dungeons and dragons tabletop rpg theory

Why We Should Split the Party

Quick Reading Guide:

  1. Benefits of Splitting the Party
  2. Preserving Player Participation
  3. Addressing Concerns of Danger
  4. Wrap Up
  5. Further Reading

“Never split the party” is a tabletop RPG adage that’s old as dirt. Allen Varney joked about it in a Dragon Magazine review of Blood Brothers back in 1991. A couple of years later, William J. White wrote a defense of the concept in the same magazine, re-examining what even by then was considered classic role-playing wisdom:

While you can guess from the title of this article that I’m here to convince GMs that letting the players split the party is a good thing, there are two compelling reasons why the zeitgeist has for so long told us not to.

The first and most important concern is that splitting the party means that some players cannot participate in some scenes. If Kelly’s Rogue is on dungeon level 2, she can’t realistically interject while Lawrence’s Fighter tries to negotiate with a vampire on dungeon level 1. And if Lawrence spends 20 minutes talking to this vampire, Kelly’s going to spend a minimum of 20 minutes twiddling her thumbs or perhaps playing a few turns of the new Fire Emblem on her Switch.

The second concern is one of expectations and challenge. In most of the games where this advice is common, challenging the players and inviting them to risk the death of their characters is a big part of the game. Why put your ninth level Fighter in unnecessary danger by having her walk around with half a party? Why not always travel as a Wizardry-esque blob when the game allows it? In games like this there’s at least some element of a tactical challenge being posed to the players. Keeping their characters within range of each other is seen as smart play that minimizes the possibility of death. No one wants to accidentally roll a dragon on the wandering monster table and then get “monster immediately attacks” on the reaction roll when their character is alone and two days from retirement.

And yet, the benefits of splitting the party are great enough that I would suggest thinking of ways to address these two concerns just for the sake of having this happen. I think there are three major benefits to doing so:

Benefit 1: New Sources of Tension

Putting characters outside of their comfort zone is a natural source of tension. They have less resources at their disposal and they are more vulnerable. Tension in your average D&D-style adventure manifests most obviously in the fact that your characters die if they run out of hit points. If you look at the collective hit point total of your group as a “clock” on how many times you can take dangerous risks during an adventure, splitting the party effectively creates two smaller clocks.

What’s more, things that may not have ticked these clocks down may do that now. If a big burly, plate-armored fighter isn’t there to stand in front of a wizard in a split-party situation, things that would’ve likely bounced off of the fighter’s armor (dealing no damage and being essentially consequence-free) are now going to gradually tick down the wizard’s clock. In essence, splitting the party lets you create tension in places where it may not have existed before, using tools that normally would have caused no tension at all. This leads to…

Benefit 2: New Kinds of Play

Isolating characters to give them new challenges is a time-tested approach to writing ensemble stories, and while I would never suggest that you should be writing RPG plots, the process of creating scenarios to put your characters in has a lot of similarities with writing fiction.

When you play a party-based game or tell a story about an ensemble of characters, you’re probably going to see the characters start to play mostly to their strengths. And while there’s nothing wrong with that, it’s easy for a player in this situation to limit their participation to challenges specific to their character’s spec. A character who normally plays the “brawn” of the party may step aside when it’s time to sweet talk an NPC. After all, that’s usually a job for the party’s “face” character, right? Alternatively, let’s say you’re playing a game with less focus on character stats and one player is considered the “best” at bullshitting the NPCs. You might notice that other players let them handle that kind of challenge when it comes up. And there’s nothing wrong with either of these situations! But what if sometimes the characters and players can’t do that?

When you deprive a character of a commonly used solution, they have no choice but to find new solutions. Maybe a Fighter suddenly shorn of access to the Wizard’s Floating Disc spell now needs to find materials to build a makeshift bridge across a chasm. Or maybe a normally conflict-averse Wizard who finds herself facing down a barroom bully in town decides to test the heft of her staff. On the player side, people who may be a bit more reticent in a scene involving the whole group may more quickly volunteer a solution to a puzzle or ask more questions. All of these things lead to qualitatively different play than you normally see.

This feeds into…

Benefit 3: Quality Time

In your typical ensemble sitcom or movie, characters tend to be flatter when they’re sharing a scene with many others. When you’re rationing screen time between many characters, what each character does must express something simpler and easier to process about their personality. Their impact on the direction of the story is also different: in a scene involving a lot of characters, usually only a couple of them are making decisions that will drive the plot forward or influence other characters while the rest of the characters do or say things that perform other kinds of work for the story, such as establishing stakes or affecting the main characters’ perspectives. But if only two characters share a scene, there are no supporting characters and both of those characters are equally in control of what happens next. Any disagreement turns into an interesting push and pull that allows the story to explore what makes those characters different.

Player characters are in a similar boat. When they’re with four or five other people, we’re mostly going to see the “elevator pitch” aspects of their personality. And when it comes time to determine what the party should do next, the party will likely either democratically decide their course, or one or two characters will drive things forward. This isn’t a bad thing, of course. Player characters are controlled by people who know they’re playing a cooperative game and who want the game to proceed smoothly. As a result, players will naturally do things that make the most sense for the entire party rather than indulge in the most divergent aspects of their own character’s personality.

And yet, the most divergent aspects of a character’s personality tend to be very interesting! Splitting the party means that players have much less of a “majority” view to fit into, and their characters get a proportionately larger say in the fiction. We get to see characters interact with each other in ways that maybe being a group of 5 or 6 may have discouraged before. Maybe the normally gruff veteran warrior gets to show a little more vulnerability to the only other character who understands her, or maybe a party newcomer gets to ask some more sensitive questions like “What the hell’s going on with the Warlock, is he always like that?” or “Do we really trust Jerry?” You don’t need to do much to provoke this either; I’ve noticed that players who find themselves in smaller scenes naturally take the cue to speak a bit more, be a little more daring, etc. However, there are still things you can do to facilitate this! Check out “Role-Playing Incentives” a bit further down the page to see what I mean by that.

So in summary, you should split the party, even a smaller party, because it can expand the possibilities of play for your table and give individual players greater command of a scene for a period of time. But to do so, you’ll need to solve the two big issues we looked at earlier: player participation and danger.

Addressing the Player Participation Problem

Tool 1: The Spotlight

Gnome’s Fellowship offers a useful tool for maintaining consistent levels of player participation: the Spotlight. The Spotlight just means that you shift focus between players in the middle of a scene, even when those players aren’t currently in the same place. You may be using something akin to this already, but Fellowship’s section on the Spotlight gives some great advice on exactly when you should be shifting focus. Gnome recommends cutting on individual player actions, which makes sense when the Spotlight is the core source of initiative and player turns in your game. This means the Spotlight could sometimes be shifting in a matter of seconds. When you’re using a game with its own set of procedural timekeeping or turn-taking, however, you may want to adjust your use of the Spotlight to work in periods of minutes so that separated parties can make certain decisions as a single unit. Then you can move the Spotlight among individual party members in that party as you please.

In adapting this to my own OSR-style campaign, I’ve found that the best times to shift focus are when a player or sub-party has had roughly four to five minutes of solid play and one of the following things occurs:

  1. The players find themselves in danger or with a significant decision to make.
  2. The players have taken some action inviting consequences, and the players are excited to hear what happens next.
  3. None of the above has happened yet, but the players are looking at you to see what happens next.

In other words, you should be cutting between parties when players have something to occupy their minds while they’re not playing. It may be suspense about what’s going to happen to their characters, or it may be the cogs turning as they try to figure out the best move while you put the other party in a similar situation. Number 3 on that list is a bit of a fail state — you probably don’t want to go 5 minutes without some kind of interesting decision point or other source of suspense, but if it happens you want to “cut” during that lull. You don’t want to keep the other party waiting too long just because you’re focused on delivering one of the first two items on that list.

One last note before we move on to the next tool: some systems operate in 5-10 minute dungeon turns, which players may spend so they can thoroughly search a room or look for a way to bypass a trap. It’s okay to have one of these pass without passing the Spotlight back to another party (particularly if this happens very early in passing the spotlight), but just remember that this happened when you do return to the other players. Either let them know that at some point they’ll have to decide what they spend that dungeon turn on, or, if they conduct a lot of activity without explicitly spending one, just say that this activity is what occupied them for the turn.

Tool 2: Role-playing Incentives

When players split the party, consider using elements of the scenario that you have explicit control of to reward them. If the players are wandering around a dungeon you already designed, you probably don’t want to go changing rooms to create uniquely tailored challenges to that player, because that can be hard to improvise and, if it’s too obvious, may even hurt the player’s suspension of disbelief. But in a less concretely-defined area, such as while the party is in town, you have far more freedom to put NPCs and situations in front of them that uniquely play off of the composition of the smaller party occupying this scene. This tailoring of encounters should make more sense in town where the smaller party is going to be mentally filtering out or walking by events and activity that don’t interest them anyway.

Think about what your players have discussed or maybe implied are interesting questions they would like to explore through roleplaying and put things in front of them that help that happen. Maybe one of the characters are off on their own in town and you remember that they have a bit of a Robin Hood complex. Their time alone might be when you decide to show them some NPCs in trouble with a rich merchant. With no one else there to discourage them from drawing attention, this might be their chance to get into exactly the kind of trouble they wanted to. Or maybe one of your players is a Rogue who wants to engage in some good ol’ thievery without being weighed down by their plate armor-clad friends. Splitting the party may let them pull off some minor caper at night without the rest of the party hampering their stealth skills. In summary, whenever you’ve got a player character by themselves, ask yourself what kinds of play you now have access to that you didn’t before, and see if you can introduce that while you have the chance.

Whatever you do, don’t forget to make liberal use of the spotlight. The idea here is that you’re getting more out of the same time with the player, not allotting them more time.

Addressing the Danger Problem

The first thing to ask yourself when considering the danger problem is whether you’re playing a game where players leaping into danger is an expected or unavoidable thing. If you are, you probably don’t need to worry too much about players shying away from this for tactical reasons. In games where cautious play is encouraged, however, you’re generally going to do one of three things: 1) set a ceiling on how cautious the group needs to be, 2) give them situations where splitting up is a necessary risk, and 3) use the danger of a split party as a consequence to a decision they make.

Tool 1: Telegraphing Danger

Players are more likely to split the party if they know that they are playing a game where the choice to split up is not seen as a single action that may acceptably lead to unavoidable death. Many GMs will give hints to the players of impending danger to give the party a chance to avoid it — a concept often called “telegraphing” danger. This is less of a tool specifically for encouraging a split party and more of a generally effective piece of GMing advice, but if you endeavor to give your players more information on the risks they’re taking, they will take more risks. And splitting the party is of course a significant risk in an adventure site.

It is important that you know what it means to your group to telegraph danger. In a game where players are expected to play cautiously, there needs to be a consensus on what kind of risk-taking may acceptably lead to death. An obvious example of this is combat. If a player’s character has 4 HP, and they know that all weapons deal at least d4 damage, then they can probably be expected to know that being hit by a weapon will kill their character. But what if the player character is looking at a suspiciously-placed rug that seems to be cut down the middle set in front of a treasure chest? Is it fair if they walk onto the rug and it turns out to be a trap door, dropping them to an instant death on sharp spikes? The answer to this question is going to depend on each table’s shared assumptions on what level of caution is expected from the players. If you want the players to choose to split the party, you’re going to want to establish that the act of choosing to split the party still leaves success on the table in a way that walking on to that rug may not have.

Telegraphing danger is commonly understood to mean environmental hints as to what lies ahead, such as scattered corpses preceding a powerful monster. However, I’m also including ways out from danger as part of this tool. Give your players chances to resolve things without resorting to direct combat, let them flee from dangerous enemies, and give them more than one path (including physical paths in a dungeon) to their goal.

This section of the map from Magical Murder Mansion features a necessary key item in location 71, but players can follow at least three or four paths to it while feeling out each room on the way for possible threats. Players exploring this map get a lot of opportunities to reduce risk and choose paths of less danger.

Communication is key to accomplishing this: tell your players right from the start (such as at session 0 or maybe in your pitch for how the game will work) that you intend to telegraph danger proportionately to its lethality. An instant kill trap should provide an obvious tell or chance to mitigate its harm, but a trap that simply puts the party in a disadvantageous position might only need subtle clues or warnings. For example, players going into a mine should probably know (or be told their characters know) before going in that tunnel collapses are a thing that could happen, and a tunnel may collapse multiple sessions later. However, players about to open a door that will electrocute them for fatal damage should probably see some wires or other such obvious hints that they can interact with.

As for how the principle of telegraphing danger applies to split-party situations, the key thing to remember is that most combat- and damage-related danger is going to be more lethal for a split party than for a full one, which may affect the way you telegraph it.

If you’re playing an old school-style campaign, you’re likely going to be cautious about introducing situations that need to be resolved via violence anyway, but I’d recommend being a bit more alert about remembering to give the smaller parties chances to run, hints about what threats exist in the distance, etc. I’ve found that more puzzle-oriented dungeons are easier to do this with than dungeons full of monsters looking for a snack, so you may simply choose to let the players know early on whether this dungeon features a large number of prowling enemies that will attack on sight, and give them a chance to gauge the strength of the average encounter there. That may seem like a lot to tell them in the early stages of the crawl, but I find that the fun of the players choosing to split and gamble on the smaller strength of the individual parties often creates scenarios that make this worth it.

If you’re playing a less OSR-style campaign, or you tend to GM as more of a DJ putting stuff in front of the players on-the-fly than as a Referee running pre-written content, you can probably guess how to adjust here based on the above advice: give the players opportunities to parley, think hard about whether the monsters would retreat upon taking damage, convey the strength of an enemy group relative to the party’s, and let the party run. If forced combat encounters are a thing in your game even for split parties, your players should know that splitting the party doesn’t invite the same kind of encounters.

Again, I realize that this is all standard practice for OSR-style dungeon crawling games that rarely force combat, as well as for games like Dungeon World that feature strong guidelines on how “hard” a GM can hurt the party in a given circumstance, but it is worth pulling out of the toolbox for a different style of game if you find that your players fear death too much to split up.

Tool 2: Mechanical and Scenario-based Incentives

One of the most reliable and system-independent ways to encourage players to split the party is to reward them for doing so. The easiest way to do this is in a dungeon crawl where time expenditure can lead to forfeited treasure, more random encounters, and unfavorable changes in the dungeon. Parties will naturally split up if they realize they stand to gain more by exploring more of the map in a shorter period of time. However, you can write non-dungeon scenarios with a similar incentive easily enough. You just have to introduce multiple threats or problems to the party at the same time, and require the players to pick between one or the other or split up to do both. It may help to let them know that they aren’t expected to tackle such a decision in any given way, because some tables may assume that trying both options with half a party each is guaranteed to fail. As always, communication can help your table dispel notions of some choices being traps or otherwise not “real.”

In case you’re unfamiliar with enforcing time pressure in dungeon crawl scenarios, the classic method of doing so is the encounter roll: every 10 minutes or so of in-game time (a dungeon “turn”), the party has a small chance of encountering a wandering monster. Because wandering monster encounters can consume the party resources via combat, bribes or other conflict resolution, this encourages the players to weigh the consequences of thoroughly searching every room, always moving cautiously to avoid traps, etc. against having their resources depleted before they reach their goal. Classic editions of D&D also encouraged the Dungeon Master to change the dungeon over time, such as by having a rival party empty some treasure piles or by having some areas become more dangerous. However, this was usually a feature of the dungeon itself and not mechanically integrated into the systems’ timekeeping procedures.

To try and blend all sources of time pressure into the encounter roll, I like to roll the encounter die and assign to each number a random event that may result in any of the above-mentioned things. For example, a 1 may result in me rolling on the dungeon’s random encounter table and a 4 may result in some aspect of the dungeon escalating (e.g. a room slowly filling with water may flood). There are a lot of ways you can implement this idea, so I’d encourage you to explore it further if you’re having trouble finding ways to enforce time pressure in your dungeon crawls. Just remember to be as transparent with your players as you can while holding back the surprises you deem most necessary for that dungeon. Walking up to an empty treasure chest is meaningless if the players don’t connect it to their choice to search for traps one too many times.

If you’re not into the encounter die or are maybe not playing a game where resource management is as important, I would recommend instead checking out clocks in Blades in the Dark. Clocks in Blades in the Dark are a visual method of communicating increments of progress toward some kind of major change in an NPC, the environment, or something entirely abstract. Making noise while sneaking through a mansion might fill more “ticks” on a “security” clock that culminates in the entire location being put on high alert. Alternatively, players finding useful bits of refuse or other resources on a deserted island may progress an “able to build a raft” clock. In other words, you can use clocks to show your players whatever kind of change you want them to keep in mind when making decisions, and use progress on potentially “bad” clocks as an incentive to get more done in fewer ticks.

Whatever option you go with, just remember that the party needs to be able to tangibly appreciate what they gain and lose for splitting up. They need to know what kinds of things trigger encounter rolls, tick clocks, lose them treasure, etc. so that they feel like their decisions have meaningful tactical effects. If they don’t see an appreciable difference between splitting up and staying together, inertia will likely keep them together.

Tool 3: Dividing the Party as a Consequence of Risk

The easiest way to split the party is by force. You probably already agree that this is a sometimes food, but when is it appropriate? Like any potentially lethal consequence for the party, it should follow logically from the actions they took and the risks they accepted. Cave-ins, an enemy’s teleportation spells, hidden greased slides into lower dungeon floors, and trick rotating walls are all flavorful ways to force party division to provide a challenge. Applying what we discussed in Tool 1 (telegraphing danger), danger and duration of such splits should generally correlate with how foreseeable they are: a rotating wall in a mansion may only split the party briefly, but a party-initiated explosion that causes a cave-in can reasonably come with heightened and prolonged danger. Whatever you do, just make sure that the table can logically follow the connections between their choices and any harsh consequences that follow. In my opinion, a calculated risk that ends badly tends to be more exciting than spontaneous and unrelated calamity.

Wrap Up

Don’t be afraid to split your party. If you keep your focus moving between groups of players and time your cuts properly, you won’t meaningfully lose player participation and may in fact see flavors of it that you otherwise wouldn’t. Make the players feel comfortable with how danger is foreshadowed and delivered, give them good incentives to split, and you might find that they jump at the opportunity in spite of the risks. Finally, don’t forget that splits work best when they end in reunion. After all, these characters chose to be with each other for a reason! A fellowship broken may find itself stronger in spirit when mended.

Further Reading

Fellowship by Gnome. Fellowship’s Spotlight tool is one of the ways in which it handles disrupting its central motif with finesse.

DM David wrote a more succinct version of this article that I sadly discovered after already starting on it. While I’ve tried to expand on it as much as possible, his take is also worth reading.

William J. White’s article “Divide and Conquer” in Dragon Magazine #190 (mentioned above) has a lot of tips as well.

The Principia Apocrypha is full of good advice for running old-school D&D-style games, referenced above in the “telegraphing danger” section.

Dungeon Meshi by Ryouko Kui is the source for the header image. It doesn’t have much to do with this article, but the art is always a pleasure to look at.

1 comment on “Why We Should Split the Party

  1. A very good read – I will also have to read up on the Spotlight – I think most of us GMs knew about this instinctively, but it’s something worth exploring.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: