Abstract University


I've been developing and testing an abstract strategy game for two players called Druids for months now. We've been testing it furiously, almost daily, due in large part to us enjoying the game and it being addictive to play.

I've never designed an abstract before, nor have I honestly played that many. Doing so has taught me many things, and I wanted to take a moment to note my reflections both to share and to think about myself.

You can take a look at Druids here. That link provides you with the rules, a video rules explanation (just a hint out of date), and the Print and Play files. You can print, learn, and be playing the game within 10 minutes. Honestly!

That being said, here are some things I've learned. What's even better about them is that they are lessons for so many designs, not just abstracts.

Status at a glance. When playing Onitama, an excellent game, I noticed how easy it was to quickly take stock of the game. It's very easy to identify threats and opportunities, then cross check them against the available moves. Both players have one big piece, who is central to either victory condition. Their placement instantly reveals much. But also, the graphic designer did an excellent job with a subtle color cue: red moves tend towards the left, blue the right, and green are neutral or middle focused.

Druids has the neat (I think) core mechanism that you cannot move in the direction you're facing. By using triangle shaped playing pieces, it's very easy to see where an opponent, or you, cannot go. If you examine the board below, you'll notice that brown isn't that threatening to tan here -- they can't move forward!


No exceptions allowed. The biggest changes to Druids in the last few weeks have come about to remove exceptions. I'm not an elegance at all cost designer. I respect those who are, but I'm actually fine with appropriate and justified exceptions. But, no exceptions is a key part of the abstract genre. It's an expectation of its audience, and it's one you should respect. Therefore, the bar to allow exceptions is far higher and must be respected.

Some of these exceptions, in hindsight, seem foolish.

  • Previously, 2nd player began the game with a specific form, and three more were placed on the starting board. There are only three spaces to place new ones as they are used and cycled, and it wasn't uncommon to need a place to place a fourth. Therefore, I made it so you didn't cycle the specific form when first used. Which is odd to remember. I simply made it so there were never more than three forms on the board. Exception removed!
  • There used to be Terrain and Animal Forms. Animal Forms begin face down and are cycled, Terrain Forms begin face up and are not cycled. This is a weird exception, and I removed it along with a larger strategic change to the game.
  • Previously, if you move your piece onto an opponent's, their piece is removed (as in checkers, chess, Onitama). But, if you captured a piece with a form, you'd get one of your pieces back...if the Temple was clear. If not, no. So, that's two exceptions. And, one that makes the game last forever.

Be mindful of exceptions in your abstracts, but really, in all games. I'm fine with them, but they need to serve a purpose. You need to be able to justify them, seriously.

Think about the first player problem. This is true of most games, but especially true of abstracts. The first player is effectively always one move ahead of their opponent, which is a major advantage. You need to think about this.

In Druids, we always gave the second player a random Form. However, some were way too good and some weren't good enough. We stumbled across one which felt like a perfect solution: Turtle. The Turtle lets a Druid move in the direction of its facing. This is a beautifully mathematical counter to first player always being one step ahead. It's basically a free bonus move when used.

How you handle the first player problem in your abstract, or any game, is up to you and must be considered in light of your design's nuances and strategy. But, don't overlook this problem or shrug it off. It's a very simple one to test -- how often does first player win?

Incredibly simple goal. Abstracts should have an incredibly simple goal. Let players focus on all the ways to accomplish a task, not the task itself. As so many abstracts are two player affairs, you're able to leverage concepts like player elimination, or first to meet a goal. These are often less fun in games with more than two players, but with two players? Fair game.

In chess, you capture an opponent's king. In Onitama, the same, or moving to the temple. In Druids, I looked to the greats: capture all three, or move to the temple with a form. I say that, then people get it immediately. Then, they can focus on the five animal forms in play. Or, how to maneuver three Druids and the Terrain to pincer their opponent.

As soon as you begin thinking of rounds, and phases, and points, and end game points, you've probably gone too far with your abstract.

Find ways to add variety. I believe modern abstracts have a great opportunity to move the genre forward with more variety. Checkers is a game the computer can solve. Chess has been studied and memorized for centuries. But, adding variety means you must do so within the bounds of the genre. Yes, few exceptions, simple goals, elegance.

Onitama does it well with having five of 15 move cards every game. Druids borrows this idea using four unique animal forms every game. Neuroshima Hex does it with different armies, and Theseus does it by adding new factions, but also varying a player's cards every game.

This is an area I've really tried to add some freshness for Druids, so I'm going to walkthrough a few more elements.

  • During setup, two randomly chosen, face down forms are in play. The other forms are cycled in randomly once the current ones are used. This can really change the game, even if you select the same four Forms every game.
  • During setup, four terrain pieces (two clouds, a forest, a mountain) are placed by the players.
  • During the game, the terrain pieces can be moved when players move onto specific spaces. This is a great layer that's not very complex.
  • The forms enter play face down, so only their owner knows what they are. Players can choose to reveal them, however, for a benefit, but also, the ability to affect their opponent's behavior if they know the outcome of a move.

Players greatly value variety. Finding a way to do so for your abstract is very important.

Be intuitive. Your core movement or turn choices should just make sense. In the same way you need to be careful how you use game terms like shuffle, hand, and discard, you need to make your piece movement, or placement, very intuitive and simple. Lean on classic spatial relationships first and foremost for your abstract.

For other mechanisms, look to nature, history, or other concepts to drive home your mechanisms and make them more intuitive. For Druids, I thought about the fantasy classic shapeshifting Druids. When I say "You gain new animal forms," many players are with me. then, I tried to make sure the animal behaviors match the form mechanism. The eagle can move around the board rapidly, like flight. The rhino charges forward until it hits a wall. The snake slithers along orthogonally. The rat plagues the opponent by removing its form. The frog jumps from point to point.

I did this with Terrain as well. I chose their design based on the piece, and again, people get it. Mountain blocks. You cannot move through it. Forest hides. You cannot be captured inside it. And clouds whisk you away. They let you move between them.

It's subtle, and it's tough to prove or gauge, but abstracts can really benefit from intuitive mechanisms.

Add layers, not complexity. When adding new things, try to add new clean, elegant layers that neatly combine and complement each other. Try to avoid complexity.

With the Terrain in Druids, I was worried it would greatly complicate things. But, I kept it very simple. If you move onto a Terrain symbol, you move the associated Terrain. People see mountain symbol, and move the mountain. They were already used to moving onto a form and getting something, so this tied in nicely.

The entire game is about positioning and movement, and all of the Terrain affects precisely that.

The Terrain helps break, or create stalemates and moves the game forward. People immediately pick up that clouds let you outflank and opponent, or move to their Temple quickly to score. They get that mountains help you block opponents into corners. They get that Forests let their Druids move near an enemy in safety.

My point is that this addition to the game is dead simple, intuitive, and does not hinder the flow of the game or add too much more rules overhead. It's a new layer, something players must consider, and it fits nicely.

If you're looking to design checkers plus, or chess with a twist, remember, a layer, not an automobile engine.

Brevity makes mistakes more palatable. The shorter the game, the more okay it is for players to screw up and rethink their next game, or be the victim of just tough luck. This is true of randomness. The shorter your game, the more acceptable it is for luck to drive the outcome of the experience. This is why Monopoly and Munchkin garner so many eye rolls from players -- they aren't short experiences, and luck plays a large factor in the experience.

If your game has simple goals, few exceptions, and elegance, then decisions should be decisive. They should have clear outcomes. If your game is brief, it'll make these decisive moments more palatable for the loser.

Which leads me to...

Move towards a conclusion. Your game should always be moving towards a conclusion. Do not let your design get to a position of spinning and repeat. A few weeks ago Druids could go this way at times, and I knew it had to be solved.

In checkers, normal pieces cannot move backwards, which slowly forces captures and progress. In chess, pawns cannot move backwards, which has a similar affect. In Onitama, playing chicken and not losing pieces is fine, except you stopped paying attention to their piece rushing towards your temple!

Your game should always move towards a conclusion, or a decisive move. And per the previous point, if it's relatively quick, people won't mind that one mistake costs them the game! Though, if possible, it's nice sometimes when mistakes are actually feints...

Your game should hasten towards a conclusion. It should end, whether the players want it to or not. Build your mechanisms accordingly!

What did I miss in explaining these lessons? What else do you think is key?

Join in via the comments.