More than a few DMs struggle with making boss fights that serve as more than a speed bump on the way to a ton of XP. They want to challenge their players, but they don’t want the campaign to end in a TPK. What ends up happening is they repeatedly think this spell or that ability is too much for the party to handle so they avoid it during the session ultimately turning a CR 15 monster into a CR 3 monster. After the session, they say something like “This should have been a deadly encounter, but my players defeated it in one round.” I did this a lot, and it’s only by converting and creating lots of monsters that I’ve come to understand how CR works and how easy it is to make a powerful monster useless by changing one simple ability.
CR is Fragile. To challenge your players you have to understand how CR is calculated. If you don’t have a basic knowledge of it, I suggest you read the section on creating monsters in the DMG. I’ve also written a post elaborating how to use CR that might be helpful to read as well. To reiterate, CR is essentially the average of a monster’s defensive challenge rating and its offensive challenge rating. Some monsters have a huge gap between the two, such as spellcasters. A spellcaster typically has a high offensive CR and a low defensive CR. Lets take the archmage for example. It has a CR of 12, even though it only has an AC of 12 and 99 hit points. To play this monster at a CR of 12 you need to make sure that it has mind blank, stoneskin, and mage armor up before combat. Without its spells (mind blank doesn’t affect CR but it lasts 24 hours so any self-respecting archmage would keep it active) , the archmage’s defensive CR is 2, but with them it goes up to 4. You see, most of an archmage’s CR comes from its spells, and this is even more true about its offensive spells. If you don’t play a monster optimally, it could be half the CR.
To Tweak Or Not To Tweak. A lot of DMs like to take a stat block and tweak it but they don’t recalculate the CR after doing so. For instance, if you look at the archmage’s 9th level spell time stop, it might be tempting to switch it out for something that deals damage, or to use it to cast a 9th level cone of cold. However, time stop allows the archmage to take 1d4+1 (an average of 3) turns after its current turn. If you’re switching it out for meteor swarm you’re going to have nearly double the damage output. Now if you were making your own monster this would increase the CR by a few, but the archmage actually has a CR of 9 with the spells it has if used to maximum effect. If you swap time stop for meteor swarm the CR goes up to 12. My bet is during playtesting someone swapped that spell and they recalculated the CR, since even if you pretend the archmage has enough action economy to cast mirror image, fly, fire shield and globe of invulnerability in addition to stoneskin and mage armor, it still doesn’t bump the CR up to 12, but I digress. Let’s say you have second thoughts during the game and decide the archmage won’t cast anything above 5th level spells because the party is level 10 and they don’t have anything beyond 5th level spells and it wouldn’t be fair. You’ve now reduced the archmage’s CR to 5 which is no challenge for a level 10 party at all. The point is, if you want to tweak something and you don’t want to worry about CR, don’t be surprised when what you end up with is not at all what you expected. At the very least while trying to calculate CR you’ll get a feeling of what a monster needs to do on the first three turns to have the biggest impact.
Making a Boss Monster. Now that you know how to challenge your party, how do you make a monster into a boss? Well you also need to know how your party handles a creature being played at it’s full CR. If you’re uncomfortable in deciding what is a deadly encounter (and keep in mind a boss encounter should be deadly, not hard) and what is an impossible encounter, have a session where you have the party fight a couple random monsters and have it be non-canon so if they die it doesn’t matter. Don’t hold back. You might be surprised at what they can handle. If played right, a deadly encounter can definitely function as a boss encounter without necessarily adding legendary actions and extra hit points. I blame myself for when my players steamrolled a fight I thought was difficult because either it was because I held back and didn’t play optimally to avoid a TPK or it was because I gave them too many magic items at a low level when I was new to 5e and didn’t realize how easy they made encounters. The biggest piece of advice I can give you is to not be afraid to TPK your party with a boss fight. They should know it’s going to be a difficult fight and should plan accordingly, and if it isn’t it’s going to be anticlimactic. If they die they’ll still have a memorable battle, but if it’s a cakewalk they’ll be disappointed. I’ve already written about how to make lower CR monsters a challenge and how to add class levels to a monster to increase its CR. You can do all these things to enhance the difficulty of an encounter but be mindful of how your party would handle the base creature particularly if the base creature has a higher CR than their level.
Legendary Monsters. I find that making a monster legendary often isn’t enough to make it a boss monster. I find that the overall CR has more to do with it than whether or not the monster is legendary. If you take a low CR monster and make it legendary and now it has a CR of 6, it’s probably going to be no more powerful than another CR 6 monster without legendary actions, though it will probably be less susceptible to getting whooped in between turns while it can’t do anything. There’s a better solution to that problem, however, and that’s to add extra monsters as minions during the fight. One monster is almost never a fair fight for a party unless it has some ridiculously high CR in which case the party almost has to kill it in a round or two because otherwise it’s a definite TPK. If a party is going to face a boss alone, it should be because they were able to get it alone somehow through trickery or by quietly taking out its minions, not because the DM was afraid to take off the training wheels. Or if the party is going to face a red dragon and you want it to be a solo fight, consider having the party have to go through a gauntlet of kobold traps to get to the main lair of the dragon. They’ll still be able to focus fire on the creature, but they’ll be short a few hit points and spell slots. If they try to rest have the kobolds flood the tunnel with lava.
Extra Hit Points. Giving a monster extra hit points might help if you gave out too many magical items (or if you like giving out lots of magical items), but never do it simply to make a boss fight harder without changing anything else. I recommend using the maximum hit points for a creature’s hit dice if you’re making it a boss, but don’t double or triple the hit points just to stretch out the fight. This doesn’t make the fight harder, just more tedious. There need to be other things to challenge the players, and simply increasing hit points by some multiplier is just going to turn what could be an epic boss fight into four or five people bashing a piñata for several hours to get the candy inside (in this case, XP and loot).
Non-Combat Bosses. A boss encounter is usually a difficult encounter that must be overcome to achieve a goal. The boss is not always a villain to defeat it can also be a king who needs to be persuaded to not attack a neighboring city-state, or a sphinx with a riddle that needs to be answered before the party can enter the tomb. Consider non-combat options especially if you’re running a campaign that has been mostly combat up to this point.
What to Do if the Boss Still Sucks. If you’ve tried your best but you just haven’t nailed down the appropriate challenge level for your party, or didn’t understand the villain’s mechanics, and the fight was anticlimactic, don’t worry. Think about what you can do to make the campaign less disappointing. Maybe the PCs find a letter revealing that the villain is working for someone more powerful, or perhaps the villain has had a clone prepared using the clone spell and is resurrected in its clone, ready to continue its plans (just make sure the PCs can find some information that will let them know this might happen). A disappointingly easy fight isn’t a bad thing unless it’s supposed to be the climax of a campaign or important story-line (such as a PC’s personal quest). As long as you allow the PCs to figure out your plot twist it can help keep the players engaged and allow you to run the encounter again at some point armed with the information on what didn’t work.
What to Do if the Boss is Too Epic. So you made an epic boss for the players to defeat but either you went overboard in giving it stats and traits or the players were just having an off-night and it ended up killing them all. Well, if they’re all dead there’s not much you can do besides introduce a deus ex machina that may make the players feel like there’s no real consequences. However, if you keep in mind a few simple alternatives, you can believably end a fight without ending a campaign. When I ran Lost Mine of Phandelver from the Starter Set back when my campaign first started, the party were all knocked unconscious by the goblins in the first encounter. I decided that it didn’t make sense for the goblins to kill them on the spot because they didn’t pose a threat and the goblins didn’t have orders to kill them, and they weren’t important enough to kidnap so the goblins just looted all their good stuff and left them laying by the road. The goblins weren’t intended to be a boss encounter but they very well should have been for a level 1 party. With me using their hide ability to full effect the party didn’t stand a chance. Early editions of D&D talked more about what monsters would do in combat so you could play them at a difficulty in line with what the author of the adventure intended. In 5e, you really don’t see that much although it still exists in some encounters. Intelligent monsters have varied motives and it helps to know their motives before combat. They may not necessarily need to kill their opponent if knocking them out and looting them or taking them prisoner is sufficient. Beasts and other low-intelligence monsters may intend to kill them, but they might drag one of them off to a lair and leave the rest behind, saving the kill for later to keep their quarry fresh. I encourage you to avoid any solution that doesn’t make sense, like a starving wolf leaving a PC behind and not killing it or an evil NPC repeatedly knocking out the PCs and locking them in the oubliette even though they’ve already escaped four times. Sometimes a TPK can actually be the best option so consequences and believability are not impacted to the point where the players actually wish their characters had died.
Now get out there and stop holding back!