Tribal Knowledge: Sculpting a Boss Fight

I’m taking a week off from the history of the Ranger class to talk about encounter design, an under-appreciated art within the broader work of game-running. Tabletop games, like video games, have long had a concept of a “boss fight,” a climactic end to an adventure or series of adventures. Starting from that similar position, though, video games have done far more creative things with the concept, while tabletop games have scarcely advanced beyond what MMO players would call “tank & spank.” There are exceptions, and I’ll be noting the ones I can think of, but in general I’m interested in expanding the range of options people consider when they design climactic encounters.

The simplest phrasing of this that I have heard is “The point of a challenge is to make you uncomfortable.” Breaking that down a little more, the goal of climactic combat design might be, “Disrupt the PCs’ optimal combat tactics in at least one significant way.” This highlights two things: what makes a challenging, exciting encounter for one party may be impossible for another group of the same level, and trivial for a third. (Correspondingly, take my ideas below with a grain of salt.) Secondly, it helps us brainstorm by breaking down the PCs’ approach.

As a basis for discussion, let’s assume you have a classic D&D four-man band, or close enough as makes no difference: a fighter, a rogue, a cleric, and a wizard. Actually, using 4e’s combat role titles will make this more generally applicable; whatever you may think of 4e, the most iconic party composition in every edition of D&D does include a defender, a striker, a leader, and a controller. (5e definitely doesn’t break if you’re missing one or more of those roles.)

In general, the players want the defender to close with as many of the bad guys as possible, as quickly as possible, and stop them from getting to the less-armored party members. If the striker is melee-based, they want the striker to flank the enemies and deal damage, disengaging when it looks like someone is going to go house on them. If the striker is ranged, you want her to make sure she’s getting clear shots. In either case, everyone who is doing damage at all wants to focus fire on one target at a time, though if the controller and others splash damage on secondary targets, it’s all good (just watch out for excessively close air support). Most leaders are melee types, but with some number of good options for ranged attacks; I think 4e and 5e are the only editions with substantial ranged healing (other than really-expensive-for-what-you-get area healing effects in 3.x), so most healing output requires getting perilously close whatever the melee types are fighting, and possibly casting spells while threatened. The leader may do some off-tanking, keeping opponents tied up while everyone else chews up and spits out other enemies. Usually the leader has sufficient armor, if not quite sufficient hit points, to do this for a good while. There may be debuffs or other conditions that party members, and especially controllers, can throw out to make things easier on everyone; slowing down the bad guys trying to mulch the defender is a good idea. Keep enemies off of the controller at all costs, and things should turn out okay in the end.

By itself, there’s nothing wrong with this structure. It has a job for everyone in the party, and it feels pretty good when it works right, because it’s designed to dismantle the opposition’s plan and put them in the ground with a minimum of fuss. What happens when you’re forced to ditch this plan and come up with something new on the fly, though?


Defender Closes with the Bad Guys

The fighter-types probably can use a bow or other ranged weapon, but if they’re the party’s meat shields, it’s melee all the way*. If you’re going to mess with this part of the plan, make sure the defender still has a good way to contribute and have fun, because an enemy that just makes the defender useless is not a good time. If you’re playing the defender, put some tricks up your sleeve so that you can contribute even against teleporting or phasing bad guys.

*: This is not the time to mention your Warlock 2/Sorcerer 2/Paladin X that is a better defender than any fighter could ever be.

  • Immunity to weapon damage, or to the defender’s weapons at least, certainly puts the defender out of commission as a threat to the bad guy. The DM might do this, though, if there’s a situation where the spellcasters need to lock down the bad guy with stuns, hold person spells, or whatever, while everyone else works on disabling the Widget that makes the bad guy(s) immune to damage. Letting someone else play the defender for a fight while the normal defender solves a skill challenge or breaks a Mystic Orb of Noonah is good times. The bad guy here should probably be immune to all damage while the protection lasts, not just weapon damage (but vulnerable, maybe even with a saving throw penalty, against debilitating status effects).
    • For the entirety of this discussion, “immune to damage” here mean different things: actual immunity, functional immunity through massive regeneration, or just “always gets back up after getting killed.” What matters is that conventional approaches to killing the bad guy are infeasible.
    • On a related note, I am a big fan of bad guys, especially top-end bad guys and most types of undead, having a True Death – a particular condition that is the only way they stay dead. If villains have True Death conditions that are not widely known, it drives research and, ideally, gets players to engage with the lore, while also possibly changing how they approach conversational encounters (since there’s little point to escalating the encounter to a fight).
  • Hard-to-Avoid Area Attack, or: Let the Rogue “Tank” This One. So you’re fighting a blind opponent that has a highly damaging attack such as a breath weapon. It targets places where it has been attacked and places where it has heard movement. The goal, then, is for someone with Bluff, Deception, or maybe justghost sound to misdirect the opponent while the rest of the party deals damage, sneaks past, or whatever their goal might be.
    • Setting the goal of the encounter as anything other than “kill the boss” is good, though for this article I’m mostly concerning myself with killing the boss in an unusual way. Changing the players’ or boss’s goal to something other than the destruction of the other is actually the easy way to shake up a battle.
  • Cycle the Defenders: In 5e terms, give the boss a combination chill touch (preventing the target from recovering hit points for one round) and stacking damage-over-time effect, such that the damage-over-time is only stopped by receiving a healing effect. To make your life a little easier, make the chill touch effect into an aura. This should force the melee types to step back long enough to shake off the chill touch and receive some healing. Might be a good time for the melee strikers to look up the Dodge action.
  • Flying Lessons: The boss is strong enough to pick up and throw, shove, or otherwise manhandle anyone who closes into melee, and there’s a readily accessible pit full of minions (it’s probably snakes) for him to throw them into. When no one is up in his face, the boss throws boulders at any ranged attackers, because there are firing positions that are out of his reach. This structure really needs to give the melee types something else to do, because standing back and using their ranged weapons can be pretty dull for characters who are just not good at that.
    • It helps a lot if the out-of-reach firing positions have cover points that the boss’s boulders or other ranged attacks can destroy, forcing the ranged types to keep moving.
    • Maybe there are support beams down in the snakepit that the melee types can destroy, so as to drop the boss into a pit that is one Sarlacc deep?


The Striker Moves to Flank

The directive to outflank an enemy whenever you can applies to everyone in the melee, of course – even if the system doesn’t grant a bonus to attack rolls, you’re in better position to cut off escapes. This is also about focusing fire on opponents, because most editions of D&D, and indeed most tabletop games, avoid death-spiral mechanics; thus an opponent fights at full strength until the moment he keels over. (Mechanics like “bloodied” in 4e could have worked around this one way or another, but that was relatively rare.)

In fact, 3.x and later editions of D&D go much further in encouraging PCs to focus fire, such as regeneration, Sneak Attack, and any marking mechanic (common in 4e, optional in 5e). 5e’s Concentration rules and 4e’s non-stacking damage-over-time rules are notable countervailing elements.

  • Making flanking much more difficult, or outright impossible, is okay if approached with care. For example, how about a giant spider fighting from the ceiling, which is low enough that the PCs can attack in melee… but the floor is not solid, and there are only a few jumpy-stones where melee characters can stand to make their attacks?
    • You’re potentially taking away the Rogue’s primary damage mechanic, so whatever you’re including for the rogue to fight that isn’tthe spider boss, it should probably be retuned to not just eat the Rogue alive. (Less necessary if we’re talking about Barbarians, say.) It’s also rough on high-level melee Rangers, who may rely on Whirlwind Attack for much of their damage output.
    • This kind of terrain feature is also good for rewarding Athletics and Acrobatics proficiency, but a roll every time a PC moves may grow tedious.
  • Lifeleech Aura: A boss could have some kind of necrotic aura that deals damage to creatures within 5 feet, and heals the boss for the amount of damage dealt. In case that isn’t enough, base the amount of damage per target on the number of enemies within the aura. This is absolutely brutal on rogues; I hope they have some way other than flanking to gain their Sneak Attack damage. Especially if it’s a melee-heavy party, make sure the boss has some minions around so that PCs who aren’t swarming the boss have something to do.
  • Triggered Effects: If the boss has particularly nasty abilities that trigger the first time it falls below 50% of its hit points, and again when it dies, the PCs are likely to want to spread those two events out a bit more, so that the leader has time to pick people back up, remove status effects, or the like.


Clear Shots for the Ranged PCs

Ranged strikers and controllers mostly want to pick one spot where they can see the whole fight, but they are completely insulated from it aside from an arrow slit. A really small one. From this position they rain death and mayhem upon the DM’s glorious armies, et cetera. The DM’s goal, then, is to force them to move, because moving (especially moving away from cover) exposes them to risk. Something to think about with their shots other than murdering the boss is good too.

Look, everyone here has played at least one video game shooter, right? I don’t need to explain about the shooting gallery of balloons or whatever that drop cool power-ups or weaken the boss for whatever reason? Okay, good.

  • A boss that let the PC try a much harder shot for a big damage bonus is something I don’t think I’ve seen, other than the inadvisable Called Shot rules of 2e. In 5e, it feeds nicely into the Archery Fighting Style of the fighter and ranger granting an attack bonus, and the higher-level warlock’s eldritch blast making a lot of separate attacks.
  • Giving the boss a large piece of cover, a fog bank, or the like that he can move behind (relative to the archer types), forcing the melee characters to follow him, is not bad in itself. It’s like body pulling in an MMO, but because the boss doesn’t operate on an aggro table, he can use those tactics on the PCs.
  • A small skirmishing party that can circle around and run down the archers worked in Braveheart, and there’s no reason it shouldn’t work for the boss. NPCs that were not revealed at the start of the battle is a great way to raise the tension, in any case. It’s also a great counter to focused fire – the back ranks often must handle the new threat immediately, or they will be dead before they can finish off the boss.
  • This is as good a place to mention this as any: I dimly recall reading once that no Chinese battle plan was complete if it didn’t include fire or water in some terrain-altering way. I have no idea what that was from anymore, but it stuck in my head. Anyway, this is D&D, so we can do two elements better.
    • Ceilings collapse when subjected to the tremors of the titanic battle going on nearby… especially if the bad guy has weakened the structure to make sure they’ll fall far from where he plans to fight.
    • A drifting cloud of deadly neurotoxin that the boss keeps well away from herself by controlling the winds is good fun. Making her immune while the PCs struggle to hold their breaths is great, but much higher stakes.
    • Other than arctic or underwater scenarios, I’m not sure there’s ever been a bad time to raise the tension by lighting some part of the terrain on fire. Even if this is more bad than good for the boss, it’s just about always interesting.
    • A valley or chamber that floods with water, sand, or lava round-by-round is also a good move, as the DM can probably set up the field to create an ever greater divide between the melee and ranged characters (each scrambling to find higher ground).


The Controllers Hurl Spells

Editions of D&D that are not 4th don’t share its clear delineation about what is a controller and what is a spellcasting ranged striker. That’s fine. For the sake of argument, understand that this section and the one preceding it might apply to all spellcasters, but this section applies much less to archers, slingers, and gunslingers.

In combination with focused fire, there’s a substantial incentive for both sides of every fight to use their biggest attacks first (“go nova”) – especially the NPCs, who probably do not have long to live, so if they want to be memorable or impressive, they had better get it out of their systems in the first two or three rounds. PCs at least have the possibility of more encounters later in the day… but this is a post about boss fights. The ones at the end of an adventure.

I’m intrigued to see a few mechanics in 5e discouraging PCs from going nova as soon as they see the boss: legendary resistance. A certain number of times per day, some top-end probably-a-boss monsters can turn a failed save into a successful one instead. While seemingly heavy-handed in execution, this is a pretty clear reaction against the degree of saving throw penalty that could be levied against solos in 4e, trivializing even climactic encounters. To a much lesser extent, counterspell also discourages going nova, though that’s costly enough to be closer to a wash.

  • Spell Turning is an effect found (as far as I know) only a certain ring, thus far in 5e, but giving the boss an improved form of this ability as a special ability might be interesting. I’d be more inclined to use the mechanics of earlier editions, though: incoming spells are automatically reflected, up to a certain number of spell levels. This encourages the party’s spellcasters to spend a few rounds wearing down the effect (ideally, with magic missiles that fizzle out on their shields) before they throw the bigger effects. I might have cantrips not interact with this effect at all. (Be aware: this screws warlocks over like whoa, so use it with utmost caution if there’s a warlock in your group.)
    • Honestly, spells are complicated enough that figuring out smart ways to apply them is already pretty compelling gameplay, and boss fights only have so much to add to that. Giving spellcasters an idiosyncratic magical effect in the area to worry about or resolve would be pretty cool.
  • Along those lines, maybe a local environmental effect is generating damaging spell-like effects (lightning bolt springs to mind) that the spellcasters can counter, thus letting them defend the defenders (or whoever). Reveal this battleground feature before the PCs’ last long rest before said battle.
  • Nothing goes great with fireballs like chain-reaction explosions. Swamp gas, barrels of lamp oil, and absolutely any kind of vehicle (what, you didn’t know that chariots explode?) can be counted on to make fireballs dangerous to far more people than the caster originally intended. There’s a huge potential for this to just be the DM screwing over the evoker, so use this with consideration.


The Leader Heals or Buffs Someone

Oddly, this is one of the hardest ones for me to do anything interesting with, but that’s okay, because leaders already have to worry about moving with the melee combatants and/or hanging back with the spellcasters. I’m not seeing that many interesting ways for the boss or the terrain to influence one PC casting a spell on another, other than the chill touch thing I mentioned earlier. Frankly, that was my experience of playing a priest in WoW too – a lot of times I didn’t really watch the fight except to avoid standing in fire (unsuccessfully, let’s be honest), and just try to keep those health bars topped up. D&D is substantially different, in that you won’t cast a healing spell in every round of combat until high-level play, if even then.

Increasing the healing demand on the leader classes is not a great direction to take things – they’re casting out of spell slots they could be using to feel directly effective in the battle.

  • In terms of altering their primary functions, consider a special terrain effect that significantly boosts healing word and similar ranged healing effects, encouraging the leader to go further from the melee… and risk getting cut off or swarmed, but that comes down to risk/reward tuning.
  • Either for one battle or for the whole campaign, bring back bloodied effects – something that changes in a PC’s stats when they fall below 50% of their maximum hit points. As long as this change is a net gain, it encourages leaders to hold off on healing until the target really needs it, again increasing risk in exchange for a reward. (If you’re still playing 4e, then… keep doin’ what you’re doin’, I guess.)

The most important thing I can say to conclude this: don’t negate someone’s gameplay and stop there. If you close one door, you are obligated to open another. Players also shouldn’t make their characters one-trick ponies – if your class makes both ranged and melee attacks an option, be prepared to use both. Let your character be practical about it.

Okay, that’s what I’ve got. See that Comments field below? It would be totally cool if you added something to it that constituted an awesome boss fight structure, adapted to D&D. People are going to have all kinds of different tastes as regards mechanics becoming intrusive on the epic showdown with the boss, and if all of this sounds terrible and game-y to you when what you want is a pure narrative experience, that’s cool too: but talk about how you set climactic encounters apart. If not mechanically, then what narrative tricks do you use to engage all of the players in the intensity of facing their nemesis? (Nemeses. Nemesises. Whatever.)



More Tribality Articles You Might Enjoy


Brandes Stoddard enjoys games of many kinds: video, tabletop, board, card, and live-action games. He runs Dust to Dust, a fantasy LARP in Georgia, and works in freelance game design and writing. He blogs about games at