Creating Bosses: Actual Numbers

This post is the reason I’ve been avoiding writing about boss monsters for so long. While I’ve run hundreds (literally hundreds) of sessions in 3E/Pathfinder and 4E, my 5E experience is much more limited. I wasn’t sure I had the experience with 5E to offer such blanket advice. But now that I’ve delved deeper into it and done some research, 5E isn’t that imposing. It doesn’t work like other editions, but that’s why we need guidelines for it.

The trick about balancing boss encounters in 5E is to use some of the things about higher-Challenge creatures without using the creatures themselves. As we’ll see, the way 5E handles the stats that scale with Challenge means it obliterates parties who fight higher-Challenge monsters, and the experience point values are really more useful measures than the Challenges themselves. For example, a Challenge 1 creature is not for L1 parties. By the experience point budget, it’s a deadly encounter not suitable for everyday use. It’s a bit of a rude awakening for DMs used to earlier editions where 1 = 1. Boss monsters aren’t just high-Challenge creatures, and they aren’t high-Challenge versions of low-Challenge creatures. Instead, boss monsters are monsters who imitate some, but not all, of the traits of higher-Challenge creatures. We need to figure out which ones to take and how to apply them.

Let’s start with one thing outright: we can’t change the numbers we can’t change. This sounds tautological but it’s also important. 5E has certain numbers on lockdown; saving throws, attack bonuses, and ability DCs are always based on a creature’s proficiency bonus and its ability score modifier. There’s no capacity for “this creature gets an arbitrary +2 to Dexterity saving throws”. It can have advantage and/or proficiency but nothing else. What we can do is change a creature’s ability scores, and we can do that fairly easily, but we should only do it with purpose.

One of the explicit design goals and selling points of 5E is that numbers, by and large, don’t scale with level. Recall that a character is only guaranteed to get +1 to attack rolls every four levels. Everything else has to come from ability scores (which have a hard cap), magic items (which exist only at the pleasure of the DM), and a thimbleful of class powers. It’s perfectly reasonable to have a fighter whose attack bonus is +6 (18 Strength, +1 proficiency) at L1 and only +10 at L10 (20 Strength, +4 proficiency, +1 magic). Monster AC is built on this:

AC goes up with Challenge, but it’s not steady. There’s no hard-and-fast rule that associates AC with Challenge, just a general trend. The average increase from level to level is 0.389 AC per Challenge, which means increasing a creature’s AC by 1 is roughly the same as giving it the AC of creature whose Challenge is 3 higher. Given that, we can safely increase a boss’ AC by 1 or 2; any more than that and we have to compensate somewhere else. This does vary by the creature. The lower the AC your creature has, the more you can increase without blowing them out of the water, but a high-AC creature should stay right where it is (or even lose AC, if you want to give the players a different sort of challenge).

So our basic number tweaking doesn’t work. We don’t have a lot of room to change AC, or attack bonus, or saving throws, or most of a creature’s round-by-round numbers. That’s because in 5E, level really only affects three things: hit points, damage, and ability power. High-level play is about monsters and characters who hit hard and take hard hits:

Ignoring Challenge 18*, there’s a clear growth in hit points as Challenge increases, and that growth looks like it ramps up as you hit epic levels. But when you consider the actual numbers, HP growth is faster at lower levels than higher levels. In general, increasing a creature’s hit points by 15%, or 25% at early levels, is roughly the same as giving it the hit points of a creature whose Challenge is 1 higher. We can increase monster health by 30 to 50% without going too far afield. That’s because hit points don’t really affect what a monster does, they affect how long a monster does it. The more hit points it has, the more rounds it’s likely to live and the more chances it has to hurt the party:

This chart takes into account each monster’s primary and secondary attacks, including bonus effects like poison damage that only take effect some of the time. It does not take into account spellcasting; that’s too variable to quantify.

On average, even a Challenge 20 creature only hits for nineteen damage. The only attacks I could find that dealt more than fifty damage were the purple worm’s tail sting, which does a bonus 15d6 damage on a failed saving throw (and is part of that Challenge 15 spike), the mind flayer’s extract brain, which only works on targets it’s already grappled for a turn, and the remorhaz’s bite, which is its only attack. And that’s really the thing; high-level monsters don’t get big attacks as much as they get more attacks and more riders on those attacks. Level-by-level damage growth is less than ten percent, which isn’t even enough to consider. Increasing single attack damage is not a large factor in Challenge. This doesn’t mean you can triple a monster’s damage without a care. In fact, it means the opposite; the game balance isn’t equipped to handle changes like that. The most damage you can safely add is a bonus damage die. Each attack looks scarier, and they can add up, but they’re unlikely to make the difference between an easy fight and a total party kill. You can usually get more impact by adding an extra attack, especially if that attack gives the monster a new targeting option. For example, a creature whose Multiattack lets them makes two melee attacks and one ranged attack can threaten much more of the battlefield, but not enough that it invalidates a tank.

That leaves us with the power of the creature’s abilities. There’s no measuring stick for that; we can’t say one ability is high-level and another is low-level without exhaustively rating each of them. But we can judge universal abilities like conditions. The more a condition limits a character, the more powerful it is; a penalty to attack rolls is just not as impactful as the inability to make attack rolls at all. We can also judge how powerful certain non-universal abilities are by how they affect the characters:

In general, anything that takes a PC out of combat (or nearly does) is major and anything worse that than is devastating. Players make more attack rolls than skill checks or saving throws, so a penalty to attack rolls is a bigger deal than a penalty to other types of rolls. Disadvantage on one type of roll is mathematically equivalent to a -3.375 penalty on that roll**, so a penalty of -3 or -4 to AC is about the same as giving all monsters advantage on attacks. The weight of bonus damage is dependent on the number of attacks a creature makes, so consider how the damage affects the creature’s total per-round damage assuming all attacks hit. And the danger of recurring damage is obviously dependent on the amount of damage, but it’s usually fairly small and it’s easier for a healer to deal with it than burst damage so it’s not as bad as other options.

These don’t map easily to Challenge, but from how they affect encounter pacing and a character’s ability to react to enemies, I’d say that if you escalate a creature’s abilities from a value in one column to a value in the next column, that’s consistent with increasing Challenge by 1 or 2. Try to stay within thematic lines; it often doesn’t make sense to take a big, burly monster who knocks enemies prone and instead have it poison targets.

Now you can think about whether you want to change ability scores to add a +1 or +2 to a monster’s attack bonus, saving throws, or ability DCs. Again, remember that a PC’s ability to handle changes is limited. Adding +2 to an ability score is often safe, and adding +4 is acceptable for major bosses, but anything more than that is too much. There’s no graph here because changes like this usually aren’t fun. They mostly affect whether you say “yes” or “no” to a player, nothing else, and players aren’t likely to notice the difference. Honestly, a change like this is more to make the DM feel better than it is for encounter balance. And remember that some ability scores affect multiple values. If you’re increasing a monster’s Dexterity by 4, you probably don’t also want to increase its AC by 2. The former usually includes the latter.

Finally, if you look at these numbers and you don’t like them, do what feels right. As I said before, you’re the DM. You know what your group can handle. No online advice is as good as knowing what you and your group want from the game.

Let’s look at an example. Say your players are a group of industrialists looking to despoil a forest, because in my experience players love being the bad guy, and say they’re just starting out at second level. You want to give them a nature-themed antagonist, and you want it to be around Challenge 3 so it’s a big problem for your L2 party. You could just get a Challenge 3 creature, but those things are numerically scary and conceptually boring. You want something special and unique but not as party-killing as a standard Challenge 3, so let’s instead take a dryad (Challenge 1) and scale it up.

The dryad’s AC is 11 when she’s surprised and 16 after she casts a spell; 16 is already 2 higher than the average Challenge 3 creature, so you can leave that alone unless you plan to switch out her spells, in which case you can bump it up to 14 to represent natural armor. She has 22 hit points, and we can increase that to 36 safely. You can give her three more Hit Dice (bringing her to 8d8, or 36 exactly) or give her one Hit Die and increase her Constitution by 1 (6d8 + 6, or 33) to also buff her Constitution saving throw.

Her attack has +2 to hit and deals 1d4 damage. That’s terrible. But if she casts another spell it becomes +6 to hit and deals 1d8 + 4 damage. That’s enough to wound almost anybody and kill a low-health PC, so it’s probably fine unless you want to add 1d4 poison damage. But you could let use use her fey charm ability as part of a Multiattack. That adds the possibility of charming enemies while she’s attacking, and charmed is a minor condition we can add without tipping the battle too much, especially since she can only charm one PC at a time. Now she can fight the barbarian but also convince the wizard to come to her aid, involving and threatening multiple PCs without doing tremendous damage. It’s especially great if she attacks the party with charmed wolves early in the adventure, foreshadowing her charm ability but escaping before the party can fight her directly. Then the party isn’t surprised when she starts using it against them, and they can even prepare for it if they’re clever (and have the spells or abilities to do it).

Compare this to an actual Challenge 3 nature guardian, like a veteran who might be a local woodsman. Our dryad has far fewer hit points (36 vs. 58), slightly lower AC (16 vs. 17) and much, much less average damage (8 vs. 20). But remember, you don’t want a Challenge 3 creature. You want a scarier version of a Challenge 1 creature. Your dryad won’t last as long or hit as hard as a veteran, but she also won’t kill any non-tank player in one turn, she isn’t limited to melee, and she has more interesting powers. A veteran is numerically superior, so it’s a deadly fight. Your dryad is powerful and interesting and forces the players to think on their feet, so she’s a boss.

The numbers above are consistent with existing monsters and roughly on par with the advice in the Dungeon Master’s Guide, but they’re not absolute rules. You can give a monster +4 AC, or triple damage, or a petrification aura. That’s your prerogative. But for everything you do like that, have a justification for it from an at-the-table perspective. The goal of the game is to have fun, so know why you think huge numbers make the game fun for the players before you use them. You (probably) didn’t read this far because you want a justification for that kobold tank with AC 24. You read this far because you want to make a boss fight that doesn’t sacrifice the game at the altar of your system mastery.

Or you did it so you knew enough to yell at me meaningfully in the comments. And I have to applaud that level of effort, I suppose.

* — The numerical information in this post is weighted toward lower levels. High-level monsters suffer from the curse of small sample sets; the Monster Manual has fifty-eight Challenge 2 creatures, but only three Challenge 12 creatures and one creature each at Challenge 18 and 19. This isn’t unique to 5E. Low-level play is more popular than high-level play, and it’s easier to throw ten low-level creatures at a high-level party than throw one-tenth of a high-level creature at a low-level party. The math still seems to work out, but if you’re wondering why monster hit points drop precipitously at Challenge 18, that’s not a quirk of monster design. It’s because the only data point we have there is a demilich, one of the smallest creatures in the game and a monster with a host of immunities and resistances to increase its longevity beyond what its hit points might indicate.

** — Similarly, advantage is equivalent to a +3.375 bonus on the roll, in case you ever wondered just how much advantage or disadvantage actually matters.