by Mike Shea on 8 August 2016
Actual monster difficulty in the Dungeons & Dragons 5th Edition Monster Manual varies within challenge ratings. Monsters such as the beholder and the wraith punch well outside of their challenge rating when compared to equally-rated monsters. The banshee is another such monster, due to its devastating wail attack. Though only a challenge 4 monster, the banshee's wail attack can drop an entire group to 0 hit points if they fail a DC 13 Constitution saving throw, regardless of their hit points.
This variance of monster difficulty in D&D 5e is something we DMs will have to learn to deal with. We have the tools to handle underpowered monsters with tricks like increasing their hit points, increasing their ability scores, adding environmental effects, or adding other complications to the battle. Sometimes, though, things can go far in the other direction and take everyone by surprise. A bad few rolls with a monster like the banshee can lead to a total-party kill (TPK).
"3 players showed today. Entered a room with 2 banshees who won initiative. All PCs failed the save. Dropped to 0 HP."
What does a DM do in this situation? How do we recover from a total-party kill? We have a few options, including simply ending the campaign with the death of the party but there are other less harsh ways to handle it that can open up entirely new pathways for the story.
What if, instead of ending the campaign, a TPK by the banshees put the PCs in a new and entirely unexpected situation? What if they found themselves somewhere else, in some other predicament led to by their loss at the banshees?
Story-focused gamemasters and game systems refer to this as "failing forward". It's a simple idea but a difficult one to put into regular practice. In fact, we may never put it into play if we never find ourselves facing such a situation.
The core concept of failing forward is to put the PCs in a new and interesting position when they fail at something rather than have them simply fail. As a simple example, let's say that a thief is trying to pick the lock on a sturdy door and rolls a thieve's tools check with a DC of 12. If they roll a 5, we might simply say "sorry, you can't open the lock", but what if the story is really on the other side of that door? What if this failure pushes the entire game away from a fun path? Instead of pure failure, we can have them fail forward. Sure, the rogue picks the lock, even with a 5, but the racket from the clicking alerts the guards down the hall who come running. Now the duke's wife woke up and burst out of the door of the duke's longtime advisor wearing nothing but her smallclothes. Now that failed lock picking has some interesting consequences.
A banshee knocking an entire party down to zero hit points is different from failing a lock picking check, though. How do we deal with a total-party kill on the wail of the banshee? We can still fail forward, even in this extreme circumstance. Here are a couple of thoughts:
One of the PCs (whoever rolls the highest constitution check) wakes up, barely able to open their eyes but still, miraculously alive. They see the banshees slowly draining the life out of one of their other companions and has an opportunity to do something to get away, stir up trouble, or lure the banshees away from their companions so they can get them back on their feet.
What if there's something even worse than the banshees that comes for the PCs. From their dream-like state, the PCs see a hulking armored figure come in and, with a single abyssal word, send the banshees cowering. The armored figure pulls up the PCs and drags them into the depths of a ruined castle, where they are dropped into a charnel pit as fuel for a horrible undead ritual.
What if a necromancer comes, uses "control undead" to push the banshees away, and collects the PCs for his own hideous experiments within his dark laboratory. Now the PCs have to figure out how to escape from the necromancer's clutches.
Notice that the PCs aren't miraculously saved in any of these situations. If the PCs got wiped out by some monsters, we can find a way to bring them back into the story that doesn't simply make the world right. Instead, they are still in a tough situation and have to find their way out of it. There are no easy ways out when you're wiped out by a pack of banshees but that doesn't mean there isn't a way out at all.
These failing forward moments work best when we have some idea how they might come to pass. It behooves us to have some ideas prepared if we know a battle is going to be particularly tough. Before our game we can ask ourselves what we'll do if the game heads towards a TPK. We don't need to plan much, just enough to make sure we have something to lead in to if things go bad.
This is very different from expecting and planning for a TPK. In my opinion, planning for a losing battle is a lame way to force the story into one particular direction. In 5e D&D, its definitely possible to build battles that wipe out PCs (it was a lot harder in the 4e days) but doing so just to move down one path of the story is the worst form of railroading. The PCs, and by extension the players, have little to no agency in such stories. They're not even given the benefit of a false choice. They go in against insurmountable odds, wipe out, and then end up following whatever path you want them to go.
Exposing PCs to powerful monsters is fine but they should always have options to escape, negotiate, or, in some cases, actually fight their way out without the predetermined path that they're supposed to lose. Such battles hold little appeal. Instead of being afraid, players will recognize the impossible threat and just wait for the story to come along and save them.
Instead, we can offer hard situations with hard choices, some of which might lead to a TPK and some which might lead to a powerful victory. We can prepare (the lazy way of course) for numerous outcomes.
Any time we're designing a powerful battle we can do a quick test to check if we're railroading by asking "what two or more outcomes are possible from this situation?" If there are two or more solid outcomes, we know it's not a pure railroad. If there is only one, we might be pushing things too hard in a single direction.
In an old 4th edition D&D game, our band of adventurers faced a particularly powerful upgraded version of Orcus. I had really made this guy hard, with the intent of pushing our group of level 30 PCs to the brink of death. There was a good chance the PCs might all die. What then? Do we end the 30th level campaign with the PCs losing to Orcus? I don't think so.
Instead, I prepared to fail forward. What if Orcus won? What if he was the new God of Death? What if the entire prime material plane was now fueling his giant soul-eating machine?
What if the PCs awoke fifty years later, fifty years into Orcus's reign, resurrected by one of the other gods or demon princes (maybe Lolth herself!) to once again infiltrate Orcus's new bone palace and slay the new God of Death?
Just because Orcus could win doesn't mean the game is over. It might mean it's just beginning.
That's failing forward.