As part of our support for the Choice of Games Contest for Interactive Novels, we will be posting an irregular series of blog posts discussing important design and writing criteria for games.  We hope that these can both provide guidance for people participating in the Contest and also help people understand how we think about questions of game design and some best practices.  These don’t modify the evaluation criteria for the Contest, and (except as noted) participants are not required to conform to our recommendations–but it’s probably a good idea to listen when judges tell you what they’re looking for.  In today’s entry, I’m going to be discussing our thoughts on how to design the end of a game in a way that makes choices meaningful and interesting.

If these topics interest you, be sure to sign up for our contest mailing list below! We’ll post more of our thoughts on game design leading up to the contest deadline on January 31, 2018.

When the Choice of Games staff looks at an interactive fiction game, we always focus some attention on the end of the game–what we sometimes talk about as the game’s victory design. What makes interactive fiction different from plain-old fiction–novels and short stories–is that the player makes choices that matter. Those choices should matter throughout the game, but they absolutely need to matter at the end of the game. If the game always ends the same way, then the choices along the way can’t be very meaningful.

We’ve identified two ways to design the end of a game that ensure that the choices along the way are meaningful and significant. The first is to create multiple, independent goals for the player, all of which are significant for the end of the game. The second is a structural design we refer to as the arm-and-fingers structure.

Multiple, Independent Goals

One of the pitfalls in interactive fiction design is to structure the whole game as an answer to the question, “does the player win?” For example, imagine that a game is about leading a heroic rebellion against an evil empire, and that it has two endings: either the player wins and overthrows the evil empire, or the player loses and the empire continues to oppress people. As soon as the game designer adopts that structure, the choices in the game become much less meaningful. They can still be tactically difficult and interesting–which of choices A, B, and C has the best chance of producing a winning outcome?–but fundamentally, the game can tell only two stories.

A variant on this problematic design is to have multiple different goals, but to make the goals overlap so they are not independent. Continuing with our rebellion example, perhaps the heroic rebel has several goals: recruiting the allies you need to overthrow the empire; developing the military strength to fight off the empire’s troops; and becoming a hero who’s remembered in legend after the victory. At first glance, this seems to solve the problem of only having a single goal, but in reality, the problem persists. If you overthrow the empire, you either necessarily or likely become a hero who’s remembered in legend. And if you fail to recruit needed allies, you can’t overthrow the empire. There might be a little differentiation–maybe it’s possible to win without allies, just difficult–but the basic problem remains: the player either wins or loses.

The solution to this problem is to make the different goals independent and indeed often in tension with each other. My partner Jason Hill uses the example of the different goals a student might have in college. A college student’s goals might include: getting good grades and graduating with honors; lining up a good job after college; winning the big game in their sport against the rival school; keeping their student job and earning enough money to pay tuition and other expenses; going to parties and having a fun time; building friendships that will last; and having a satisfying relationship with a significant other. That’s a mix of different goals, all of which could be interesting in an interactive fiction game about attending college.

In order for this approach to work, the goals need to be independent. Getting good grades has nothing to do with winning the big game–a player could achieve one goal and not the other, achieve the other goal, achieve both, or not achieve either, and that variety of outcomes means that choices along the way have room and scope to be meaningful. The player’s choices can tell different stories, whether that’s the story of a jock who wins big but fails out of school or of a student-athlete who leaves the team entirely to concentrate on their studies. And precisely because of those tensions, the victory design sets up interesting choices along the way: should the player character head to the gym for more practice, study up for the test, spend time with their S.O., or concentrate on an internship that might lead to a good job? There isn’t enough time in the day to pursue every goal fully, and the player has to decide which goals to prioritize and how.

The goals don’t have to be completely independent in order for this design to work well. For example, earning good grades can help to get a good job offer, and being a star athlete can result in invitations to exclusive parties and impress some other students. And some underlying characteristics and approaches can help with multiple goals: a smart character will do better at both schoolwork and jobs, and a charismatic character will find it easier to get job offers as well as easier to make friends. But the player will still need to decide what goals to prioritize and how. When your S.O. has a problem the night before a big exam, which wins out: studying or helping your S.O.? When your coach wants you to practice and your boss wants you to meet your work schedule and there’s homework to be done, how do you spend your scarce time? If you try to burn the candle at both ends and just cut back on sleep, will you cause everything to fall apart? These choices are conflicting–you can’t pursue all of the goals fully–yet combinable so that the player can choose to pursue any two of them.

By designing goals that are independent, a game designer gives the player’s choices space to be meaningful.

Arm-and-fingers structure

An important technique for ensuring that player choices have a major impact on the conclusion of an interactive novel is to have a major branch point late in the game with several different possible final chapters–what we refer to as an “arm-and-fingers” structure. To understand why that structure is valuable and how it works, it’s easiest to start by considering the problem that it solves.

One of the traditional tensions in interactive fiction design is between introducing branches and forcing the storyline to largely progress in a fixed way. Some of the earliest examples of interactive fiction relied heavily on branching, which cause player choice to have a larger effect on the game but also require much more writing and shorten the length of each playthrough. If every choice produces another branch point, than a game that has 10 choices each with 3 options ends up with 3^10 different branches–nearly 60,000 branches! As a result, some of the old “choose a path” books were a hundred or two hundred pages long, but with average playthrough lengths of 5 or 6 pages. The other extreme from branching on every choice is to make the game design linear, with each choice leading to the next in lock-step, which makes the game design and writing process manageable and feasible but can make multiple playthroughs feel repetitious.

Choice of Games recommends in general designing games as a stack of bushes. Each scene has branches, but the branches merge back together at the end of a scene. Then, variables and delayed branching can be used to make choices remain meaningful beyond a given scene. A stack-of-bushes with delayed branching is a good basic technique, but it can still feel frustrating when every playthrough of a game ends with the same climax.

An arm-and-fingers structure is a game with several different final chapters where the player’s decisions determine which final chapter they experience on a given playthrough. Most of the game is the arm, with chapter leading to chapter more or less automatically, but the structure of the end of a game is like a hand, with entirely distinct and separate fingers branching off in each direction. Kevin Gold pioneered this structure in Choice of Robots to great effect, and Lynnea Glasser also used it well in The Sea Eternal. It can maintain a manageable structure that does not require writing thousands of different branches, while still creating the feeling that the end of the game depends on the player’s choices, not just in determining a final outcome, but in determining the entire feeling and plot of the game’s climax.

By introducing a major branch point before the last chapter, the arm-and-fingers structure underscores the importance of the players’ choices. Not only does the outcome of the final conflicts of the game change, the nature of the climactic conflict changes as well. As an example, imagine a fantasy game in which the player plays the heir-apparent to a monarchy. Depending on the player’s choices, the final chapter could be one of four entirely different choices. If the player built a strong base of support in their court and among the nobles of their country, the final chapter could be a conflict with a neighboring kingdom that could be resolved through warfare or through diplomacy. If the player focused on the study of wizardry, the main character could renounce the throne altogether and pursue true mystic power on a personal quest. If the player fostered new ideas about politics and rights, the last chapter could be about fostering a new more democratic regime and breaking the power of the high nobles. And if the player lost control of their country, the last chapter could be a story of a monarch in exile fighting a civil war to retake power. Every one of those chapters is a satisfying, dramatic conclusion to the story, but replays offer wide variation and the player’s choices have meaningful impact by determining which branch the story goes down (and of course how that branch resolves).

At the same time, the amount of additional writing required is manageable: instead of a 10 chapter game with each chapter following linearly, the last chapter might be replaced by one of 4 possible end chapters, requiring writing 13 chapters total. That’s not a trivial increase in work compared to writing 10 chapters, but it is a far cry from an exponential explosion of different branches. And the pay-off is very substantial, making each playthrough of the game feel very different and making the player’s choices drive the outcome of the game. The arm of the game should be a traditional stack of bushes, and the introduction of a set of separate branches at the end–each with multiple choices and telling a satisfying climactic story–will make all the difference in making the choices feel meaningful and different.

—-
As you think about outlining a ChoiceScript game–and as you think about how to maximize your game’s score on the “conflicting goals with satisfying endings” criterion in the Choice of Games Contest for Interactive Novels–we urge you to design a game with multiple independent goals and to implement an overall arm-and-fingers structure. Neither is strictly required, but if you implement both your game will be miles ahead of a game that has a linear structure with a “do you win?” goal design and maybe a separately tracked measure of romance success. So think about how to make a compelling set of goals, each independent from the others, and then think about how those goals and player choices can drive the game to one of several different climactic chapters.