Dragna's Blog: Tension Curves in Castle Ravenloft

Arc P: Ravenloft Heist is probably one of the strangest arcs I’ve had to write so far in Reloaded, structurally speaking.

At its core, Castle Ravenloft is a romp through a giant dungeon sandbox. It’s a labyrinthine, mazelike apparatus where often the only way to go to another room on the same floor is to go one floor up or down first. Numerous junctions and dead ends dot the keep’s interior, and countless secret doors make it easy to criss-cross different areas if the players successfully spot them.

In the original module, the players’ exploration through the castle isn’t a story so much as a pseudo-random walk through unfamiliar, alienating territory. Random encounters abound throughout the keep - some dangerous, some not - and several rooms include small squads of monsters that serve to ambush the players and immediately die (e.g., the Strahd zombies on the chapel balcony and the torture chamber; the Barovian witches in the cauldron room.)

That’s not to say that there are no notable “narrative” events in the castle. Players traversing the keep might encounter:

• The illusionary Strahd, followed by the raising of the drawbridge, if they enter the dining hall on the first floor

• Rahadin, who randomly appears, directs them to a random location in the castle, and then fucks off

• Ezmerelda (if they haven’t met her already), who can be recruited as a permanent companion

• Strahd, who randomly appears and then attacks

It’s noteworthy, though, that of these, Rahadin’s encounter is the only one that provides any sort of narrative structure (“Strahd wants to see you. Go to X location to meet him.”) - and even then, it’s a complete non sequitur/red herring that accomplishes nothing but waste the players’ time (especially since most of the encounters in the castle aren’t very dangerous for a 5th-level party or above, especially if they have a cleric with spirit guardians, a paladin, the Sunsword, or the Holy Symbol of Ravenkind).

Now, this style of random/emergent narrative through gameplay is a valid way to run the game; it’s been the backbone of D&D sandbox design for decades. (See also: Dungeon of the Mad Mage, Tomb of Annihilation, and so on.) At the same time, this “location-based” approach to game design severely limits a Dungeon Master’s ability to create tension and catharsis for their players. Reloaded as a whole departs from this style of module design and focuses on a “narrative-focused” approach instead, organizing its chapters and arcs not by rooms, but by scenes.

But how do you preserve the player agency inherent in a sandbox while also transmuting it to a narrative-focused design? The whole point of a sandbox, after all, is that the narrative is emergent from the players’ actions and decisions. Doesn’t imposing a particular narrative upon the sandbox destroy the very thing that creates it?

The Hitchhiker’s Guide to Castle Ravenloft

Back when I wrote the original Reloaded, I spent a good amount of time trying to figure out exactly how this giant, confusing castle fit together.

As I did, I began to notice that there were two “halves” of the castle: all of the areas that were accessible via K20 (Heart of Sorrow), which connects to some hidden bedrooms and (via a secret door) the servants’ quarters on 2F, Strahd’s study and bedchamber on 3F, and the guards’ chambers and servants’ hall on B1F; and all the areas that were accessible via K21 (North Tower Stair), which connects to the main entry and dining hall on 1F, the audience hall on 2F, Strahd’s dining chamber on 3F, the guest suite on 4F, the kitchen on B1F, and the dungeons on B2F.

Put differently, there’s a “private” half of the castle (accessible via K20) and a “public” half of the castle (accessible via K21). These two only intersect at three particular points: (1) the secret door and staircase that allows Strahd to go straight from the audience hall to his private chambers; (2) the door in K20 that allows servants to deliver food and sundries to Strahd’s study and bedchambers; and (3) the servants’ hall, which connects the kitchen to both the guards’ quarters and the dining halls above.

This was all very well and good for figuring out how to navigate the damn place, but how did it help me figure out how to structure the narrative of the Ravenloft dungeoncrawl? Well, in revisiting the castle in RRL, I decided to re-examine the castle layout to try and figure out if there were a better way of doing things.

As I reviewed the castle maps and built a boxes-and-lines map of the castle, I realized that the castle is actually much less of a sandbox than people tend to think. Lots of rooms, such as Strahd’s chambers, the chapel/audience hall/foyer area, and the kitchen area, form a self-contained loop. Others, such as the guest suite or witches’ chambers, are simply dead ends.

I began to suspect that I could segment these areas into self-contained “sub-dungeons” - self-contained areas that the players were likely to traverse in full before proceeding to the next area. (Having run the Castle Ravenloft dungeoncrawl seven or eight times, I can say with complete confidence that this is how exploration tends to go in practice as well.)

After organizing the rooms of the keep by utility, connections, and location, I came away with ten self-contained “regions” of the castle:

• The Castle Grounds, the outdoor areas of the castle (e.g., the front courtyard, overlook, and garden)

• The Castle Commons, the areas of the castle expected to be frequented by members of the public (e.g., the dining hall, foyer, chapel, and audience hall)

• The Servants’ Wing, the areas of the castle populated or worked by the castle servants (e.g., the servants’ quarters, wine cellar, and kitchen)

• The King’s Chambers, the areas of the castle where Strahd resided while alive (e.g., the study, dining hall of the county, and bedchamber)

• The Treasury, the areas of the castle concealing Strahd’s treasury (e.g., the false treasury, belfry, and treasury)

• The Guest Suite, the area of the castle where invited guests stayed (e.g., the portrait of Strahd, lounge, and guest bedroom)

• The Coven’s Quarters, the area of the castle where the coven of Barovian witches worked (e.g., the familiar room, element room, and cauldron)

• The Barracks & Fortifications, the areas of the castle where the castle guards once lived and worked (e.g., the hall of bones, kingsmen hall, archers’ posts, and Heart of Sorrow tower)

• The Dungeons, the areas of the castle meant for imprisonment and torture (e.g., the north dungeon, south dungeon, torture chamber, and brazier room)

• The Castle Crypts, the area of the castle in which the dead were laid to rest (e.g., the catacombs, Strahd’s tomb, and Sergei’s tomb)

Once I had these regions down on paper, it was time to figure out how to fill them. But how?

The Tension Curve

As I looked to work out a narrative for each of these regions, the obvious starting point was to begin with a dramatic question for each one. In addition, following my recent DMing handout regarding anticipation and catharsis, I also wanted to experiment with creating “tension curves” for each different area, slowly ramping up the tension each time the players entered an area before eventually reaching a climax that represented the players “clearing” that area.

Before I could do that, though, there was a preliminary issue to deal with. Each of these regions has multiple entrances - for example, players can access the Servants’ Wing from the Castle Commons (via K21, the North Tower Stair), the Barracks & Fortifications (via K67, the Hall of Bones), and the Castle Grounds (via K23, the Servants’ Entrance). How could I plot out a linear narrative curve for an area when its rooms could be traversed nonlinearly?

I took another look at the castle maps and noticed that most of these areas had (1) dead ends that players couldn’t approach from, (2) major nexuses that the players would have to traverse eventually, or (3) both. In particular, these dead ends provided great locations for area “climaxes” (since the players would have to go through the rest of the nearby area first), while the nexuses provided great locations for area “inciting incidents” (since the players were all but guaranteed to wind up there, and would likely do so fairly quickly).

For example, let’s take a look at the castle commons, which includes the entry, great entry, guests’ hall, dining hall, grand landing, audience hall, king’s hall, king’s balcony, and chapel. Since the staircase from the catacombs to the chapel is blocked off and the players generally have to pass through the great entry to reach or exit the chapel, the chapel is functionally a dead end. That made it a great place to put a climactic fight with Ludmilla (who I’d wanted to place there anyway because her flight + mist powerset made its wide-open space and tall ceilings a perfect battleground for her).

Meanwhile, since every obvious possible path through the castle commons (i.e., coming from the south tower stair or the castle courtyard) inevitably passed through the great entry before going anywhere else, I could very easily place this area’s “inciting incident” (i.e., the gargoyle attack) in the great entry as well. From here, it was just a matter of figuring out how the area’s dramatic question evolved to ensure the players had a reason to travel from the great entry to the chapel once the inciting incident occurred.

Okay, But The Arc Itself Still Has a Climax, Right?

It’s one thing to say “this dungeon is flexilinear with sandbox elements, which allows the players to roam around and do different chapters in whichever order they please, depending on which choices they make at which dungeons.” It’s another thing entirely to say “. . . and that means the players can do the lowest-tension area last, thereby ending the arc with all the gravitas of a wet fart.”

The nature of a tension curve is simple:

• We begin with low tension. The inciting incident drives us up to a point of greater tension.

• The players then face the first obstacle to answering the dramatic question. Through the use of their agency, the players overcome the first obstacle, defusing some of the tension and providing some amount of catharsis. (The players don’t defuse all of the tension because the dramatic question still remains unanswered.)

• The players then face an additional, secondary inciting incident or a complication to the first dramatic question. This raises the tension again, bringing the players to a second obstacle, which they again overcome using their agency. This defuses some of the tension and provides some additional catharsis, but leaves the tension (on net) slightly higher than when the players overcame the first obstacle (e.g., because the stakes are higher, the players are lower on resources, etc.).

• Eventually, the tension reaches a peak - the climax, which contains the ultimate obstacle. If the players successfully conquer this obstacle, all (or most) of the tension is defused, providing an enormous amount of catharsis. The catharsis of the climax, if executed properly, is like a drug: because the tension has been built so high, its release provides a sense of near-euphoric joy, triumph, and/or relief to the players, which creates an emotional satisfaction that makes the whole experience feel “worth it.”

Now, at least in part because most of Reloaded was written before I started developing a theoretical understanding of catharsis and tension curves, not every arc currently meets this lofty goal. However, there are a fair few arcs in the guide that currently come close, including Arc B (the climax is Doru conquering his bloodlust in his trial), Arc H (the climax is defeating Leo and returning Stella to her body), and Arc I (the climax is destroying the statue, retrieving the gem, and saving Muriel from her bloodlust).

In any case, when I started outlining Ravenloft Heist, I had the luck to already have the concept of tension curves in my toolbox. The arc’s tension curve begins something like this:

• The inciting incident takes place when Strahd departs the castle on the night of the new moon, leaving the keep ripe for the players to invade.

• The first obstacle is the players’ entry into the castle, which requires the use of agency to cross the chasm and enter the keep without alerting or being defeated by the wight guards patrolling the walls.

From here, we enter into what screenwriting calls the “fun and games” of a screenplay - the assorted obstacles that the players must overcome as they navigate the castle. The main obstacles include:

• Ludmilla, who’s surveilling the main entrance with the castle gargoyles

• Anastrasya, who’s guarding Gertruda with her wight servants

• Volenta, who’s watching over the dungeons with her ghoul playthings

• Rahadin, who’s meditating in King Barov’s tomb near the catacombs

But there’s a problem - if all of these encounters are the obstacles, then what’s going to be our climax? Rahadin seems like the most suitable climax - he’s far stronger than the Brides, hasn’t yet been defeated by the players, and is Strahd’s right-hand man. But if we want him to come last, we need to make sure that the players (a) have no ability and/or (b) have no reason to visit the catacombs before clearing the rest of the castle first.

In Dinner with the Devil, I allowed players to send Ezmerelda to scout out the castle grounds in an effort to find alternate entrances. One of the entrances she could convey was the windows leading into King Barov and Queen Ravenovia’s tomb below the overlook. Speaking from experience, in every campaign I’ve run where Ez told the PCs this entrance existed, the party inevitably chose it to break in.

That meant, unfortunately, that players who learned of this entrance would inevitably find themselves led straight to Rahadin. Oops. Guess this “helpful hint” was out.

So Ez can’t tell the players about the tomb windows. Okay; that’s fine. But how do we stop players from going to the catacombs before first obtaining the skull, Gertruda, and Emil?

Well, let’s think about how the castle is structured. There are two other entrances to the catacombs: the high tower stair (connected to the chapel) and the stone tunnel (connected to the brazier room). The high tower stair is blocked off by a mortared wall, and I doubt players would spend a dimension door on getting past it (or try to blow it up and risk getting caught), so that’s not a concern.

What about the tunnel? The brazier room has three other entrances: the secret passage from Rahadin’s office, the staircase from Strahd’s bedchambers on the third floor, and the balcony overlooking the torture chamber by the dungeons. Put differently, the players might arrive here (1) after obtaining the skull in the Hall of Bones (but before doing anything else); (2) after rescuing Gertruda (but before doing anything else); or (3) after rescuing Emil from the dungeons (but before doing anything else).

(Admittedly, #3 is less of a concern because the slide trap means that most parties will wind up in the dungeons after entering the tunnel anyway, but we’re going to disregard that for the moment.)

So how do we avoid the problem of the players going, for example, skull > brazier > catacombs? Well, just as each of these three routes provide alternate entrances, so too can they provide alternate exits leading to their respective plot hooks (the skull, Gertruda, and Emil). This works as a solution because, as a general rule, players will almost always explore a route with a more immediate hook over a route with no hook at all.

(The main exception lies in obvious dead-ends; some groups will prefer to take a quick look at any small, non-notable routes - such as closets or side chambers - before departing for a different area. The catacombs, however, lie at the end of a long, misty tunnel; most players will prefer to mentally file that off for later and return once their other objectives are complete, especially if we add elements of urgency to the other pathways and make the non-urgent pathway somehow unpleasant.)

So how can we use the brazier as a junction that encourages the players to investigate the spiral stairs (leading to Gertruda), the torture room (leading to Emil), and the secret staircase (leading to the skull)? Let’s add some immediate hooks to each one:

• The Spiral Stairs. I like to think that Gertruda used a hairpin to pick the lock on this door while unsupervised. She made her way down the stairs and found the brazier room, but was captured and returned to her chambers before she could escape. Perhaps she lost one of her shoes/slippers in the struggle, which now lies halfway up the first circuit of stairs. (Note that the positioning of the slipper is critical - it has to be partially obscured by the curve of the staircase to make sure the players climb the stairs to inspect it (thereby setting a default behavior/momentum of “ascending the stairs”), and it has to be pointed upstairs, rather than downstairs (to signal to the players that its wearer was going up, not down)).

• The Torture Room. If the masked Barovian from the dinner with Strahd is still alive, we can easily have his moans of pain attract the players into the torture chamber. If he’s dead, we can have his mask hanging from the door or scattered on the floor nearby.

• Rahadin’s Office. This is a tough one. We could have recent footprints descending through the thick dust on the stairs, but that doesn’t give us a direct hook up. We could have inspection of the footprints evince a very faded echo of Rahadin’s ghostly screams, but there’s not really a Watsonian explanation for that. Ultimately, all we really need is to get them to the first landing (since then they’ll notice that the stairs end at a “dead end” and naturally investigate), so let’s instead place a shadow watching the players curiously from around the bend. If the players befriend it, the shadows can show them around the Garrison, including Argynvost’s skull.

Finally, how do we discourage the players from going down the catacombs tunnel? The original module already does this for us - by preventing the players from seeing more than a few feet ahead in the fog, the tunnel creates ambiguity, removes any prospect of instant gratification, and raises the dread the players feel about what lies beyond, encouraging them to return later, once they’re sure they’ve checked all other paths.

So there you have it - a solid tension curve with a modular structure that heavily incentivizes the players to do the catacombs last, thereby allowing for a final climactic confrontation with Rahadin before fleeing the castle. Will it work out in practice? I’m looking forward to hearing about it!