Dragna's Blog: The Seance Sins

In Curse of Strahd: Reloaded, unlike RAW, the players have an opportunity to heal the Abbot of the Abbey of Saint Markovia of the corruption that has tainted his soul (courtesy of the Dark Powers and Amber Temple) since he first entered Barovia. The way this arc unfolds is simple:

1. The players arrive in Krezk and learn that the Abbot is a Bad Dude.
2. Ezmerelda uses a seance to contact Saint Markovia, who tells the players to retrieve the Icon of Ravenloft from Castle Ravenloft, beat up the Abbot, and use the Icon to restore the Abbot’s divine grace.
3. The players retrieve the Icon of Ravenloft from Castle Ravenloft, beat up the Abbot, and use the Icon to restore the Abbot’s divine grace.

Seems simple enough, right? But as I’ve previously noted, Step 2 - the exposition phase - can often be the most challenging to get right. After all, D&D is a game, not an instruction manual, and so every scene - including those that dump lore and plot on your players! - must still be dramatic.

Put another way - the outcome of an expository scene must still somehow rely on the players’ actions. It doesn’t have to be challenging, and it doesn’t have to be significant, but there must be a real, actual dramatic question that drives the momentum of the scene forward.

This meant that I couldn’t just write out a cool seance narration and have Saint Markovia’s ghost show up to monologue at the players before vanishing. I needed interactivity. I needed gameplay.

But how?

The Bad Spirits

Here’s the first ruleset I came up with for the seance:

• The leader of the seance (Ezmerelda) must bare her soul to the spirits she calls, inviting them to possess her and so deliver their knowledge. However, many spirits from the western half of Barovia will hear her call when it rings out across the veil—and not all will be friendly.
• As each spirit arrives, its form will be indistinct, obscuring its true nature. The players must question the spirits to determine their identities, and tell Ezmerelda to dismiss them if they are not the one they are searching for.
• However, the spirits can lie. Furthermore, each question will focus Ezmerelda’s attention more deeply upon the chosen spirit, drawing it closer toward the circle and drawing its form into sharper view. Should a malevolent spirit draw too close, it will be able to cross the boundary of the circle and attempt to possess Ezmerelda’s body.
• A spirit that crosses the circle without Ezmerelda’s consent exposes its mind to the participants in the seance. To drive a malevolent spirit away, the players must delve into its memories and locate its name, which they can speak aloud to banish it.
• The players must choose one spirit at a time to question. When one spirit is chosen, the other spirits step back, remaining silent until the chosen spirit is accepted into the circle, dismissed, or banished.

The dramatic question is clear: Can the players find the right spirit and learn its secrets to save the Abbot before a malevolent spirit seizes control of the seance and possesses Ezmerelda? There’s clear (social) gameplay here: the players must question/interrogate the spirits one-by-one in order to identify which spirit is safe and which spirits are not.

Okay, this seemed fine enough. But which Bad Spirits to choose?

Well, the seance was taking place in Krezk, so I figured it’d be best to draw spirits from across the western half of the valley. I briefly pondered my options, then started with the following list:

1. Svarog, the leader of the druids at Yester Hill in Arc K: The Missing Gem
2. Valentin Vallakovich, the former Baron/burgomaster of Vallaki and Vargas’s father
3. “Sister Constance,” an original nun of the Abbey of Saint Markovia who perished during Strahd’s torments after Markovia died
4. Khazan, the archmage who built the tower at Lake Baratok, served as Strahd’s magic advisor, and eventually became a lich at the Amber Temple
5. Sergei (you know who this is)

Why these five?

• I knew that I wanted to take spirits from all across Barovia’s history, and Svarog seemed like a good option to echo the players’ recent actions and decision.
• Valentin seemed fun because I hadn’t really used him anywhere else (despite his historic impact on Vargas and Izek offscreen).
• Constance seemed necessary to me to show the fall of the Abbey instead of telling it.
• Khazan just seemed good to have, especially given his connection to the (broken) Sunsword.
• It just seemed good to start foreshadowing Sergei’s spirit now, given that he’d show up again later. (To clarify, I wasn’t sure whether Sergei would be a Bad Spirit, or whether he’d just have a cameo in another spirit’s banishment - more on this later.)

I brought these to the Patreon Discord server, and the members had a grand old time playing a demented version of Wordle to try and guess which ghosts I’d chosen. Once all the spirits had been revealed, however, an issue cropped up - namely, that I couldn’t really justify why I’d chosen some of these spirits and not others.

Khazan and Sergei seemed fine and good, since they clearly set up the Sunsword arc later on. Constance didn’t really do much other than give exposition about a place the players were currently standing in. And Svarog and Valentin had literally nothing to offer the campaign ahead of the players - at best, they gave little nuggets of information relevant to arcs that the players had already completed. Given my motto of, “No content for the sake of content!”, that just wasn’t good enough.

So I went back and picked some new ghosts. Out went Svarog and Valentin; in came Baba Zelenna (Baba Lysaga’s disciple, who “stole” the Forest Fane from Strahd and clashed with Lugdana in the battle that lost the Holy Symbol of Ravenkind) and Brother Grygori (the monk who persuaded the burgomaster of Berez to kill his daughter, Marina, to prevent Strahd from taking her). (Constance got to stay because at least she was relevant to the emotional stakes of the current arc - the corruption of the Abbey.)

Okay, so I had my ghosts. Now what?

What’s Your Name?

Now, it was time to hammer out the actual execution of the seance.

The questioning part was a bit weird at first - obviously the players will figure out that the spirit isn’t Saint Markovia? But I remembered two things: (1) the players don’t know they’re looking for Saint Markovia; and (2) the spirits can lie.

But what about the Abbey connection? Well, I figured that each Bad Spirit could pretend to be a different person who had some excuse to have a connection to the Abbey:

• Zelenna could pretend to be Lugdana,
• Constance could pretend to be Markovia herself
• Grygori could pretend to be Lazlo Ulrich (Marina’s father)
• Khazan could pretend to be Ciril Romulich, the former prefect of Castle Ravenloft

(Sergei, I decided, would just have a cameo somewhere, and wouldn’t actually be an antagonistic Bad Spirit. Less messy that way, and it preserved the positive tenor of his reveal at the blessed pool later on.)

Each Bad Spirit would have some excuse as to why they were the one who could help the players figure out the Abbot, and each one would be lying, waiting for the right moment to slip across the circle and steal Ezmerelda’s body.

That meant my next step was figuring out what to do when they did, inevitably, slip across the circle and begin trying to steal Ezmerelda’s body. As noted above, the idea was simple: the players had to enter the Bad Spirit’s memories and uncover its True Name. This had a double purpose: It gave the players meaningful gameplay and also provided additional background about the world they were in, and the narrative arcs they were pursuing.

Here’s what I worked out first:

• For Baba Zelenna, the players had to relive the moment when she defied Baba Lysaga at the Forest Fane in her efforts to summon the Green God Chernovog and overthrow Strahd. To find Zelenna’s name, the players had to read the contract she had signed with Chernovog.
• For Constance, the players had to relive the days when the Abbey fell from grace, meeting Constance in a fit of cannibalism beneath her desk in the North Wing. To find Constance’s name, the players had to read the carvings on the underside of her desk, which she had scratched with her fingernails.
• For Grygori, the players had to relive the night when he and Lazlo murdered Marina, Tatyana’s first reincarnation, to prevent Strahd from claiming her. To find Grygori’s name, the players had to read the author’s name on a book that Marina was reading.
• For Khazan, the players had to relieve the day when Strahd delivered the Sunsword to the archmage’s tower at Lake Baratok to be destroyed. (Finding Khazan’s name would be easy - just read the carvings on the lintel over the door like they’d previously done in Arc F: The Missing Vistana.)

This was a good start, but the gameplay wasn’t quite there; these segments were little more than a point-and-click adventure, and a laughably simple one at that. So I decided to take this a step further: the Bad Spirits were obfuscating their names, blurring them and leaving them unreadable.

To implement this, however, I still needed a gameplay loop. I turned once more to my memories of Diane Duane’s Young Wizards series - specifically, The Wizard’s Dilemma. In that book, worlds and universes - including the “world” within a soul - were controlled by “kernels”: tiny sparks of light that contained the intricate divine “code” that perfectly described the physical laws that governed those worlds.

Why not replicate that here, so that the players could find a spirit’s spiritual “kernel” in order to make the names legible again? I decided that, to read a Bad Spirit’s name, the players had to find the “spark of memory” that it had hidden somewhere. Once the spark was found, the players could hold it up so that its light illuminated the name, thereby revealing it.

But how to find the spark in each memory? My first thought was to implement a kind of “hot-cold” game: the players would receive some kind of item that would get hotter the closer they got to the spark, and colder the further they got. Ezmerelda already had a glass eye component for use in her clairvoyance spell, so I decided that this component was some kind of “eye of locate object” that could cast a very limited version of the locate object spell, and that Ez could give the players a copy of that item while exploring the spirits’ memories.

I brought this back to the Patreon Discord and some folks (very reasonably) pointed out that implementing a “hot-cold” mechanic is already awkward enough in real life, let alone when you move to the context of a point-and-click adventure - and even weirder once that point-and-click adventure is entirely verbal, as D&D is.

So I threw out the eye of locate object idea and decided on something simpler: The sparks would emanate a unique silvery light that would lead the players to the right room or area. Upon their arrival, the players would need to find the spark’s hiding place. To retrieve it, they’d need to overcome some kind of combat encounter (or encounter modeled as a trap) that could also be bypassed by solving a puzzle of some kind. (For example, in Baba Zelenna’s memory, the spark would be protected by a pair of root blights in a cavern beneath the Gulthias Tree, which the players could defeat by destroying the root connecting them to the tree.)

And then I brought this back to the Discord, feeling fairly pleased with myself, and someone (again, very reasonably) asked:

“Do we really need all this content?”

My Original Sin

Before I go any further, there’s one big thing I want to draw attention to here - a mistake I made that, had I caught it, might have saved me a lot of grief later.

I’ve occasionally described gameplay challenges as having “ranges of acceptable outcomes,” with a scale ranging from “unsuccessful” and “mostly unsuccessful” to “mostly successful” and “successful.” As much as we try to make our players think otherwise, most challenges/encounters fall very cleanly in a fairly narrow band of these outcomes.

For example, a random combat encounter on the road should almost always range from “successful” to “mostly successful,” simply because the dramatic foundation and stakes are sufficiently low that it makes no sense to risk a “mostly unsuccessful” outcome - even if our players are willing to accept that risk, we, as the DMs, are not usually willing to throw away our entire campaign just because the players got killed by a pack of wolves on the road.

Similarly, a battle with the final villain of the campaign should almost always range from “successful” to “mostly unsuccessful” (where “mostly unsuccessful” means “you didn’t get what you wanted but there’s a silver lining/you get to try again”), with the probability of “mostly unsuccessful” being very very low (since the players, presumably, want the campaign to end and don’t want to feel crappy after they’ve spent a lot of time and emotional effort trying to get to this point).

These possible outcomes, even if we don’t consciously assign them, should always be tied to probabilities. For example, let’s look at the battle with the night hag coven in Arc I: The Lost Soul in Reloaded. Since the arc comes with a built-in way for the players to recover if they lose, let’s say that we want there to be a real possibility of defeat.

Let’s arbitrarily put the possible outcomes at Successful: 60% (hags die and players live), Mostly Successful 30% (hags die and most players live), Mostly Unsuccessful at 10% (players TPK but survivors wake up as prisoners), and Unsuccessful at 0% (players TPK and die forever). To balance things accordingly, we want to make the fight tough enough that individual players are in danger of getting KO’d, but not so tough that the players are very likely to TPK.

Importantly, we can apply this process to the phases of an encounter as well. Which is what brings us back to my first mistake with the seance.

The challenge is simple: The players need to pick out the Bad Spirits to find the Good Spirit. But the challenge also comes with a fail-forward mechanism: if the players let a Bad Spirit into the circle, they can kick it out again by finding its true name and banishing it.

If we take the challenge on its own terms, this suggests a very classic Successful/Mostly Successful range: the players obviously need to get the exposition from Saint Markovia to proceed the quest, so we’re going to make identifying the Bad Spirits very easy.

But once I added in the fail-forward mechanism, I unconsciously made the challenge a lot more like the escape from Old Bonegrinder in Arc I. With the night hags, even though there’s a real chance of TPK, you have a very high chance of escaping once it’s happened. Here, too, I wound up designing this encounter under the assumption that, because the “banishing Bad Spirits” stage was relatively easy, the “identifying the Bad Spirits” stage must therefore be relatively hard.

This led to sloppy design because I wanted to make sure the players got to my fail-forward content. Put differently, I designed the entire challenge under the foundational assumption that the players would obviously fail to identify the Bad Spirits and the bulk of the “real” gameplay would take place in banishing those Bad Spirits, one-by-one, before eventually stumbling upon Saint Markovia at the end of the Spirit Queue.

This was bad! This was really bad! This was a Fundamental Sin of game design, and by god, I was going to Designer Hell for it.

This Original Sin also led to a secondary Sin - as suggested in the previous section, I’d violated my holy precept of “No content for the sake of content.” To see how, let’s compare two types of content: the gameplay to escape the windmill post-TPK in Arc I: The Lost Soul, and the gameplay to bypass the Bad Spirits in the seance.

In the Old Bonegrinder escape, it was important to flesh out the content for escaping the windmill in order to satisfy the need for meaningful consequences and semi-permanent narrative penalties following the players’ failure of an important quest arc; otherwise, the sense of stakes would be totally destroyed. If the players could have escaped the hags’ clutches as easily as Houdini escaping a pair of handcuffs, every potential future TPK would lose all sense of proportion.

But in the seance, the players didn’t particularly care about getting past the Bad Spirits, which were a gameplay construct I’d invented for the sole purpose of giving the players Something To Do in this scene. (Before this scene, the players didn’t even know the Bad Spirits existed, let alone that they were potentially dangerous.) Unlike Bonegrinder, there was no need to create “deep, meaningful gameplay” to create a sense of narrative consequences for failure; I could have just as easily dispensed with the risk of failure altogether!

In summary: I’d made the following errors:

• Because I wanted to add “gameplay,” I’d designed a scene that required the players to pass an arbitrary first stage in order to reach the second.
• Because I added an arbitrary required first stage, I made that first stage “fail forward,” such that the players could proceed even if they failed.
• Because I made the first stage “fail forward,” I added in complicated, meaningless gameplay for the players to complete when they inevitably failed.

The Problem Is the Solution

Notably, this all led to an additional problem that I haven’t yet mentioned: Because I’d offloaded all of the gameplay onto the first half of the scene, the second half - the players’ conversation with St. Markovia - was still nothing more than an expository infodumping lecture without drama or stakes.

It’s worth noting, mind you, that this conversation was the entire point of this sequence - and it had less dramatic impact and took up far less space than the random filler I’d plopped in front of it.

So I took a step back, trashed the entirety of my work on the seance thus far, and asked myself a very simple question: How do I add real gameplay to the interaction with St. Markovia itself?

This reminded me, very vaguely, of the “spirit board” seances in Van Richten’s Guide to Ravenloft, which got me thinking about ideas for “spirits communicating information through symbolic items.” And then I remembered that Ezmerelda owns her own Tarokka deck and everything fell into place.

Why is Madam Eva’s Tarokka reading fun? It’s full of exposition, yes, but that exposition is encoded.  The gameplay comes from the players’ clueless efforts to figure out what the heck the reading actually means.

Same thing here. Ezmerelda lays out her Tarokka deck flat on the ground and invites nearby spirits to communicate by flipping the cards over. Markovia can’t do much in the physical realm, but with Ez’s help and the magic of the seance, she could certainly do that much.

Of course, Tarokka cards don’t have any exposition on them - just cool pictures depicting things like “the Shepherd” and “the Anarchist.” Because there are so many of them, however, it becomes a fun puzzle to encode and decode different kinds of Tarokka messages. (For example, Markovia could communicate the Abbot’s history by first flipping the Shepherd card, then flipping the Broken One immediately afterward.)

Because Markovia can’t speak unless she’s using cards to answer the players’ questions, the scene also put the players directly in the drivers’ seat, keeping their attention and engagement throughout the process. After figuring out the minimum number of questions the players needed to ask to get the information they needed (seven), I decided to let them ask thirteen questions per seance, just to give them a bit of padding and additional freedom. (If they started going off-track too much, Ez could chime in and gently help nudge things back on track.)

I don’t know about you, but I immediately felt a lot happier with this version of the seance than the prior ones. It felt elegant in a way that the originals didn’t, focusing on core narrative and thematic elements while presenting the players with a clean, simple, but engaging gameplay loop that also promoted agency and exploration.

And in the end, is there any more you can really ask for?