“Never split the party.”

It’s a maxim of the roleplaying hobby.

There are two primary reasons why it’s a good idea not to split the party. First, it divides the party’s strength, making each smaller group weaker and more vulnerable. (This is a particularly bad idea if you’re playing in a paradigm in which (a) adventures are primarily a string of combat encounters and (b) those combat encounters are all carefully balanced to threaten your full group with destruction. Nothing good can come from splitting up in those situations.)

Second, the concern that splitting the party places an undue burden on the GM, who now needs to keep track of two separate tracks of continuity while making sure to juggle spotlight time between the groups.

There’s some truth to both of these, which is why “never split the party” is great advice for newbies.

But in actual practice, there are A LOT of caveats.

As a GM, for example, I absolutely love it when the party splits up. With just a modicum of experience, you’ll discover that running simultaneous scenes – which is the end result of splitting the party – is basically Easy Mode for effective pacing (which can be one of the trickier skills to master as a GM): You’re no longer limited to cutting at the end of scenes, and can now use interior cuts to emphasize dramatic moments and create cliffhangers. It also becomes far easier to smoothly cut past empty time and refocus on the next interesting choice. (I discuss this in more detail in The Art of Pacing.)

As for the players: If you’re in hostile territory… yeah, splitting your strength is generally a bad idea. But once your adventures leave the dungeon, there are going to be lots of times when you’re not operating in hostile territory and can gain huge benefits from multiplying your active fronts.

You can actually think of this in terms of action economy: If you all stick together, you’ll be stuck doing one thing at a time. If you split up, on the other hand, you can often be doing two or three or five things at the same time, stealing a march on your opposition or just moving quicker towards your goals. (This can also make it easier for different characters to pursue actions that play to their unique strengths, making it less likely for some players get stuck in “passive mode” watching other players do everything important. When two actions of the same type need to happen at the same time, it’s also a great opportunity for someone who’s second-best at something to get a chance to showcase their skills, whereas normally you’d usually default to having the character with the best modifier make the check.)

Two easy examples where splitting the party can be favorable are personal scenes (hard to woo your lady love with four wingmen/women hovering over your shoulder) and time-crunched objectives. I find this style of play is more common than not in urban adventures, where meeting with multiple people at the same time can be very advantageous; or one group can be handling the barter while another group is gathering rumors down at the Docks. Being able to pursue multiple leads simultaneously in a mystery scenario is another common example (you’ve got to catch the serial killer before they strike again!), but can be strategically a little more complicated as you run the risk of inadvertently stumbling into hostile territory on the wrong foot.

Now that we know why splitting the party can be awesome, here are some quick tips I’ve learned over the years when GMing split groups.

Tip: Split spotlight time by player, not group.

Regardless of splitting the party, you generally want every player to be contributing equally and to have an equal amount of time in the “spotlight” (getting to show off the cool stuff they can do, being responsible for the group’s success, etc.).

An easy mistake to make, though, is what I call the “lone wolf spotlight,” in which one PC wanders off by themselves and ends up getting fully half of the GM’s attention. This can sometimes be a symptom of disruptive play (with the lone wolf’s activities interfering with the other players’ fun), with the increased spotlight time inadvertently rewarding the bad behavior, but this is not always the case.

Solving the problem is just a matter of keeping in mind the general principle of everyone getting an even share of the spotlight: If one group has four PCs in it and the other group has one PC in it, you should usually spend four times as long focusing on the larger group.

(There can be lots of exceptions to this. Juggling spotlight time is more art than rigid turn-keeping. Maybe the smaller group is doing something much more important and gets a lot of focus in the first half of the session, and then you can make sure the other PCs get extra spotlight time in the second half of the session or at next week’s session. But the rule of thumb here is still to split spotlight time by player, not group.)

Tip: Put the PCs under a time crunch.

The party doesn’t have to be constantly facing do-or-die deadlines, but there are A LOT of benefits to making time relevant to a scenario. It makes every choice between A or B significant, because you may not have time to do A and then do B.

Bonus Tip: You don’t need to set a plethora of explicit deadlines to make time relevant. Just having a campaign world which is active – in which the players can see that significant things happen over time; that situations evolve – will create not just the perception, but also the reality that time matters.

One of the things that will happen as a natural result of this is that the players will be encouraged to split the party: Faced with a difficult choice between doing A or B, they’ll cut the Gordian knot by breaking into two groups and doing both.

So have multiple stuff happening at the same time: Two different patrons want to hire them for jobs on the same day. An unexpected crisis breaks out just as they’re getting ready to go do a thing. When they’re in the middle of doing something, use a cellphone or sending spell to have a desperate friend call them for help.

Tip: Cut in the middle of scenes.

I touched on this earlier, but it can be easy when running a split party to fall into the habit of playing out a full scene with Group A and then playing out a full scene with Group B. (Or, more generally, resolving everything Group A wants to do before switching to Group B.)

This approach can work (and there may be times when it’s exactly what you want to do), particularly if you’re blessed with a group who thoroughly enjoys audience stance and can happily sit back and watch the other players do their thing, but more often than not this will result in Group B becoming bored for half the session and then Group B becoming bored for half the session.

So read the room and cut before people start tuning out. In my experience, it’s almost always better to cut too often than it is to wait too long, so err in that direction.

Tip: Swap groups when you’re not needed.

Are the players in Group A discussing their options? Or rolling dice to make their skill checks?

Cut to Group B!

There are a bunch of more advanced techniques you can employ for effective, dramatic cuts, but watching for those moments when the players don’t need the GM to continue playing are absolute gimmes.

(It’s also another example of why splitting the party can be awesome: If Group A is roleplaying amongst themselves while Group B is resolving stuff with the GM, then everyone is magically getting bonus playing time in your session.)

Tip: Scale time to balance table time.

Another common pitfall is to think that you need to keep time between the two groups strictly synced. So if one group is staying put while the other is driving fifteen minutes across town, you would need to resolve fifteen minutes of activity with the first group before you could even think about cutting to the other group.

This can very easily make it impossible to effectively split spotlight time or cut between the groups.

The solution, of course, is to simply not keep time synced: Start running the scene with the first group, then cut “forward” to the other group arriving at their destination. Run a bit of that interaction, then cut “backwards” to the first group.

This can sound complicated, but in practice it really isn’t. You can stress yourself out thinking about all the ways that the PCs could hypothetically violate causality, but you either (a) say that isn’t an option or (b) perform a simple retcon or flashback to resolve the conflict (whichever is most appropriate for the situation). Your players will help you do this in a way that makes sense for everyone. (Plus, in practice, timekeeping in the game world isn’t that precise to begin with, so you’ve usually got a pretty wide margin of fuzziness to fudge things around.)

Note: What if the PCs have cellphones or some other form of continual communication unlimited by distance that can trivially breach the continuity between the two groups? Often, the net effect here is that the party ISN’T actually split: Even though they’re in two different locations, each group is able to participate in what the other group is doing (by offering advice, expertise, etc.). This makes it substantially less important to balance spotlight time between the groups, since players can grab a slice of spotlight for themselves over the comms line.

Tip: If the PCs split up and get into two fights simultaneously, make a single initiative list.

And then you just swap between the fights whenever the initiative order tells you to.

(This generally works regardless of what type of initiative system you’re using. If you’re using hot potato initiative, for example, you and the players can choose to throw initiative between the two fights. If you’re using the 2d20 System, in which PCs always go first each round but the GM can spend meta-currency to seize initiative for one of the NPCs, you can simply do that for whichever fight you want at any time. And so forth.)

Running two combats at the same time is often seen as one of the hardest possible thing for a GM to do with a split party; a kind of worst case scenario. But the structure of the initiative list can actually act like training wheels. If you’re trying to get a feel for what running simultaneous scenes should feel like – cutting between scenes, balancing spotlight time, etc. – I actually think running simultaneous combat is a great way to do it.

The Art of the Key breaks down how you can organize and format the room keys in your dungeon to make them easy to prep and run. But what do you actually put in the key? What do you fill your dungeon rooms with?

Every dungeon is a unique snowflake, but my general process is:

1. Make a list of ideas. “Section of dungeon that looks like a Mongolian temple,” “dracolisk hatchery,” etc. Often this is just a list of cool/necessary rooms that I want to include, riffing on the general theme or purpose of the dungeon. (You can use techniques like the Goblin Ampersand to juice a basic concept and help spur your creativity.)
2. Draw the dungeon map, making sure to include all the cool ideas. (This will often involve combining multiple ideas into a single area to make them cooler in combo.)
3. Flesh out the key.

“Fleshing out the key” covers a variety of sins, but our key tip today for making awesome dungeon rooms is to avoid having rooms that are just one thing*.

It’s not just the room with the goblins; it’s the room with the goblins and the bubbling stewpot and the chandelier made from kobold bones.

It’s not just a room with some bookshelves; it’s a room with bookshelves, a slate table with a hidden compartment, and a looted rug from the Cambarran Dynasty.

A good rule of thumb here is to include at least three interactive elements (i.e., things that require or reward the players for doing more than just looking at them).

There are a couple reasons why this is good praxis.

First, the interplay between the elements tends to result in gameplay that’s greater than the sum of its parts: A fight with goblins is interesting. A chandelier made from kobold bones is a nice bit of set dressing. Goblins swinging on a bone chandelier during a fight? That’s awesome.

Second, a room with multiple interactive bits (i.e., things that are worth checking out) will encourage the group to split up and all check out different stuff.

Lisa: I’m going to scan the bookshelves for interesting titles.

Roberta: Can I make a History check to learn more about the antique rug?

DM: Go ahead and give me that check. Sandra, what are you doing while they do that?

Sandra: I’ll examine the table in the middle of the room.

If each room just has one thing for the PCs to interact with, you’ll often find the group falling into a routine where one character defaults into being “the guy who checks out the one thing in the room worth checking out.” (Usually the rogue.) If you break that up with multiple, simultaneous interactions, over the course of the dungeon you’ll make it more likely that everyone in the group will feel engaged with the exploration.

* The exception are empty rooms. In my dungeons, empty rooms are rarely truly empty (in the sense of being barren and featureless). Instead, “empty” rooms commonly feature one flavorful thing which does not have a significant interaction. (“Cool mural! Okay, let’s go left.”)

The encounter creation guidelines in the Dungeon Master’s Guide and Xanathar’s Guide to Everything are both based on the idea that you know how many PCs are in your group. Then you do a table lookup, do a little math, and – presto! – you have a budget to spend on your encounter, expressed as either an XP amount or a number of creatures of a particular challenge rating.

But what if you’re using a published adventure with a party that’s a different size than the one recommended? For example, what if you’re running Curse of Strahd (“for four to six players characters”) but your group only has three PCs?

Going strictly by the book, you would need to deconstruct the encounters in the book to calculate the original XP budget, determine what difficulty the encounter was designed for (Easy/Medium/Hard/Deadly), calculate the correct XP budget for your PCs, and then rebuild the encounter using the new XP budget. (Which may or may not be possible with the original creature(s) used in the encounter.)

Here’s the tip: It’s a lot easier to adjust the level at which your PCs play the adventure than it is to redesign every single encounter.

• For Tier 1 & Tier 2 characters, increase their level by +1 for each “missing” PC from the party.
• For Tier 3 & Tier 4 characters, increase their level by +2 for each missing PC.

Or vice versa for additional PCs.

To put that another way, if a published adventure’s recommended level is X, then at Tier 1 & 2 use it for PCs who are level X + 1 per fewer PC and Tier 3 & 4 use it for PCs who are level X + 2 per fewer PC.

So if you’re running Dragon Heist, which is recommended for five PCs, with a three-person group, you’d either want to start them out as 3^rd level characters (instead of 1^st level characters) or run a prequel adventure or two to level them up to 3^rd level before using the published Dragon Heist campaign.

BONUS TIP #1

Challenge ratings are not that precise. They’re not designed to be a guarantee (nor can they be). They are a very rough approximation of “on average.”

Some “balanced” encounters will be easy. Some will turn out to be surprisingly difficult.

That’s okay. No game, no adventure, no session is about a single encounter.

The flip-side of this is that you don’t need to worry too much about getting an encounter exactly right. It’ll mostly get washed out in the general noise – the imprecision of the system, encounters being designed over a spread of challenge levels, situational conditions of the battlefield, and so forth.

This is also why Wizards of the Coast can release adventures “for four to six player characters” of a given level. Such adventures are designed for five PCs. They’ll be a little bit harder for four PCs and a little easier for six PCs, but it’ll be just fine.

BONUS TIP #2

In terms of strict math, the rule of thumb described here breaks down for very large groups of 9+ PCs. But there are more significant balancing issues based on action economy that make creating and running encounters for such large groups more of a special snowflake in any case. (Short version: Ten PCs, with all their attacks and all their special abilities, are able to wreak an amount of havoc that is out of linear proportion to a group of four. But, conversely, you can’t just use more powerful creatures, particularly at lower levels, because the monsters can one-shot individual PCs before they go down.)

For groups of 8 PCs, rather than running higher-level adventures, you can get pretty good mileage out of taking an adventure designed for five PCs and just doubling the number of creatures in the encounter.

For groups of 9+, adjust the encounter based on the difference between the PCs’ group size and a group size of eight, and then double the number of creatures. (This breaks down somewhere in the teens, but I would… uh… strongly recommend not running groups that large.)

This can create some weird narrative challenges if the encounter was, for example, with a solo boss or the like. But those are the types of encounters which really don’t work with large groups in any case, so you’ll just need to give them a little more TLC.

BONUS TIP #3

Whatever approach you’re taking to encounters — prebuilt, custom built, or otherwise — remember that you can always dial it in over time: If the encounters you’re building are too hard, trim the XP budgets in the future (no matter what the by-the-book math says). If you’re running a published adventure and the PCs are steamrolling the opposition, hold back on leveling them up. Or, vice versa, level them up faster if they’re struggling.

As you’re getting a feel for things, keep in mind that you have to miss by A LOT and for a very long time for “too easy” to not be fun.

You only have to miss once for “too hard” to be a campaign-ending TPK.

So erring on the side of easy is recommended. You can dial it up from there.

Combat is ubiquitous in tabletop roleplaying games. A little bit because the modern RPG evolved out of wargames; a little more because combat is mainstay genre element in the pulp fiction and procedurals that underlie most RPGs; but mostly because simulated combat is fun!

As a GM, therefore, a lot of your time will be spent running and describing combat scenes. To do that successfully, you’ll first need to visualize the battlefield. Only when you can truly see what’s happening will you be able to evoke it for your players.

ON THE BATTLEMAP

Part of this, of course, is keeping your vision of the battle clear and consistent. If the continuity of the game world is constantly morphing without rhyme or reason, it will undermine everything you’re trying to accomplish.

One way to objectively achieve this consistency is through the use of a battlemap, particularly in combination with rules that are designed to take advantage of the battlemap’s precision. When everyone at the table can clearly see the map of the battlefield and where everyone is located on that map, it’s trivial for everyone to stay on the same page.

When using a battlemap, however, it’s also easy to suffer from grid rigidity: Your mental picture of the battle gets locked into the physical, unmoving miniatures on the battlemap. This results in static, boring descriptions of combat in which the goblin is standing in their 5-foot-square and the barbarian is standing in their 5-foot-square and the two of them stand there, swinging swords back and forth.

To escape from grid rigidity in your descriptions, the first thing to realize is that characters are not simply standing in the middle of their square. Part of the abstraction of the grid is that you are SOMEWHERE inside the square, but we don’t know where. There’s even times when you’re “actually” in a neighboring square (because, for example, you’re punching someone in the face who’s “in” that square).

A miniature’s grid position on the battlemap is best understood as an approximation of the character’s location. You can almost think of it as a quantum superposition, where the character’s specific location and action only becomes certain when it is “measured” by description.

Once you’ve internalized this reality that the map is not the territory, you’ll be freed up to describe dynamic, active interactions on the battlefield. For example, let’s consider our goblin and barbarian once again:

If we imagine these characters standing in the middle of their squares and exchanging blows, we might say something like, “Athargeon lifts his two-handled blade high above his head and swings it down! The goblin tries to squirm out of the way, but isn’t quick enough. It’s gutted from shoulder to nave.”

And that’s cool enough. Nothing wrong with it.

But if we can escape grid rigidity, then we’re freed to introduce descriptions like this: “Athargeon dashes forward. His two-handed blade plunges through the goblin’s belly, driving it back into the bookshelf behind it. A shelf shatters and crystal balls fall from the ruin, splintering to pieces on the stone floor. The goblin scrabbles at the blade pinioning it to the wall, knocking more books to the ground, before giving a final bloody gasp.”

The mechanical resolution here is the same in both cases (Athargeon rolls a successful hit and deals enough damage to kill the goblin), but note in the latter description how the characters are now dynamically moving through the space and interacting with the environment.

Athargeon can’t drive the goblin back and the goblin can’t smash into the bookcase if they’re both stuck in the middle of their squares with their feet figuratively glued to the floor.

THEATRE OF THE MIND

If you’re not using a battlemap, of course, then grid rigidity isn’t a problem. But without the reference of the battlemap, you’ll need to keep track of the entire battlefield and everyone on it in your head.

With a little practice, this isn’t too difficult to achieve if you’re only dealing with a handful of combatants. It’s also possible to “cheat” a little by doodling a little sketch for yourself (and/or your players) to help keep things straight.

But as the number of combatants grows, this will become more and more difficult. You’ll start making mistakes, confusing your players and leading to a frustrating and unsatisfying experience. The other problem I have is that the more of my brain is tied up trying to keep the continuity straight, the less focus and creativity I can give to providing vivid descriptions of the fight! Ironically, the more epic a confrontation becomes in terms of scale, the more pedestrian its execution threatens to be as we all get tied up in the logistics of what’s happening.

What I’ve found useful is to think in terms of melees and battle lines.

A conceptual melee basically takes a group of combatants and says, “You’re all fighting with each other.” Anyone in a melee is assumed to be able to attack anyone else in that melee as the chaotic swirl of combat brings characters within reach of each other.

For example, a fight might break down into a “big melee in the middle of the room” between the fighter, the paladin, and a half dozen goblins, while the wizard and cleric are holding back outside the melee. But you might also have a fight with multiple melees: There’s a cluster of characters fighting at the top of the stairs and another cluster of characters fighting at the bottom of the stairs while the rogue, who isn’t in a melee, is swinging on the chandelier.

The point, obviously, is to conceptually simplify how many variables we’re keeping track of. We don’t need to keep track of exactly where the fighter, the paladin, and the half dozen goblins are in relation to each other; we just need to know where they ALL are in relation to the rest of the fight.

A melee should not, however, be used as an excuse for your descriptions of the fight to become muddy or generic. This can be an easy trap to fall into, but just as abstracting the goblin’s position to being “in that 5-foot-square” shouldn’t preclude us from describing him getting slammed into the bookcase, so abstracting the goblin’s position as being “in that melee” shouldn’t preclude vivid, specific description.

In fact, the flexibility of the conceptual melee can actually free both you and your players to greater heights of creativity in the actions you take and how you describe them.

There are limitations to the utility of the melee, however: situations in which the abstraction of the melee is either not useful or confusing. The solution, of course, is to simply not use melees in those situations.

(It feels like it should be more complicated than that, but it isn’t: Just Don’t Do It.)

One common case where melees aren’t useful is the formation of a battle line. In my experience, this most commonly happens in the tight quarters of a dungeon: In order to protect the squishy spellcasters, the melee-types form a line across the dungeon corridor to block the bad guys. If the party is large enough, you may see the formation of a second line with reach weapons.

Meanwhile, on the other side, the bad guys usually outnumber the good guys, so they’ll funnel in, forming an opposing line.

What I generally find useful at this point is to simply write out the line from left to right:

Athargeon           Edvinas                Nazli

Goblin 1               Orc                        Goblin 2

Jotting this down only takes a couple of seconds and can be done as the players declare their intention to form the line. Generally characters can attack anyone directly in front of them, or immediately to their left or right. (So in the above battle lines, for example, Athargeon could attack Goblin 1 or the Orc, while Edvinas could attack anyone in the opposing line.)

What you may realize is that you’ve effectively created a tiny two-by-three grid, just like the one you would see on a battlemap. In order to respect the tactical intention of the characters (to form a defensive line), this shouldn’t be handled as a conceptual melee. The relative positions of these characters should be locked in, just like they would be on a grid.

There will now be a temptation to really lock into the imagery of these two static lines exchanging blows back and forth. And there can be a lot of good material to work with in that.

But you’ll also want to be aware that this can easily loop back once more to grid rigidity: Bring the environment into your description. Let people take dynamic actions within the line. Explore the verticality of the line. Have the line ebb and flow. You can even threaten the integrity of the line, with enemies threatening to overwhelm or break through it.

PRACTICE, PRACTICE, PRACTICE

All of this, of course, ultimately boils down to what the characters are actually doing during the battle. It’s easy enough to say that there should be a cool finishing move when a PC kills an NPC or that you should incorporate the environment into your descriptions, but it’s ultimately up to you to look at the barbarian, the sword, the goblin, and the bookcase, and then put those together into something awesome.

And that’s more art than science.

But it IS something that can be practiced.

The most straightforward way to practice, of course, is to simply run more combats: Schedule more sessions and play, play, play, play.

But you can also practice between sessions. One exercise I’ve found useful is to watch a really great fight film — John Wick or 300 or Fellowship of the Ring or Police Story — and simply narrate the action as it happens on the screen, as if you were describing it your gaming group.

This exercise will build your repertoire of fight moves, develop the vocabulary you have for describing a fight, and even teach you a little bit about the effective pacing of fight sequences (although the pacing of an RPG fight will usually be distinctly different from that of a film).

You’ll be able to take those skills and use them to fuel your visualization of the battlefield.

Creative Commons Icons: Goblin by Linector, Barbarian by Delapouite.

Cartography by Dyson Logos.

In the first episode of The Mandalorian, the titular character parks his ship and heads off on an adventure. In the second episode, he returns to find his ship being stripped for parts by opportunistic jawas. This prompts an incredible action/chase scene, followed by an adventure hook which results in some startling revelations.

While keeping in mind the Principle of Using Linear Mediums as RPG Examples, let’s assume that this isn’t just an example of the GM having a cool idea and making it happen. (Nothing wrong with that, obviously.) If we wanted stuff like this in our games, what could we do to make that happen?

Non-focal random encounters.

Random encounters, of course, can be used to achieve several different effects. But one of the ways they can be used is as a procedural content generator, providing a creative prompt to the GM for an interstitial event. Because the “camera” of our game session is virtually always focused on the PCs, we tend to think of the events generated by the random encounter tables as intersecting the path of the party; it’s something that happens randomly in the place where the PCs happen to be.

But it doesn’t have to be!

You can just as easily use procedural content generators to model events happening off-screen.

For example, if the PCs leave mounts and/or henchmen at the entrance of a dungeon while they go delving within, I’ll simply make random encounter checks for the group left behind. A notable example of this occurred when I ran The Sunless Citadel as part of my first D&D 3^rd Edition campaign. The PCs left their mounts up on the surface while they went down into the citadel, I rolled regular random encounter checks, and when they returned they found the horses still there calmly munching grass… surrounded by a dozen scorched goblin corpses.

What the hell had happened?!

Ultimately, what I’m suggesting here is pretty simple:

Roll random encounters for locations/people that aren’t the PCs.

That’s it. That’s the tip.

You can probably usefully generalize this by identifying what the PCs care about and then rolling encounters for those things. This might include people, places, organizations, etc. The rate and nature of these encounters will depend on what and where these things are. The henchmen at the dungeon entrance are easy because you can just roll on the dungeon’s random encounter table (perhaps at a reduced rate if efforts have been made to conceal their camp). But what about the PCs’ favorite tavern? Or their emotionally troubled ward? Or their political patron?

In the most generic version of this, however, you can just create the list of Important Things in the Campaign and then roll encounter checks for everything on the list as part of your session prep. If an encounter is indicated, that simply means that this element of the campaign world has seen some sort of interesting development: What is it? And, importantly, how will the PCs learn of it?

This can be a really easy way to keep a big, complicated campaign world in motion without needing to constantly grapple with the almost impossible enormity of the whole thing. It can also just be a good way of reminding the players that the campaign world does, in fact, continue to exist even when they’re not looking at it.