Set Balance

The majority of games that use decks of cards do not use every card in every game. In a game of Settlers of Catan, the progress deck contains 25 cards, but in the average game only half of that will be used. At the start of each game, as the deck is shuffled, the invisible hand of fate makes decisions like “For this game only, Year of Plenty will not exist”.


So what does any of this have to do with The People’s Orchestra project?

The idea that I’m trying to introduce is that rather than thinking of a game in terms of “This game has such and such amount of stuff”, it can be conceptualised as “At the start of each play you randomise what the contents of this game are”. The importance of that is that in designing the game as a whole, it’s necessary to consider what a player might experience if different elements are included or eliminated by the invisible hand of fate.

This effect is more pronounced the more impactful the random deck is and the smaller the set of cards that are drawn from it over the course of the game. That means that it has an absolutely massive effect in Shenanigans: The Musical, as the card draws determine what is possible in each round and you’ll draw less than a dozen cards from a set of over fifty.

In playtesting we occasionally run into what I’ve started thinking of as a “degenerate set”. That is to say that we run into a set of abilities that practically guarantee an outcome, regardless of the behaviour of the players. For instance it’s possible, but unlikely, for a game to be populated entirely by players who can look at face down cards and win if the bad guy is caught – in this situation it’s hard to imagine how they might escape.


Obviously this is undesirable. The game is at its best when everyone has a chance to win and players can come up with ruses that will allow them to push the odds in their favour. Presently these situations are rare and the game has a short enough playtime that when they emerge they tend to end quickly. So I would describe the situation as “tolerable”, but it would be better if the possibility of degenerate sets were eliminated completely.

The most common approach to this problem in social deduction games is to ask players to eliminate it manually. Games like Werewolf have players select which roles will be in play at the start of the game and provide guidance in how to choose a set of roles that’ll lead to a fun game. Generally degenerate sets exist, but most beginners will play sets suggested by the rules (and avoid them) and most experts can pick sets of cards that lead to fun play (and avoid them). The degenerate set problem is solved and play can be focused on the most entertaining sets.

This is a good solution for most games, but it doesn’t really suit Shenanigans. We have many more roles that most games of this type, so solving the problem manually will take proportionally longer. It benefits a lot from being a quick game, so requiring a lengthy consideration period between rounds to change the sets hurts it. Also – unlike other similar games – it doesn’t benefit from players knowing which roles are in the game. Each player will probably see two or more cards by the end of the game and need to combine information to find the artiste, being able to lie about which roles are in the game and generate uncertainty about which roles exist is a part of the experience.


An alternative solution is desirable. Editing the available cards seems like a red herring, if the cards were made so inoffensive that no combination of four of them could create a degenerate game it’s likely that the game would be so boring that nobody would care.

Limiting the set up offers some possibilities. Cards are already sorted into those that win by finding the artiste, those that win if he escapes and those with another type of victory condition. Perhaps enforcing a particular ratio between these will create a version of the game in which degenerate sets are much rarer or absent all together.

As always I’m writing this update on a Wednesday and have the opportunity to recruit playtesters for a few dozen games. So rather than theorising we can give it a go and see…



Hmm, so we had a shortage of people in the office today which meant playtesting at the minimum player count: Four.  It turns out that with the best will in the world there is no way to set a ratio between the card types that works at that player count, as you always have the manager, the artiste and two others  – but there are three other types of card!

We tried combining the “wins with the artiste” deck with the “alternate win condition” deck and taking one from the combined deck and one from the “wins with manager” deck. That didn’t cut out the degenerate sets though, the problem was that some of the ‘alternate win conditions’ may as well be ‘manager win conditions’.

Some characters have goals like “Be the first to correctly identify the artiste”. That’s a valid win condition and adds some gameplay in having someone who’ll form a rushed conclusion to make sure that they’re the first to say it – but still means that the game can wind up skewed by situations in which everyone is against the artiste.


Really to help with avoiding degenerate sets it’d be better for the cards to be categorised in to one of two sets based on whether they win with the manager more than half of the time” or not. The problem with that is that the existing categorisation is already pretty helpful in a different way, players like being able to tell at a glance whether they have a non-standard victory condition or not.

Two solutions present themselves. One is to raise the minimum player count to five and figure ratios that don’t involve having to shuffle the “artiste friendly” and “nonstandard wins” together. The second is to add some sort of iconography making it possible to split the roles based on their likely impact on the game to make it easy to create non-degenerate sets by saying “Draw evenly from each deck, take the extra card from the unfriendly deck in a tie”.

Of course there is also the third option of doing nothing. The game is already testing very well and the reviewer who’s seen it was very positive so it’s already doing very well, but I think it’s a curse of game designers to always be looking for ways to make something better than it is. I’m sure that this game can be improved in this way, I’ve just got to hit on the right idea to do it.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>