Players most strongly remember two kinds of fights: the very easy and the very hard. If a fight seems difficult on paper and the players blow clean through it with good rolls, good strategy, and/or a good builds, they’ll tell that story for years if only because it’s hilarious. If a fight pushes them right to the edge, they’ll remember it too, either because they overcame a difficult challenge or because they lost. Many fights fall in between, and that’s fine. Each fight should mean something to the game, but not every fight has to be a life event.
Boss fights are different. They should be memorable, which means they shouldn’t be humdrum, by-the-numbers encounters. The more important a boss fight is, the more it should test the players’ and character’s limits. I trust this concept is not controversial. The problem is doing it. How do you make a fight hard enough to test the party but not so hard you accidentally wipe them out?
For the most part, you have to avoid giving bosses sheer, unfair numerical advantages. A single zombie with 30 AC is an interesting puzzle, because the impact of the change is limited to a single monster with low damage and hit points. An undead master necromancer with 30 AC is a frustrating power play, because it can use that near-immunity to PC damage to wade into combat and sling deadly spells while the players can’t do anything about it. Major stat changes are best done in small doses, and a boss is by necessity a big dose. We’ll talk about changing a boss’ stat block later, but for now we need to look at other avenues.
What we should do is look at how bosses become challenging encounters, and they do that by taking players out of their comfort zone. If the players can leverage the same strategies, abilities, battlefield positioning, and witty banter they always use, a boss is just the fight before a long rest. It’s supposed to be different, which means we need to think about how the players’ gaming instincts and self-preservation strategies can, inadvertently or intentionally, make fights boring, and how we can route around those easy paths to success.
Problem: The simple beatdown. The party’s toughest person walks up to the enemy and locks it down while everybody else does whatever they want. In World of Warcraft they refer to this as a “tank and spank” (or they did when I was playing); as long as the tank is doing her job, everybody else can abuse the boss in whatever way most suits them. The boss doesn’t meaningfully affect anybody except the character best able to handle it, and he doesn’t prevent the players from bringing their most powerful abilities to bear. With clever placement the party can even hit the boss with area-effect spells without touching any allies. This doesn’t challenge the players because they’re not actually doing anything; they’re just standing around and hitting the “win” button. This style of fight encompasses many of the later problems, and it’s probably the least satisfying type of encounter.
Solution: Add minions. Give the boss some allies of his own to harry the party. This is the single most popular piece of advice I see for creating bosses in every edition, and it works. Minions give the boss’s team multiple actions, from multiple places, at multiple times. It lets him stand and face the tank without ignoring everybody else. Minions are most relevant when they’re not just lesser forms of the boss; instead of adding backup sword-and-shield fighters to give your sword-and-shield boss flanking, add archers to give the team some ranged attacks or a healer so the party feels the need to kill the minions first. Minions can be any lower-powered creature thematically linked to the boss, from undead for a necromancer to charmed townspeople for a fey to automated defenses for a kobold trapsmith. Your goal is to split the party’s attention and resources so they aren’t all focused on one place.
Take care not to add too many minions or minions that are too powerful. A fight is about the boss and his team, not a faceless army of 5-hit-point orcs or a surprisingly effective ranger cohort. If your boss is a single creature, the minions all combined should be no tougher than he is without them; this means a CR 5 boss in Pathfinder can have two CR 3 allies or three CR 2 allies, which are EL 5 on their own. If your boss is two creatures, all minions combined should be no more than half as tough as the boss creatures combined; two bearded devils in 5E (1400 xp combined) can have three imp minions (600 xp combined) but not four. Do account for the minions in your difficulty calculations. You are dramatically increasing the fight’s challenge by design and you don’t want to overcompensate.
Problem: Initiative blocks. At the beginning of the fight, everybody rolls initiative. At some point the boss takes his turn. Then the entire party takes their turns in a block before the boss goes again. No matter who rolls what for initiative, this is guaranteed to happen with a single boss creature and it’s a possibility with multiple creatures, especially when the players specifically engineer it. You don’t want the combat to devolve into “I go, then you go, then I go” pacing. It removes tension from the player ’s side by giving them a buffer, letting them know the boss can’t do anything else to them until they’ve all had their say.
Solution: Off-turn actions. The way you mess with initiative varies with edition. In Pathfinder and 4E you can just have a monster delay to break up an initiative block, though that only works if you have multiple creatures in the first place. In 5E you can’t delay at all. But you can give a boss limited legendary actions, letting him take one or two actions from a small list when it’s not his turn. The boss can still be a threat even during the PCs’ initiative block, and if his legendary actions aren’t attacks he can hit them in interesting ways. You can also let him respond to player actions with immediate actions (in 4E), reactions (in 5E), or whatever random ability you want in your heart (in Pathfinder).
I want to make a special note about giving a boss multiple initiative passes, which is a mechanic I learned from Dragon Quest. Here you just give a boss two turns; he either rolls initiative twice or he takes a second turn at his initiative minus ten. It’s not as powerful as doubling the monster count because he still has the same number of hit points and he can still run out of non-consumable powers, but it is powerful. If you do this, consider limiting it in some way, like only letting him take certain actions on his bonus turn or rolling for recharge abilities only on the main turn. The 4E behir is a good example; it has three turns, but it can either move or attack on each turn, not both. This mechanic is also not for monsters with high burst damage because those bursts come too frequently for most parties to handle. Giving a boss multiple initiatives is a simple way to make it interesting, but you have to apply that power wisely because the boss will affect combat much more than any player can.
Problem: Automatic fights. Once everybody figures out their rhythm in a fight, there’s little reason to change it. The fighter does full attacks, the rogue does full attacks from flanking, the cleric heals and does other things when there’s nobody to heal, and the wizard either casts bread-and-butter spells or works his way through his big guns depending on what sort of player he has. Even the boss settles into a routine, swinging twice at the fighter and pinging somebody else with a status effect. Either both sides are keeping pace with each other, which means the fight ultimately comes down to luck, or one side is coming out ahead, which means the fight has a foregone conclusion. Either way, player agency and interest are gone because there’s no further thought involved.
Solution: Dynamic elements. Include some way to mix things up if the fight gets too boring. The most common recommendation I see is some sort of battlefield effect, like dense moving clouds that block line of sight or periodic lightning that leaves scorching ground. Rulebooks call these “terrain”, but that always bring to my mind a static effect like a magic circle or a lava flow. We want exactly the opposite, like a magic circle that summons minions unless the players stop it or a lava flow from which gouts of poisonous gas erupt. If you don’t want unnecessary complexity in your fight by default, provide the interesting terrain but don’t have it do anything right away. If the fight gets boring you can suddenly have it take effect, changing the encounter by adding a new (but foreshadowed) issue for the party to deal with.
Again, you don’t want the dynamic elements to be a bigger deal than the boss, and you don’t want them to lock combat down. Avoid dynamic elements that create permanent effects unless those effects are low-impact; a permanent zone of difficult terrain is fine, but a permanent zone of 4d8 lightning damage is usually not. It’s safer to have the effects fade or move slowly enough that players can get out of the way. If they deal damage, don’t make them any stronger than one of the boss’ attacks. And if you can manage it, try adding an element of randomness to them. The players will panic when they see meteorites falling from the ceiling, but if one of those meteorites hits the boss, they’ll love it forever. You can probably get away with using a stronger dynamic effect if it’s as likely to hurt the boss as hurt a PC, especially if the players can manipulate the boss into it.
The boss can also have dynamic elements built-in. He could summon his minions on certain rounds instead of having them all at the beginning, or he could have an ability that lets him teleport when he reaches certain hit point thresholds. This sort of ability shouldn’t consume his actions for a turn, and it lets the players put him in a bad situation for a while but doesn’t let them keep him there indefinitely. Multi-round actions are also good for this; the boss ends his turn by glowing blue, then starts his next turn by dealing cold damage to everybody nearby before he does anything else. The next time he glows blue, the PCs have to either get out of the area of effect, find a way to mitigate the damage, or hope the healers can keep up. Either way, they’ll react to a change in the state of the battle, and a change is what you want.
Giving a boss multiple phases also counts as a dynamic element, but I want to talk about that in a dedicated post.
Problem: Combat-stopping powers. Any ability that takes a PC out of combat for a while is a potential problem. It essentially removes them from the game until the ability ends, and they can only sit and watch unless you give them something else to do. Similarly, any ability that takes a boss out of the game is a major threat to the challenge that boss poses, especially when the boss is the only creature on the enemy team. A stunned or paralyzed boss can do absolutely nothing to affect the fight, and even his off-turn actions are on lockdown. He has to simply stand there while the PCs are wailing on him, casting buffs, or changing the battlefield to suit them. Nothing takes the wind out of an otherwise scary fight like making him spend a turn or two as a piñata.
Solution: Cheating. Technically, there’s nothing a DM can do from a monster design perspective that’s cheating. But we generally assume a boss’ stat block at the beginning of a fight is the same as his stat block at the end of it. That doesn’t have to be true. You can decide, spontaneously in the middle of combat, to give a boss the legendary resistance ability and let him automatically succeed on a saving throws a few times per day. If you don’t like that, you can let him shrug off the effect at a cost, like spending hit points to de-escalate from stunned to merely slowed. Some players are fine with steamrolling a boss because he failed one or two important saves, but if your group isn’t, you don’t have to let it happen. You can decide that encounter pacing and satisfaction are more important than a single natural 1.
But you do have to limit yourself, hence the “X times per day” or “costs Y hit points” tradeoff. A player spent resources to use their ability and they succeeded on it, so they should get some benefit, even if it’s not the one they expected. Never make them feel like they wasted their turn. If you’re not sure what to do, watch how the table reacts when they stop the boss in his tracks. If they cheer and kick back their feet, they think they’re coasting and you can afford to up the difficulty. But if they take a deep breath and start talking about how this gives them a chance to escape, they’re already invested and they feel they need this to succeed. You don’t have to take it away from them.
If you don’t like the idea of cheating mid-battle to make bosses tougher, that’s fine. Don’t do it. Instead, be up front with your players. If you come into Session Zero telling your players that certain bosses may be resistant or immune to certain conditions, everything’s out in the open before someone builds their whole character around paralyzation effects. If you’re a couple of months into the campaign and your PCs suddenly break out stun-locking tech, explain why that’s a problem and tell them future bosses will have ways to handle it. On-the-fly cheating is a right-now solution for a problem you didn’t foresee; communication is a long-term solution once you know what’s happening.
Problem: The hot streak. This happens to everybody eventually, on both sides. There’s nothing you can do to prevent a player from rolling critical hits on three consecutive attacks or from rolling all sixes on her fireball. In short order the fight went from a well-balanced, dangerous encounter almost all the way to a comedy, and it’s nobody’s fault. It’s just the result of using a dice-based system.
Solution: All of the above. Not at once, of course. But when the party is, through no fault of their own, about to trivialize what should be a gripping encounter, you can react in pretty much any way you choose. Create minions who come to their master’s aid, especially if it doesn’t consume his action to summon them. Use terrain or automatic powers to force the PCs to back off and reassess their situation. Change the boss’s stat block when he hits a milestone, like losing half his health, to make him tougher or faster or stronger or whatever you want. All of these preserve the damage the players did (unless the boss heals to full, which is a nasty trick to pull) while keeping the fight going.
Few fights need more than one or two of these solutions, and a lot of them overlap and solve multiple problems. For example, minions can keep PCs from focus-firing on the boss, take their turns between PCs’ initiatives, and affect the battle even if the boss is stunned. You don’t want every boss to become a chaotic mess just to prevent problems that may not occur. Before you do anything, see how your groups works and how they handle important fights. If you find you’re having a problem making bosses challenging and memorable, switch things up.
The point of this is not to punish the players for success by subverting their strategies, builds, or luck. It’s intended as the opposite; the PCs are doing so well that they need new threats to keep them interested. You’re not challenging the characters by making things numerically unfair, you’re challenging the players by acknowledging that a stand-up fight is too simple for them and giving them a reason to switch things up. The key is to come at it from the standpoint of providing the best possible play experience. If you’re just trying to up the difficulty to prove you can, everybody’s losing regardless of what happens on the battlemat.
And if you do want to challenge the characters by giving your boss better-than-average numbers without killing everybody, we’ll talk about that next.