applies to most kinds of fiction, including role playing. Pace is greatly improved when you start a scene at the moment it becomes interesting, and end it at the moment the conclusion is obvious. If it’s become clear that the orcs are going to lose, or if it’s become clear which members of the Alliance are going to help the party’s cause, it’s time to wrap up, even if the orcs have some hit points left or there’s still some formalities to go through before the Alliance’s support is official.
The filling in the middle of the encounter is there to give players input on the question. The more important the question is to the overall plot-arc, the more input players need to have. It would make sense from the standpoint of realism alone to resolve a mass combat with the fate of the entire kingdom the campaign has taken place in hanging in the balance with a single opposed INT check, with some modifiers thrown in for one side or another having more or better troops, but it would not be satisfying for such a pivotal moment to come down to one die roll. Even if the players succeed, it will feel anti-climactic, and if the players fail, they’ll feel like they weren’t given a chance to succeed. Once again, the converse is also true. Players don’t want to spend dozens of die rolls on opening up a chest with a few dozen silver inside.
The length of an encounter comes down to more than just the amount of die rolls made or the time spent on it. Spending some time expositing on the battle before telling the players the results of their one die roll will help make it a little bit less anticlimactic, but not much. Likewise, making three different die rolls for three distinct stages of the battle, perhaps starting at the walls of a castle, then moving to the courtyard, and finally into the keep, will help make the encounter seem like a bigger deal, but it won’t give the players any more feeling of control. Their only influence is still just selecting the guy with the highest INT score and hoping he rolls well. Good description and NPC acting can be used to liven up an encounter for sure (we’ll talk about this more when we get to more specific encounter types), but to really get players engaged, you need to give them a game to play, and that means giving them interesting choices to make.
Many encounter types have a pre-existing structure that presents interesting choices automatically. Combat is usually the only one of these that’s present in the game rules themselves, but most GMs quickly come up with structures of their own for common encounter types (we’ll get into some of ours later on, feel free to steal them – we didn’t make any of ours up from scratch). If you’re dealing with an encounter that doesn’t have a pre-existing structure, you can use a general purpose system to paper things over. General purpose systems always pay a price in effectiveness for their flexibility, so if you notice yourself using the general purpose system for the same sort of encounter a lot, you should try to come up with a more specific system.
Generally speaking, it shouldn’t take more than five minutes to resolve a single decision point (you may need to hurry players along if they’re taking more than that amount of time to make a decision), so one or two decision points gives you (at most) 5-10 minutes of action, equivalent to a very quick combat encounter for conflicts that do need resolution, but are not very important. For example, opening a trapped chest has one, maybe two rolls, each of which should have some kind of interesting decision attached related to how the players open the chest. The worst that can happen is that the trap goes off and saps a little bit of HP, and the best is that the party get the treasure with no damage. Maybe rolls won’t be necessary at all because they opened chest from behind, so the poison dart is shooting in the wrong direction.
Three to six decision points is in the neighborhood of 15-30 minutes, which is about how long the average combat encounter takes. This works well for encounters like a chase scene or infiltrating a masquerade ball disguised as guests or servants, where the importance to the plot is about the same as any given rumble with the local dark lord’s goon squad.
Seven or more decision points means you are reasonably likely to take over half an hour (at twelve decision points, it’s about a full hour), and something that long should have similar weight to a climactic boss encounter, like escaping a collapsing doom temple after removing the magical doohickey that was keeping it up.
What exactly is a decision point? It's ultimate purpose, of course, is to be a point at which the players must make a decision which will influence the outcome of the encounter, but how does that work? First, you must come up with a decision to make. In the "escape the collapsing doom temple" example, players might be faced with a collapsing hallway and you can offer the choice to sprint straight through the bricks and hope to stay ahead of them on raw speed, or else analyze where they're falling and weave between the collapsing points. The former would be an Athletics check, while the latter would be Investigation, or perhaps Acrobatics (for guidelines on what DC to use for the check, see the Art of Rulings: Skill by Skill).
A clever player might come up with some novel way to approach the problem that you hadn't anticipated. Maybe this is just a superior way of taking one of the options you thought of, allowing them to roll the same skill with advantage, for example, a player using a class feature to rocket jump down the hallway would get advantage on the Athletics check. Maybe they have some kind of class feature that lets them flat-out ignore the problem, like a class with any kind of short range teleport just zapping themselves to the end of the hallway. Maybe they can think of a way to use a different skill to solve the problem, like using Mage Hand to toss bricks (which each individually weigh less than the 5 lbs. limit of the spell) out of the way, allowing them to roll Arcana instead. If you think that's a stretch, you might require them to roll at disadvantage, if it's particularly brilliant you might let them roll at advantage, and if you don't think it'd be particularly more or less effective than running down the hall, let them roll normally. Don't feel like you have to reward players with advantage for coming up with clever solutions unless their solution really is much better than any of the ones you thought of in advance. Being able to roll one of their stronger skills is already a reward. Contrarily, don't feel the need to punish players for thinking outside the box, either. Your default should be letting them roll normally, and advantage or disadvantage should only be applied if it's a really good idea or if it's really a stretch to apply that skill to the situation.
What happens if players fail a check at a decision point? The generalist nature of this one-size-fits-all system means that making the mechanics work will sometimes break the narrative and vice-versa. Although it's counterintuitive (being the opposite of things like the "make a ruling and move on" advice when you can't remember a rule), in this case it is better to make the mechanics work and use narration to justify the result. Specifically, if more players succeed than fail, give the players a point, and if more fail than succeed, subtract a point, and if they have a positive number of points at the end, they win. This keeps things simple which makes it easy to keep the pace up. It might not make perfect sense that the party's speed in escaping a collapsing doom temple is an average of all their results rather than being dragged down by everyone having to wait up for the slowest member (or abandon them), but a simple, generalist system was never going to make perfect sense for a specific situation. Players probably won't notice the seam if you keep things moving fast enough to be exciting, and even if they do, they're less likely to care if the escape seemed thrilling in any case.
Some groups will blow through decision points faster than five minutes a piece (especially in situations like the "escape the collapsing doom temple" where it's easy for you to find excuses to pressure them into choosing fast). Some groups will also go through combat encounters like a well-oiled machine, each player immediately ready to act on their turn and knowing exactly which dice to roll and what modifiers to add when they do so. The point of more decision points is not to make the encounter take more table time but rather to give the players plenty of opportunities to give real input so that their victory or defeat will feel earned. Don’t feel the need to stretch things out if climactic encounters regularly clock in under twenty minutes. So long as those twenty minutes are exciting and feel appropriately climactic, it’s all good.
For the rest of this chapter, we’re going to give advice on running four particular kinds of encounters, the ones that most often crop up in a game: Combat, chase, stealth, and social. Before we get into it, it’s important to note that you as GM do not set up combat, chase, stealth, or social encounters in advance. You set up opposition, and players choose how to engage with that opposition. Sometimes certain options just aren’t available (you can’t negotiate with mindless undead, nor can you sneak your way into winning the respect of the proud warrior race guy).
It’s important to keep in mind that players decide what kind of encounter it is because there’s another definition of “encounter” in a D&D context, which is an individual unit of opposition. For example, if there’s a group of orcs with an Eye of Gruumsh in one room of a dungeon, people might call writing those guys up and putting them in that dungeon room “designing an encounter,” and they’re not wrong. When you put a bunch of orcs and an Eye of Gruumsh in a dungeon you are designing opposition that may be encountered, and it’s not unreasonable to call that “designing an encounter” (although for clarity’s sake we’ll call it “designing opposition” in this series specifically), however that opposition needs to be ready to be a part of any type of encounter. We’ll talk about designing good combat opposition in one section and good stealth opposition in another and so on, but that is emphatically not because you should be thinking of a specific pack of orcs as “combat opposition,” but rather that you should make that pack of orcs suitable opposition for combat, stealth, social, and so on.