On Non-Linearity and Multiple Endings

In this article which is based on a quote of Berthold Brecht, I argue that some of the games which are called non-linear or games with multiple endings, aren’t really non-linear, nor do they really have multiple endings.

The Shortest Line Between Two Spots: a Curve?

Berthold Brecht says, “The shortest line that connects two spots is, if there is an obstacle inbetween, a curve.” I believe that most games can be summarized like this. They are basically “stories” with one problem/conflict (spot A) that call for one truly desirable solution/ending (but this is an ending that we as players usually experience in the form of an anachronism: either success (spot B) or failure (spot B’)). As a result of the reciprocal influencing of algorithmic procedures and player input, a variety of curves will emerge, which most of the time will result in an undesired B’, until the player learns to withstand the challenges of the game dynamics and manages to reach the desired B.

The Curves of Linear Gameplay

The video game as a medium has a great advantage over Film and TV in that it allows the player to interact with the dynamics that carry the story from A to B’. You have algorithms that can endlessly reproduce the dynamics of a world (and the potential stories that could emerge in it), and you have human players with the desire to understand and achieve, and the ability to learn. So you don’t really need to define every detail of how this interaction between human and machine takes place. You could just define the options that are provided and the overall processes that articulates the chosen options. Then, in the build-up of the game, you’d try your best to make the player adopt the problem (to achieve B) and let her work towards this solution by allowing her to discover the tools and methods to manipulate the dynamics of the game, meanwhile keeping her happy enough to repeatedly send herself through various A–>B’ curves.

This type of interaction or storytelling is something computers and therefore video games are very good at, for Film and TV productions consist of “records of the past” and therefore are to a great extent constructs created through the one-time arrangement of recorded events, which after that are not really futher open to aesthetic or narrative manipulation; while on the other hand the algorithms that manage a game are rather “blueprints of a future” (roadmaps on how things would/could/should unfold, which are yet to be negotiated with the player), and in that sense they are almost predictable but not fully predetermined procedural systems with an agenda of their own, which are however open to manipulation through player choices articulated into this process as input, therefore all this being a reciprocal (or interactive) process of becoming.

Probably the line between A and B’ is shortest when the player does not try to change his algorithmic fate: Then the game will straight go from A to B’ (Just watch how blocks pile up in Tetris). All other situations mean that there is a curve, not a line.

Linearity as Controlled Freedom

Many players, designers and marketing departments call the various “curves” that emerge in a game “non-linear gameplay” and based on this they claim that their story is non-linear, which I think does not reflect the truth. In one of his articles on level design, Cliff Brezinski uses the words “controlled freedom”, which I find a very good description: You seem to be free to make many curves, but then all you actually try to do is to connect A to B (the plot being a controlling force of how we bend the curve in most of our ‘free’ attempts.). We can compare such game stories to a maze with one entrance, and many forking paths that lead to or away from the only exit which we search for (meanwhile facing the danger to get lost on the way, so that we find ourselves frozen to death in the morning… yeah, yeah, The Shining ) They are linear stories in the sense that, there is only one truly desireable line to draw, that between A and B, despite the fact the we musn’t follow it; and indeed, once we accept the role of the player, we most often find ourselves trying to draw the narrowest possible curve around the obstacle, from A to B, and in this our attempts we often end up in a B’.

The Challenge of Multiple Endings

The challenge of multiple endings lies in the fact that each different pair of endings (B and B’) will require a different set-up or plot of its own (the A), because an ending makes only really sense if it relates to the problem that lies at its core. In that sense we must plant the seeds of more than one row-of-crisis + climax in the build-up of the game.

How to Overcome the Challenge?

One method for this would be to use subplots that unfold differently depending on how the main plot or other subplots develop, thereby changing the dramatic direction of the story (altering the nature and scope of the problem) and turning the game towards a different climax or maybe multiple climaxes (maybe a better word for “multiple endings”?). Another way could be to manipulate exposure in a way that results into different knowledge regimes as I go through the game, changing my experience as individual, although the event unfolds the same (For example I might or might not know that there was an agreement with the cops while I play a character in a Mafia game, depending on if I’m a cop or a mafia member)

Designing this maze of mazes is different from designing the pathways of the single maze with its single entrance and exit (A, B, B’). In many of the games based on moral choices, the design effort actually goes into the pathway of a single maze, and not into multiple storylines and multiple endings, but still these games are called non-linear games or games with multiple endings. But multiple storylines and multiple endings is to have multiple mazes (packed into one big maze, or arranging them as overlapping/intersecting mazes, or have them as parallel mazes). The problem now is how to intersect these mazes, if you should do that at all and not instead prefer to make two or more different games.

This multi-maze is different from the single-maze, because it has many points to enter and many points to exit; maybe some parts of the path would need to be walked through in each one of the stories (like if it is the only bridge over a river that runs through the maze) while certain parts of the maze might be not connected to any other path, once you decided to turn into that branch. In some designs it would be difficult to tell which path connects which entrance to which exit.

A story that aims for multiple endings would be served better, so it seems to me, if it has multiple antagonists at once, with their unique motives. This would make it easier to plant the seeds for multiple storylines and enable the writer to weave compelling story-branches out of them. This character-engendered approach could be accompanied by a variety of event-engendered situations, alltogether creating a huge story.

How the Evaluate the Branches

I think stories with multiple branches need evaluation criteria not only on the horizontal, but also on the vertical level. You could have a variety of story-track parts, A, B, C, D, E…. etc etc… but what matters is not just if B logically follows A, or if we can trace back A coming from B. It also matters what emerges through the combination of A and B, i.e if AB propels the act onto a higher level. There could be situations in which E would propel B onto an unexpected height, but not D, of which it is just a repetition. So when we combine many storylines consisting of many story-track parts, which are the track parts that ensure a climbing story when combined and which ones would cause the story to stagnate?Would there be a way in which a narrative engine could evaluate the ‘topos’ of the particular track parts of every story line and direct the player towards those parts that ensure that the plot keeps climbing and away from those that would stagnate story development, depending on the current topos that the player holds?

A last note

One risk you always face while having multiple storylines or many perspectives at once is that the story could lose its focus and go off-track. What if the story just feels like branching away from what has been perceived as the initial conflict? The player would probably think after a while: “What was I trying to solve? Was this my problem at the beginning? What the heck is my goal now? What is this game about?? Am I character X trying to save Y, or am I character Y, trying to save X?”


Games like Tetris (and many of the old coin-op games) have no B at all, and rather follow a proverb of Samuel B’eckett: “You’ve lost. Good. Lose again, lose better.” They are “A–>B’ Only” games, but the B’ can be converted into a B with the help of highscore lists or a hall of fame, which means that performance feedback is presented in the format of agon, so that the player still can compete with others or her past performances. Also each level that we finish in Tetris can be seen as a subgoal, meaning that we achieve a row of succesful A–>B curves, but not in an ultimate sense. We can win a lot of levels until we lose, which is quite different from losing directly. (And the reverse of many FPS type games were we lose, until we win, which is different from winning directly .

Was this article useful? Please leave a comment and share your thoughts.

4 Responses

  1. When did you write this piece? I’d quite like to cite some ideas in here for an essay that I’m currently writing.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: