Name of the Game: Blood on the Clocktower
Game’s Creator: Steven Medway (Designer)
Platform of the Game: Physical Board Game (with a Web App for online play)
Target Audience:
- Player Count: 5 to 20 players (includes 1 Storyteller).
- Age Range: 15 +
Player Profile: Fans of party games, advanced social deduction, complex logic puzzles, and story-driven or moderated gameplay.
I played Blood on the Clocktower (BotC) with friends from my undergraduate club via a web conference. This social deduction game shares a core similarity with “The Shrimp Game” that I am currently designing: the game features many characters, but players do not know exactly which characters are in play for the current session.
Figure: BotC Character Sheet
As shown in the image, BotC includes four types of characters: Townsfolk, Outsiders, Minions, and Demons. The first two belong to the good team, while the latter two belong to the evil team. The goal of the Townsfolk and Outsiders is to find and execute the Demon. The goal of the Minions and the Demon is to destroy the town, leaving only the Demon and one other player alive. In the game I participated in, there were 8 players in total, corresponding to 5 Townsfolk, 1 Outsider, 1 Minion, and 1 Demon.
Figure: The character composition of my game session
In classic social deduction games like Mafia, Werewolf, or Avalon, the types and numbers of roles in play are fixed. For instance, in Avalon, all players know that “Merlin, the Assassin, and Morgana are in play,” and the core mystery is “which player is Merlin.” In BotC, however, the game adds an extra layer of fog: players do not know which characters are actually in the game. For example, there are 13 Townsfolk characters on the sheet, but the GM will only select 5 to include. Although the game mandates the presence of 1 Minion, there are fully four possibilities for what specific Minion character that could be.
In our actual gameplay experience, we found ourselves endlessly tangled in the problem of “deducing the Minion’s character.” We discovered that two players claiming Townsfolk roles had contradictory information—if this were Werewolf, it would be obvious that a werewolf was masquerading as a good player and lying. However, in BotC, there are far more possibilities: First, the Minion could be the Poisoner, who poisoned a Townsfolk (game mechanic: a poisoned player receives false information). Second, the Minion could be the Baron, who added a Drunk to the game (game mechanic: the Drunk thinks they are a specific Townsfolk character, but they are not, and the information they receive from using their ability is false). Third, one of the “Townsfolk” was a Minion or the Demon in disguise, intentionally feeding false information.
Reading this description of the experience, one might critique: “There is an information overload, too many possibilities, the game is too complex, this is not good design.” However, while immersed in it, I didn’t feel it was overly complex; instead, I actively threw myself into the deduction. I believe this mechanism of “information overload” actually enriches the social dynamic: First, every player has an ability and possesses their own unique information. Therefore, every player has the opportunity to be a “one-minute hero,” giving them sufficient motivation to share information, and giving others good reason to listen. Second, because there is so much information and so many possibilities, the difficulty of “synthesizing all information for a perfect deduction” is much higher than in games like Werewolf, but the barrier to “synthesizing partial information to propose a possibility” is much lower. This encourages more people to present, rather than having “one player lead the whole room.” Finally, on the path to the ultimate goal of “finding the Demon,” there are many mini-goals, such as “identifying the Minion’s character” or “identifying the Outsider’s character”—completing each mini-goal gives players a sense of achievement, filling the game’s lengthy hour-plus duration with “micro-loops” and “real-time rewards.”
Based on the MDA framework, the mechanics and dynamics described above create three types of aesthetic (fun): challenge, fellowship, and narrative.
Regarding challenge, it mainly stems from the confrontation between the two teams. Because there are many characters and their abilities all lean towards being overpowered, the powerful information-gathering skills of the good team and the potent obfuscation skills of the evil team both put immense pressure on the opposing side. This is a clash between titans, rather than the strong bullying the weak; one could simplistically understand it as “when both sides are overpowered, it creates a form of balance.” Of course, there are risks here: if the GM’s selection creates a distinct imbalance between the two teams’ strengths, the game experience will be mediocre. This is also a key contradiction (and the key solution) in BotC: using the GM to balance the game, from before it starts (character selection) to during the process (information distribution).
Regarding fellowship, this game contains elements of “building meaningful friendships.” Players can chat privately (as shown in the image, sending private messages when playing online, or moving to an adjacent room when playing offline), which creates conditions for “proximity.” Furthermore, as each character exercises their abilities and enjoys their chance to be a “one-minute hero,” they can feel their contribution to the collective, thereby strengthening their sense of belonging (reciprocity).
Figure: Screenshot of playing BotC. Left: chat box; Middle: game interface; Right: character sheet; Bottom: online meeting. Screenshot from Bilibili
Regarding narrative, this is the most subtle part: the GM in BotC is aptly called the “Storyteller.” Although the game itself does not provide a storyline, the interactions among players construct a wonderful story. During the deduction process, players hypothesize various possibilities, forming rich narrative chains in their minds. After the game ends, these possibilities converge into one definitive reality, and players will then deduce backwards to uncover the truth of what just happened, re-evaluating their previous reasoning (whether it was dead-on or laughably wrong). This experience is akin to reading an excellent detective novel.
Returning to our Shrimp Game: our game includes 6 characters, with 4 players participating in each session. Players all receive a character sheet and know their own roles, but they do not know the roles of others or exactly which characters are in play—this design is the same as BotC. We designed a “Detective” character: they can guess the identities of all other players at any time. If entirely correct, they win immediately; if wrong, they must reveal their identity and receive a very difficult victory condition. Combined with the “identity puzzle” inherited from BotC, this character will face their own “hero or clown” moment. It also acts as a Sword of Damocles hanging over every player’s head: if others expose their identities, it’s equivalent to handing over the victory—even if the “Detective” is not actually in play, everyone must still hide and lie.
But one difference is: BotC is a two-team confrontation, whereas The Shrimp Game is a free-for-all. In BotC, there are some pure-information characters (for example, knowing whether two players are on the same team at the start of the game, after which the ability becomes inactive) available for players who are unwilling to deceive and conceal; in The Shrimp Game, however, all characters must continuously hide their identities and guess the identities of others until a single person achieves victory. This amplifies the challenge attribute.
A stronger challenge attribute aligns with our design philosophy. Borrowing The Molecule analysis framework: our target audience is casual board game users. For occasions like parties or social ice-breaking, we provide a solution that is short, exciting, and allows players to moderately communicate and get to know each other. BotC, on the other hand, targets a more hardcore demographic, aimed primarily at accommodating large groups for lengthy social activities. From this perspective, our game should maintain (and even deepen) the characteristics of being short, fast-paced, and supportive of small player counts.
Figure: The Molecule analysis framework
In conclusion, through analyzing BotC, I have learned the potential of the “unknown character pool” mechanic in stimulating players’ social dynamics, and I have also clarified the differentiated design direction for The Shrimp Game. I hope to create a thrilling, fun social deduction game suitable for casual players.
