What makes an interesting game?
Tough question. In the book A Theory of Fun Raph Koster writes that a game is interesting when we can recognize a pattern. The pattern is the collection of constraints and challenges that we face while we play. If the game pattern is too obvious, the design ends up with being boring and unnecessary. On the other hand, if the pattern is not intelligible, we cannot act intentionally and the game proves itself frustrating.
A game is intriguing when we understand its rules or laws but this understanding does not result in a trivial experience. Rather, its ruleset generates an environment where we see the obstacles coming and the attempt of overcoming them rewards us.
At Tablescope we develop digital hybrid games that rely on cutting edge computer vision technology. Each iteration requires us to act on a complex machinery. The concept of game pattern allows to direct and optimize this process.
As an exercise, I analyse the evolution of my personal understanding of the patterns in the videogame Lumberjack, a very simple design playable within a Telegram chat.
Evolution of game patterns in Lumberjack
Lumberjack does not come with a tutorial.
When I start the game I see the trunk of a tree in the middle of the screen. The lumberjack shaped avatar with his hat, plaid shirt and axe in hand waits on one side of the plant. At the boundary of the game area two arrows point respectively to the left and to the right.
I push the right arrow key.
The axe hits the tree and takes a piece of it away. I push the left arrow key. The lumberjack moves to the other side and takes down the next chunk. Each cut increases the score by one point. Meanwhile, the upper part of the tree shifts towards the ground after each cut.
The tree does not seem to be provided with a top: up to losing, the game is infinite.
On the other hand, the plant has got branches. Just to be on the safe side (ahah), I try to avoid cutting at branches. There seems always to be at least one side with no branches.
After some ten logs, I realize that right over the score there is a green bar, which gets shorter over time while recovering some length after each cut. I assume that using up that indicator is not a good idea: I can’t just take my time.
Pressure, this find puts me under pressure.
Right, left, left, BRANCH. The green bar is still full but the game ended when I tried to cut the tree at a branch. Cutting on the branch side ends the game.
I play a couple of times more without worrying about reorganizing the discoveries I have made about the game patterns. At this point, my approach is to decide instance by instance whether I want to press the right arrow or the left arrow, while dodging branches and being as fast as possible.
At the fourth run I start to cluster game actions.
If on the right side I see a sequence of three logs to cut away, I go with pushing three times the right arrow key. Then I move to the left and press the correspondent key the right amount of times. The elementary action still consists in pushing either right or left but now I group them in one or more pressures.
In the meantime I realize that every twentieth cut the level indicator increases by one. This does not modify any aspect of the game but the green bar on top, which now decreases faster. The more points, the faster one must play.
The theoretical find
It looks like the minimal cluster of actions always consists of at least two pressures on the same key. I verify the hypothesis in the eigth game, validating it. I cannot cut a log if on that side a branch is hanging upon the avatar. Besides, I cannot switch from left to right or vice versa if a branch is already at the ground level on the other side. The game must provide a side where it is possible to cut at least twice! (I knew that my PhD in math would come handy.)
The elementary action consists now of two or more pushes on the same key.
After a bunch of games, I start to automate the execution of action clusters.
While my finger pushes four times the right arrow key, my eyes check how many times the left key is going to be pressed next.
At the thirtieth game, clusters of actions turn into clusters of clusters.
All the information on the screen gets interpreted at once and employed to generate a left/right sequence. I both execute and update the sequence depending on the configuration of the branches coming down from the top of the screen.
After scoring 364 points in a game that ends because the green bar on top is exhausted, I feel satisfied and stop playing.
Game (patterns) over?
Yes and no. The agonistic mindset would start at this point. Training, posture, fluidity in the movements, speed engineered to minimize risk while keeping the green bar full, and so on.
Again, if Lumberjack actually aimed for longevity then we should expect more complex patterns to emerge while being exposed to its mechanics for a longer time. In chess this would be like developing the theory of openings.
Conclusions and related articles
Understanding and shaping game patterns is a crucial step when it comes to invent a new game.
Next time you put hands on your prototype, spend some time on identifying and dissecting its game patterns 😉 Which patterns are primary and which ones are secondary? Which patterns emerge in the first game and which ones in the tenth?
Finally, if you want to find out how certain games can produce frustrating patterns (and learn how to handle them with a zen approach), have a look at Four things learned by playing Magic: The Gathering.