What is an adventure? An adventure is a series of encounters that all add up to and ultimately resolve some overarching conflict. Every adventure should be significant enough to make a difference to the overall arc of your campaign. No adventure should be so trivial that no matter the outcome the rest of the campaign won’t be affected, and no adventure should have such a massive impact that it obviates the results of one or more earlier adventures. As with all rules, this one has exceptions – maybe you want to spend a few adventures patrolling idyllic forests for petty bandits before that setting is completely obviated mid-arc by an extraplanar invasion and the rest of the campaign is about resisting your new Baatorian overlords – but stray from the general rule at your own risk: Every adventure should make a difference to the rest of the campaign, and no adventure should be so important that the outcome of earlier adventures don’t matter by comparison.
That’s not to say you can’t increase the stakes and scope as time goes on, nor that you can’t threaten everything players have done so far, just that what they’ve done so far should impact the adventure that threatens to end everything. For example, if players spend low levels defending a specific town from a goblin king and the arc ends with the goblin king attacking the town, the town’s ability to defend itself from the goblin king should be informed by the players’ success (or failure) earlier. NPCs they helped should contribute their skills to defense rather than fleeing town, lieutenants of the goblin king they failed to defeat should show up again with additional forces (and not merely replacing similar generic enemies, although replacing defeated lieutenants with significantly weaker enemies is fine if that helps maintain the pace of the climactic adventure), and so on. If one adventure threatens every adventure they’ve had so far, then every adventure they’ve had so far should have some impact on their odds of success.
Each adventure should have a theme, a through-line that distinguishes it from the other adventures. You don’t want to run the Dragon Age 2 of D&D campaigns. “This dungeon is totally different from the first one, because it has hobgoblins instead of orcs!” Unique and interesting dungeon design is its own subject (we’ll be getting into it soon) that can do a lot to alleviate this, but the best thing, the thing that will make your dungeon not just fun to play in the moment but memorable afterwards, is to have a strong theme. It’s what takes a dungeon (or a mystery, wilderness journey, etc. etc.) from being a pinata full of treasure and blood and turns it into a story.
So what is a theme exactly? In non-interactive works, a theme is a statement that the story conveys indirectly. Large stories have more than one, but to give an example of a theme from Lord of the Rings, “bravery triumphs over terror.” Lord of the Rings is constantly reminding the reader (or viewer) that the orcs are driven on by fear of Sauron’s might, and how horribly outmatched Free Peoples of Middle-Earth must stand against them even when their leaders are abandoning the ship. The defenders of Minas Tirith stand on the courage in their own hearts and not what Denethor compels from them. This is a non-interactive theme, a statement that the story makes without ever saying it out loud (Lord of the Rings sometimes does explicitly state its themes, and for a sufficiently noble tone that can work, but it’s treacherous territory).
For interactive stories there is a very important difference. A theme from the Dungeon Master’s perspective is a question, and the players will complete the story by providing an answer. Your players’ answers are very likely not going to be as clear as they would be in single-author fiction and they might not all settle on the same one. Also, since you never state your question directly, they probably won’t state their answer directly. If you ask your players what the theme of an adventure was, then unless they happen to be one of those people who enjoys literary analysis as a hobby, they probably won’t have an answer for you. However, if you tell them what the theme was – what question you were trying to ask and what kind of answer you think they gave – and they tell you that your interpretation makes sense and is interesting, then you know you’re doing your job right.
So how does that work in practice? Let’s take the Lord of the Rings example and twist it. Instead of saying “bravery triumphs over terror” we can instead ask “which is stronger, bravery or terror?” To ask this question, you present them with the forces of evil being driven on by terror and you present them with allies who are fearful of the might the enemy is bringing to bear. A few NPC lieutenants answering to the PCs keep them informed of what the men are feeling and serves as a small and manageable handful of faces to attach the hundreds or thousands of soldiers the PCs are commanding. One of the lieutenants quietly suggests that perhaps they should avoid giving battle when so badly outmatched. One is convinced that it would be better to die fighting than to let the BBEG win. A third despises cowardice and suggests reminding the soldiers that the penalty for desertion is death (or implementing this policy if it is not already the case).
The PCs have to decide what to do. They have to give an answer to the question posed. They can either try to rally the men with courage and give the same answer Tolkien gave, they can try to intimidate them with fear and instead answer that terror is stronger and only greater terror can defeat that inspired by the dark lord, or they may come up with another answer altogether. Then the players face the consequences of their answer, although it cannot possibly be overstated how important it is that these consequences arise organically from the narrative and not from any kind of GM anger over players giving the “wrong” answer. There is no wrong answer to give to the question posed by a theme, but just like it’s anti-immersive (and dickish) for an option to blow up in the players’ faces because the GM doesn’t like it, it’s also anti-immersive (though less dickish) to have an option not have any drawbacks.
All options, even the ones that are decisively the best choice over all, have some drawbacks compared to alternatives. Motivating troops with fear means that they will remain motivated so long as they think the PCs can deliver on their threats, while motivating them with courage means they will remain motivated so long as the PCs seem strong in their convictions – even if physically they are literally being ripped apart by wolves, if they die shouting “freedom!” their gruesome death may only rally their troops to even greater morale. On the other hand, motivating men with courage means that they have to believe in and agree with what you’re fighting for, and rumors of treachery or cowardice or even just inevitable doom can undermine morale in a way that morale derived from terror is largely impervious to. Rallying soldiers with courage requires skill checks while chopping off a troublemaker’s head and putting it on a pike as a warning to others generally does not (nor does one soldier usually pose enough of a threat to bother with a combat encounter), but that kind of morale will fail at the most critical moments, the times when the tide of battle is starting to turn against you and it’s looking like you probably won’t be in a position to decapitate anyone by nightfall. But who’s to say the battle will ever even reach a moment that critical? And so on.
The goal here is that, no matter what option players take (including one they make up themselves), try to give them as many reasons as you can think of to convince them they made the wrong choice. The key words here are “reasons” and “convince.” In order to avoid GM dickery, you must always be able to explain to your players why the things that are going wrong are a result of their choices. Particularly if you or your players are still nervous about the whole concept, actually, literally explaining to them why their choices directly led to the consequences helps to alleviate fears of GM dickery (although the ultimate goal is to build up enough player trust that you don’t have to spoil the exact chain of cause and effect immediately). The attitude you want to convey is that all choices have drawbacks, and players are simply being confronted by the particular drawbacks of the choice they made. If they made different choices, there would be different drawbacks, and those drawbacks may have been easier or harder to manage, but no road would have been entirely free of consequences.
To ignore the drawbacks of either option (or of whatever third option PCs may arrive at) is harmful to immersion, whether the GM is ignoring drawbacks because he thinks the PCs made the “right” choice or just because he’s going to give them what they want no matter what they choose. Real consequences and drawbacks make choices meaningful, give the sense that it mattered that you chose to motivate with courage over terror (or vice-versa). Tabletop RPGs should not be nailing themselves to paragade systems that were developed for video games run by computers which need things boiled down to a bunch of numbers and true/false flags to function. One of the greatest advantages RPGs have is that they are run by a human being who can intelligently interpret the consequences of any choice the PCs make, and to reduce those consequences down mainly to cosmetics the way that video games often do is to forfeit one of, if not the, biggest reasons to bother playing an RPG over a video game in the first place.
Just like with encounters, having a framework to help guide the construction of an adventure is helpful. There are, however, a much smaller number of types of adventure than you might think. In fact, the overwhelming majority of adventures are one of three types, and one of them isn’t very good.
The first adventure type is also the most classic, the dungeon. A dungeon is fundamentally a series of encounters connected by decisions about where to go next. Traditionally those decisions are defined by physical barriers, but you can also use trails in a forest or other wilderness area, or city streets in a city amidst an invasion, or a network of teleporters that connect rooms hundreds of miles apart from one another. What makes it a dungeon crawl is that you enter a room, have an encounter, and then make a decision about where to go next. We’ll talk more about how to make a good dungeon later, but when we say “dungeon,” we’re referring to this broader category, not just subterranean labyrinths.
What separates dungeons from other adventure types is exploration. Exploring a dungeon means wondering what’s down the next hall (depending on how lethal the dungeon is, this wonder may provoke anything from excitement to dread and often settles in the middle to a sort of hopeful apprehension). A dungeon crawl is best suited to adventures where the goal is to find something amidst a bunch of other things. That something might be spread around all over the place, like “treasure,” it might be in one specific location, like “the lair of the boss,” it might be in one place but moving around, like “the captain of the hobgoblins raiding this town,” it might be basically everywhere but you have to get all of it like if you’re trying to clear a dungeon completely for use as a base of operations, but whatever it is, players are trying to find a thing, and sometimes then want to do something with that thing (kill it, shove it in their pack, whatever).
The second adventure type is the mystery. While games like Call of Cthulhu and Dark Heresy have characters investigating mysteries that are typically some kind of actual crime, like a detective novel where the culprit is a shoggoth or an evil space wizard, a mystery is used any time characters’ main goal is to know a thing they don’t already know. The heart of a mystery is just a trail of clues leading to a conclusion. You can use a mystery adventure for tracking goblins down to their lair, finding blackmail material on a reticent member of the king’s council to pressure them into voting your way, or finding the location of a long lost magical artifact. Mysteries have a bad reputation in certain RPG communities because they are hard to run well, and a poorly run mystery is definitely extremely tedious, but they’re plenty of fun when done right, and we’ll talk about that further down.
The main difference between a dungeon and a mystery is actually quite small. In a dungeon you are searching for something in a small, well-defined space and you can search that space systematically. In a mystery you are searching for something but the list of potential hiding spots is too vast, or else the thing you’re looking for is abstract or otherwise insubstantial so a physical search doesn’t make sense to begin with, so instead you have to deduce what you’re looking for from clues. It’s a relatively small difference and yet it changes virtually everything about how the adventures are constructed and played.
The third adventure type typically encountered is the railroad. The GM designs a certain set of set piece encounters, the PCs encounter them one after another, and the GM fudges the encounter if it looks like it’s going to turn out the wrong way. The players have no significant decisions to make, are never at any real risk of failure, and at best can only pick from a very limited selection of goals (i.e. help Alice get the MacGuffin or help Bob get the MacGuffin). Often their goal is pre-determined entirely, and by definition of a railroad, their approach is also pre-determined. A lot of advice is given on how to make a railroad appear to react to player choices without actually changing the pre-determined list of encounters at all. Without real choices, there is no challenge, so for purposes of this guide it is strongly recommended you avoid this kind of thing.
We’ll be adding on two new adventure types on top of the dungeon crawl and the mystery, and those are the wilderness journey and the mass combat. The wilderness journey has some similarities to the dungeon crawl (and you can actually use a dungeon crawl system to run a wilderness journey if you want), however each room contains an encounter table instead of a specific encounter. After one encounter from that table has been completed, it is crossed off the list, and from now on whenever players roll that encounter, they scroll up or down in the direction of whichever encounter is their ultimate goal, usually the exit from one area of wilderness to another. Another big difference between a relatively small scale dungeon crawl and a wilderness journey is the ready availability of long rests and the relative difficulty of retreating to the safety of a town.
The most important difference, however, is that the sheer amount of space between encounters in a wilderness journey renders many of the most interesting parts of a dungeon crawl meaningless. It’s rare that a wilderness will be any kind of labyrinth, and the question is usually not “which one of these paths leads to our goal” but rather “which one of these paths is the fastest or safest route to our goal.”
The mass combat is a battle writ large. Units trade blows with one another on a strategic map and every round of combat can contain at least one, sometimes more encounters for the players. A mass combat is for the most part a war game and should be used sparingly unless your group is the sort that would be happy to sit down and play Twilight Imperium.
D&D has a long history of death being the only meaningful way for a party to be defeated. The protagonists of other media get defeated without being killed all the time. Hollywood’s stock three-act structure involves an act two down beat where heroes are defeated, and it almost never involves one of them being killed, let alone all of them.
D&D 5e already has a few mechanics in place to make death unlikely. Death saves mean that even a character who is completely isolated when they run out of hit points won’t necessarily die. A TPK will likely result in some actual deaths, but many characters will make their saves and revive. Declaring that all PCs make their death saves and revive a few hours later whenever they go down for any reason wouldn’t be any more of a reach than making short rests five minutes and long rests an hour (an optional rule presented in the DMG). Characters stir back to life, left for dead on the battlefield, some of them dragged off into piles of corpses not yet burned, others lying in a ditch where their “dead body” was shoved to make room for soldiers marching into the town they were trying to defend, some awaken where they fell, and now the scattered party must regroup in the ashes of their failure, the town burned, the good king captured, and the forces of evil triumphant. Everything valuable, most certainly including any magic items that weren’t concealed or disguised somehow, has been looted by the victorious enemy and is now being either used by powerful enemy lieutenants or else, if they can’t use it themselves, stored away as a trophy and treasure.
You can also use the popular ruling of having characters take a level of fatigue whenever reduced to zero hit points (which will flat-out kill them if it happens six times, but hopefully they'll take the hint before then, and if they don't, they've had enough warning that it falls firmly in the category of consequences for their decision). This gives even characters with absolutely nothing to lose some kind of disadvantage for being knocked down and either left for dead or captured.
Just because no one’s died doesn’t mean that PCs haven’t failed, and indeed, the fact that the failure state isn’t immediately story-ending (if not game-ending) means that the stakes are much higher, because the GM has no reason to be extremely reluctant to allow PCs to fail. Darths and Droids jokes about how the fact that PC failure means PC death means that any potentially lethal situation is one in which PCs can use their own imminent deaths as an ultimatum to get the GM to bend the rules for them. Not only that, if PCs are facing death and the end of the story for any failure, they are strongly encouraged to try and fast talk the GM into accepting whatever insane plan they’ve concocted to get them out of consequences scot free. The incentives are much weaker if the PCs are looking at a setback rather than the loss of their character.