created by Zander Majercik, Kai Ssempa, Izzy Meyerson, Hannah Yu, & Gracielly Abreu

Artists’ Statement

Mark of the Witch is a game designed to quickly facilitate complex social deduction. Freeform social deduction games often have a lot of inertia because concrete information is revealed slowly, or not at all. This means that players often have very little information in which to begin solving the mystery.

By contrast the structured mechanics (e.g. spell cards and defined trial sequences) of Mark of the Witch allow players immediate access to information so they can begin role deduction. This access to information also allows players to bluff conditioned on information that is incomplete but definite. This allows for more nuanced social deduction and avoids the conversation devolving into one player insisting that they are innocent and no other player having any hard information to disprove them. With more definite information comes an additional social deduction challenge, as players must figure out who is on their team in addition to achieving their team’s goal.

We set our game as a witch trial because we all loved the aesthetic choices that setting permitted us, such as unsettling symbology, a minimalist aesthetic, and the framing of card effects as “spells”. Further the historical formalities of a witch trial, such as the various tests to determine whether or not someone was a witch, give us a thematic opportunity to add more structure to our player discussion mechanics in service to the gameplay goals mentioned above.

Concept Maps of Game System

Elements of the Game

Turn Circle

Initial Mechanics and Values

In our first brainstorming session in class, one member of the team suggested a Salem Witchcraft-style game. We immediately all jumped on board with the theme. At first we envisioned a board game, but settled on a card-based game as that felt more within our collective wheelhouse.

Initial Concept Document

As seen on our concept document, we initially saw the main aesthetic of play as narrative, because we envisioned a murder mystery, who dunnit-esque game that was facilitated through the use of narrative and testimony cards that gave room for players to bluff. For example, each player was assigned a character like “The Baker’s Daughter” and would have character specific narrative cards that gave increasingly more detail to the group (e.g. “I was helping my father bake rye bread at the time of the child’s convulsions, so it could not have been me!”). However, we decided to pivot away from this narrative focused game because we wanted to create a game that could be playable more than once, which a Who Dunnit game style would not have allowed for.

We liked the idea of cards facilitating movement and dynamics in the game, so we kept that element of it. We wanted to have very structured mechanics that entailed a classic turn-taking card-drawing with emergent complexity. The cards allow players to take actions that allow them to gain knowledge used for the social deduction portion of the game. The social deduction is enacted through “trials” in which players vote to execute a witch. Before the vote is carried out, each player is able to give a testimony to advocate for or against another player (or themself). This is in direct contrast to freeform social deduction games where anyone can talk at any time.

With these mechanics in mind, we felt that the target audience would be individuals 13+, slightly older than the the ages for other social deduction games. This increase in player age is because there are more complex social deduction factors at hand. Our game is best suited for game players that have some familiarity with other social deduction games like Mafia, Love Letter, and One Night Werewolf. We thought that due to the more intimate fast-paced nature of this social deduction game, that the ideal player number would be 4-7 players (with adjustments made in the roles in play).

So our initial prototypes were centered around these turn-taking, card-drawing, and limited testimony mechanisms, which we put to the playtest!

Testing and Iterating on Prototypes

Prototype 1: Love Letter Mod (Playtest 1 and 2)

With our first prototype, we wanted to test how our structured card draw mechanics would interact with the touchstone trial mechanic common to many social deduction games. Given that the trial mechanic is relatively simple (with complex dynamics arising from player discussion which are in turn facilitated by drawing and playing cards), we decided to mod a game with established card draw and interaction mechanics: Love Letter.

The following analysis is for two playtests that we did (in two separate classes). The analysis is combined because we did not update our prototype between the playtests because the first playtest did not yield enough information to us to justify any changes (most time was spent just getting the game going).

Playing Cards Modded from Love Letter

Playing Cards and Their Effects

Role Cards

These 5 players must deduce who the witch is in order to vote them out, whereas the witch is trying to survive two trial phases in order to win. These players are able to begin their social deduction by drawing and playing cards, which will allow them to take actions to gain slivers of information (ex: look at the role card of the player to your left).

For this first playtest, we just wanted to focus on how the card taking and playing would interact with the trial phase, so we kept the cards similar to those used in Love Letter, but with different names (Bodyguard, Mercenary, Priest, Drunken Brawler, Dinner Guest, Journalist, Assassin, Fortune Teller, Mob Boss/Mayor). Because we weren’t able to get any meaningful gameplay in the first in-class playtest, we kept the same prototype for the second in-class playtest.

Pros:

Once players picked up on the strategy and adjusted to the complexity of the card mechanics, they understood how those mechanics would support the social deduction aspect.

Cons:

However, the prototype did not satisfy our initial goal of getting to the trial phase (the embodied social deduction aspect) quickly. Again, we were limited in our time and there were too many rules to grasp out of the gate, leading to unnecessary complexity and a steep beginner’s learning curve. The rules and the motivations behind the rules were not put into the appropriate context, so players were confused.

Main Questions We Were Left With:

Critically, it was unclear how the witch’s goal was significantly different than any other player’s goal, and the odds seemed rather stacked against the witch’s ability to win. Finally, as our roles were similar to those from Love Letter, the roles in play were not thematically aligned with our game’s narrative.

Prototype 2: Internal Group Playtest

After analyzing the results from our first prototype and the two playtests carried out with it, we knew we had to hone our game mechanics more towards the theme and simplify the gameplay. We also wanted players to get to the trial phase more quickly. Thus, we implemented some major changes:

1. Instead of one witch, all players are split amongst a witch team and a non-witch team. Each individual knows what team they are on but must deduce the team that all other players are on. This change allowed for the witch team to have a fairer shot at winning. It also reframed the game as more of a team fight, which made the similarity in goal between the Witches and Non-Witches more intuitive. Further, this gave every player a new social deduction goal of identifying their teammates (ie the Witches did not know who the other Witches were).
2. Players can now call a trial at any time (on their turn). Whereas before players had to wait for the “trial phase,” players could now more quickly access the social deduction stage. However, this requires the players to have more information, which we address in the following changes.

We greatly simplified the card mechanics. In Love Letter, the cards present many ways to eliminate other players. We decided to focus our cards on gaining and exchanging information to be used for social deduction rather than player elimination.

New Gameplay Mechanics

We now had 2 witch roles, 1 martyr, and 2 non-witches. The non-witch team can win if both witches are eliminated whereas the witch team wins if two members of the non-witch team are eliminated. The martyr’s sole mission is to be voted out. If they succeed, they win the round.

We also now had a simplified deck of 14 spell cards that allowed players to gain information to help them deduce other players’ roles. Role cards are dealt to each player and each player starts with a spell card. The remaining spell cards are placed in a draw pile. On a player’s turn, they must draw 1 spell card. They can then play 1 of their 2 spell cards and resolve its effects OR they can discard 1 of their spell cards in order to call a trial.

EXAMPLE OF SPELL CARDS (white) and ROLE CARDS (pink)

In the trial phase, an accuser names an accused. The testimony phase then ensures where each player may say one sentence (to defend, accuse, or otherwise) or may pass. All players must then simultaneously vote using one finger to indicate guilty and two fingers to indicate innocence. If there is a majority, then the majority verdict wins and a player is either eliminated or found innocent. If there is a tie, the player is found innocent for this round.

Because this playtest was conducted internally, we were able to iterate these changes as we played, seeing what worked thematically and mechanically and what needed to be modified.

What Went Really Well:

Not knowing who was on your team added another layer of social deduction which made it so much more fun and complex! For example, Hannah was a Witch and thought that Zander was a Witch as well because he made an early move to eliminate another player with a spell card. As a result, on her next turn she called trial and accused Izzy, who she thought was a non-Witch, believing that Zander would vote guilty as well. However, as it turned out, Zander voted innocent (surprising Hannah!) and Izzy was eliminated and turned out to be a Witch (surprising Hannah even more!). From this, Hannah was able to deduce that Zander’s early aggressive/suspicious behavior wasn’t because he was a Witch, but probably because he was the Martyr.
The Martyr role was a hit! As explained in the narrative above, the Martyr role added a fun dynamic to the game as well because it prevented players from assigning suspicious behavior to solely the potential Witches. We knew we wanted to keep the Martyr role in the final prototype.

What We Needed To Fix:

We learned that we needed to fine tune the frequency of certain spell cards (making some more rare and others more common). For example, we had too many PEEK cards that just got annoying at some point, because it seemed like that’s all anyone was doing. We added more curse cards which allow players to swap roles.
The amount of silence throughout the game made it a bit… boring? Since we were only allowed to speak during the “gossip phase”, this meant that even if some of us knew really important information (like someone was a Witch), we couldn’t say anything. And if we were eliminated then that information died with us as well, which meant that the game couldn’t progress as much and have that fun conversation element that games like Mafia and Werewolf have.

Changes That We Made:

Adjusted frequency of certain spell cards
- Added more CURSE cards which allowed players to swap roles
- Decreased PEEK to 4 in the deck
Allowing players to speak casually throughout the game instead of limiting it to the Gossip Phase
Creating a page of rules to explain the structure and mechanics of our games, to prepare for playtest with new players!
Updating the playing card designs to something better than slips of paper!

Prototype 3: Final Playtest

For our final playtest, we created some new components of the game. The first was creating a rulebook for our game that explained the roles, objectives, mechanics, and win conditions of our game.

Rulebook V1

We also created a more formal playing deck of cards (as opposed to our previous slips of paper) with tarot card design backs to help our players buy in more to the theme and card-drawing mechanics.

Updated Playing Cards (More on Theme)

We tested our third prototype in class during the final playtest. We found that players were confused at the beginning of gameplay. However, this time it was not because the rules were too complex but rather that there is a lot of information to take in. It is one of those games that takes several rounds to grasp the strategy, which is common in many games similar to ours.

It is clear from our playtest video that there were quite a few questions that popped up throughout the gameplay. Some questions were insightful for us as game developers because they revealed small details that we had not thought to clarify. For example, a player asked us how long the gossip phase was—which we had never considered as an important detail! We made sure to note this down so we could add it to our rulebook. Another detail we noted to change was to make explicit that after playing the SWAP spell card that both players could view their new roles, since there was some initial confusion about this. Players also asked about the win conditions for the Witch—did they need to outnumber just the villagers, or villagers + martyr? We noted this down to clarify in the rulebook.

Another type of question our moderator received during the gameplay were questions about rules that had been explained in the rulebook. We understood that our game rules had many details, but we thought that the rulebook should have been enough to help players learn how to play the game. However, it was clear that our rulebook was not as effective as we thought it would be, given that players defaulted to asking the moderator clarifying questions instead of consulting the printed rulebook. We knew we had to make major changes to our rulebook to improve readability and digestibility.

However, we noticed that most players asked questions regarding strategy, which we found to be a positive sign as it meant they grasped the general concept of gameplay. For example, players asked the moderator “What is the benefit of being able to see another person’s spell card?” This further reflects the complexity of the game, which is neither a good nor bad thing. We are happy, as game makers, to create a game that requires thought and strategy. By the end of the game, however, players were engaged with the strategy, leading us to believe that we should just revise the rulebook to set expectations for new players.

We thought this captured a critical moment of fun in the playtest once players got a hang of the rules and could start executing their social deduction strategies!

The feedback from the Google form indicated that players had lots of fun playing our game and found Mark of the Witch to be more structured than most social deduction games, as we intended. There was more tension, a greater exchange of information and ability to gain information, and a faster pace than many social deduction games like Mafia. However, there is a steep learning curve for beginners which could be ameliorated by some rule and gameplay clarifications.

After this playtest, we knew some modifications were in order. To reduce the learning curve and allow our game to be more beginner friendly, we made MAJOR (!!!) changes to the rulebook in addition to other small revisions we thought would improve the visual appearance of our game and game flow.

Changes to Rule Book

Before and After (only one photo is included for each version)

More Photos of Updated Rulebook

For a flipbook version of our final rulebook, please visit: https://online.flippingbook.com/view/522289375/

Changes We Want to Highlight in the Rule Book

Clarifying some mechanics at the beginning of the rulebook to help players start to understand how to develop strategies for social deduction.
Adding a FAQ section to help users easily look up common questions.
Clarifying win conditions for Witch and that they have to outnumber ALL other players, not just villagers.
Including an Example Turn that highlights strategy and interesting card interactions.

Other Revisions:

Changed the Martyr role win conditions—In addition to winning alone if they get voted out, we added a win condition that the Martyr can win with the villagers at the end.
- Players in our final playtest said that once people knew who the Martyr was, the Martyr had no incentive to vote for either villagers or witches. With this change, the Martyr still has an objective even if their role has been found out.
Added a double cheat sheet card with all the spell cards, their effects, and frequency for easy player access.
Decreased Divinitation card frequency from 2 to 1, because players suggested it was a little overpowered and took some of the social deduction fun out of the game.
Changed the name of Gossip Phase to Testimony Phase, to be more on theme.