Designing new characters is both fun and interesting. The scope of Blood on the Clocktower creates many opportunities to play around with the mechanisms and themes.

I’ve been fortunate enough to have a number of my character designs go official, and I’ve been homebrewing Clocktower since 2017. I thought I’d take this opportunity to share some key principles behind character design, which have been shared with me by the designer.

Because these design principles were used as minimum standards for a character to be made official, they might not be relevant to you trying to do something very niche and experimental. If you’re only working in digital spaces and creating unique characters for unique scripts, you can probably ignore or bend some of these rules.

However, even if you want to break these rules, they remain really good guidelines on how to make sure your character is well designed.

Publishing Constraints

The primary reason behind these principles is parsimony. If multiple characters overlap, players gravitate to the better of the two, making one redundant. If two characters serve a similar function, it is often just better to use the better version.

Fewer but better characters is a practical publishing decision. Each unique character requires one or more physical tokens, potentially reminder markers, and a system of storage. The more characters, the more sophisticated and expensive any system for organising characters becomes.

I have toyed around with a set of approximately 300 different characters at one point. Not only was it tedious to organise and keep together, but it becomes time consuming to look through an archive to find the set of characters you want to run a script, only to put tokens back after the game and maybe fish out a new set of characters for the next script.

One of original design constraints by the designer ensured that no character exists in more than one official edition at a time. This way, there is no need for more than one copy of a character, and characters can be stored in edition boxes for easier set up and play.

Design by Constraint

The three rules for making an official character are intended to produce the maximum amount of value from the smallest set of tokens. This is in part a publishing constraint, but it is also a solid game design constraint. The more characters that are in existence, the greater the learning curve to master the game as a whole.

That being said, homebrew can be free from some of these constraints: character texts don’t need to adhere to a set maximum length; or they aren’t limited by the impracticalities of a large number of physical components required.

So by all means, in the context of making niche characters for a niche purpose in your homebrew script, do whatever works for the context of that script. You’re using an entirely different set of rules. However, I still think understanding the purpose of these design rules will help you challenge yourself as to the reason behind a homebrew character and potentially make it better.

Three Rules

At the moment, the odds of a genuinely novel and fun character idea emerging that isn’t already covered by some other character is extremely low. So the rate of new official characters has hit a strong plateau. The designer has stated that in order for any character to be considered as an official status it must pass the following three criteria:

Unique

The most important criterion for any character to be considered official is that it is a unique concept. It does something unlike other characters. It is incredibly easy to produce characters that are nothing more than permutations of existing characters. More specifically, the question is not whether their core mechanic is similar, but their core *function* is similar.

A really good example is Drunk vs Puzzlemaster: both have the core mechanic of one Townsfolk is secretly drunk. However, the core premise of the Drunk is for players to doubt their information in the absence of an Outsider. The core premise of the Puzzlemaster is for players to doubt their information by virtue of an specific Outsider claim. Both have the same core mechanic, but they come about in entirely different ways with different player choices.

As another example, I’ve made the Confectioner below as a homebrew character. It shares a lot of territory with the Courtier (a Townsfolk that poisons players). However, the decision matrix of the Courtier is trying to guess what kind of character is in play, and seeing if they’re right. The decision matrix of the Confectioner is instead trying to guess which player is evil (or Outsider), and seeing if they’re right.

Likewise, the Bug Eater as below shares some ground with the Snake Charmer. They both have a way of learning who the Demon is, and pay a price when they guess right. When the Snake Charmer gets it right, there is immediately a tell and the whole town responds to this. When the Bug Eater gets it right, they change teams and there is no tell. So the town has a problem of whether to believe their information or not.

Bluffable

This is a mechanical constraint for good characters only, which means they can be used as bluffs by the evil team. Mostly this means that a good player’s ability cannot be fool-proof in and of itself. Of all the restrictions, this is probably the one that really needs to remain in place for designing any character.

For the most part, this largely relates to characters that can have an observable mechanical impact on the game. Information characters don’t typically have a means of proving themselves except through their information.

An Empath or a Chef is easily simulated by an evil player as they have the information needed to simulate that info. A Saint is bluffable because the town has to risk the entire game on disproving their claim (and as long as there is poisoning on the script, it’s not fool-proof).

Of course, there are always going to be weird edge cases where a character would normally be bluffable, is not. A good example is the Fool in a Riot game, because there are no deaths at night, there is no hidden death to explain why a Fool’s second life doesn’t happen. It’s not the best, but it’s not completely a problem as Riot can afford to lose at least one demon to a bad bluff.

Side note: if your character has a cool, but unbluffable ability, then it’s probably better as a Traveller.

Fun

This is the hardest to gauge without at least playtesting the character. There are a lot of mechanically possible characters, but they’re only really useful to the game if they have a fun interaction with the game. A character whose only ability is they can never be drunk/poisoned is mechanically fine, but what does that character do?

Often the best way to approach this is to try and boil the character concept down to its core idea, and then build the mechanisms around it. I find this is also a useful exercise to help you eliminate extra parts to a possible character than can overcomplicate it. Generally, characters with a single coherent idea, with few moving parts, have a good balance of fun and mechanical elegance.

There is no hard and fast rule about what makes a character fun, and sometimes it really only becomes apparent when put into play. There are anecdotal stories from the early playtest of Clocktower, where a lot of players found the Soldier as a passive character to not be fun and unengaging. This is because even though it’s a simple idea, what you can do with the character isn’t always immediately obvious.