Last week, I talked about going from idea to prototype. Today, I want to talk about one of the pitfalls many of us fall into when prototyping.

I’m a huge proponent of variable setups, missions, objectives, factions, and characters. I feel like it’s the one thing all of my games will have in common: “20 different powers, only use 3 each game!”

But variability should be the last step when designing your game. To start, you need to make sure your game structure is solid; then, when your game is strong enough to handle it, you can start adding variables to it.

Believe me, I know all about the temptation of adding that variability early. My first game idea was a game about breeding monsters: I had probably 30 monster cards, and I wanted to have a card for a unique offspring for each of those pairs. And, of course, I wanted to have multiple of each, so you never quite knew, when mixing a dragon and a unicorn, which kind of dragicorn you’d be getting!

Yeah, it was the weirdest idea.

But it never saw the light of day. I worked on it for months, never got close to having it ready for a playtest, and I didn’t want to playtest it without finishing it. I was figuring stuff out, making dozens of cards, making diagrams; I felt like I was making progress. In a way, I was, but in a much truer way… I really wasn’t.

Why is adding variability too soon a bad idea?

There are 3 main reasons why it’s a bad idea to start with your game’s content, rather than its structure.

First, any change in the game’s structure will require changes in the elements built upon that structure. Imagine changing the structure of a game like Dominion. After a few playtests, you decide to remove the limit of actions per turn. Without even entering into the issue of game balance, how many cards suddenly don’t even make sense with the new ruleset? All of the Village cards, or the “cantrip” cards that give +1 Card and +1 Action ⁠— these cards were specifically designed  to overcome that limit, and would no longer have a real purpose.

The second reason is that creating content is a time-consuming task which does not really have an end state. Creating content is fun, it’s easy, and it feels like progress, but too much can be a trap. Even if all of your ideas would make it into the final product, you will always have more ideas for new cards, new tiles, new characters, new missions. With a Smile & a Gun is being printed right now, and I still come up with a Shadow card idea every week. If you wait until all of your ideas are in the game before you start testing, you will simply never start testing; each idea you write down will give you another one, and then another.

Finally, content becomes much easier to create after you’ve played the game a few times: you have a better feel for what would be interesting, you can poll your testers for their opinions, and you probably have a pile of mechanisms which did not fit… but maybe could become a card!

But don’t I need content?

Of course! I’m not saying “playtest your card game with blank pieces of cardboard”. Although… if you have a specific mechanism for, say, acquiring or discarding cards, it’s always a good idea to take a standard deck of cards, and try it out.

You can’t playtest your deckbuilder without actually having any cards. Here is a rule of thumb: determine the minimum amount of different elements you need to have to be able to play a single game. If you were making your own version of Dominion, you wouldn’t need to have 25 different action cards designed before you start playtesting; every game only uses 10.

Once you’ve identified that minimum number, make half of it.

No, I’m not kidding. For a first test, you want to have as few variables as you can. If during your test, the game doesn’t click for some reason, you need to be able to identify what causes those issues: it might be the structure itself, or it might be one of the cards that’s throwing the entire game off kilter. Worse yet, it could be an interaction between two cards. Having 5 cards instead of 10 means that not only do you reduce the number of variables which could be affecting the game structure, you’re also reducing the number of possible card interactions from 100 combinations to 25. Sure, 5 is too few for a full game experience, but you’re not aiming for a full game: you are testing whether or not your core idea is fun.

Maybe you can’t actually divide the number by two, but you can just cut down on how different you make these elements. Working on SuPR, a cooperative dice drafting game, I wanted players to roll a pool of dice and determine who would use what. Therefore, I needed players to have different abilities, or else the central conceit would break down. However, for the first playtest, I just changed which action was paired with which: it was very minor asymmetry, and a lot less than what I needed to make the dice selection shine, but at least it made me able to work on the bigger picture.

But I like making content!

Creating content for your game is not only important, it is also fun. It is perfectly okay for you to indulge yourself and create some of that content early. My goal with this article is not to take away your candy, but to make sure you understand that (1) you should not wait for the content to go ahead and playtest; and (2) all the early content you create is unlikely to make it through the game’s development unscathed. 

How early do you usually add this variability to your designs? Have you gotten caught in this pit before? If so, how did you get out of it?