On <em>The Archer's Flight</em>

On The Archer's Flight


Mark Keavney describes his process in composing a story in which the readers voted on plot points as he was writing, resulting in a truly interactive fiction - a narrative in which, as Keavney puts it, “[n]either the players nor I owned the story completely.”

For most of my life, role-playing games have been my passion. Nothing I knew had the same mix of spontaneity, mythic storytelling, and especially creative collaboration - the sense that together my fellow players and I were creating a story that no one of us could have created on his own. For over twenty years, I played and ran role-playing games, attended game conventions, and even wrote gaming magazine articles and game supplements.

The more I played, the more sophisticated my games became, with the emphasis shifting from killing monsters and gathering treasure to developing characters and telling a story. My last game, for example, was a quest for spiritual truth that explored themes of self-discovery, religion, and the nature of death.

But the more ambitious my games grew, the more frustrated I became with the form. For me, creating and playing these games was an art, but due to the nature of the form, that art could only be shared with a few people.

In 2002, I launched the web site City of IF to bring these kinds of interactive stories to a larger audience. I started by publishing the first chapter of a story called The Archer’s Flight. It described the character of Deica, a human girl raised in a village of centaurs, who grew up an outcast because of her strange form.

I told this story using a form of interactive storytelling that I now call “storygaming.” The story had seventeen chapters, each published about two weeks apart. Each chapter ended in a “decision point” for Deica, and each decision was made collectively by the readers (i.e., players) rather than by me. Neither the players nor I owned the story completely: the players chose Deica’s actions without knowing how they’d turn out, and I wrote how they turned out without choosing the last action (or knowing the next).

The way that the players made decisions was structured: first there was a “posting phase” in which they posted and discussed their suggestions on a web forum, and then a “voting phase” in which they voted on the most common suggestions. The winning suggestion became the basis of the next chapter.

This experiment gave me “nothing I had asked for, and everything I had hoped for,” as the saying goes. My first lesson was at the end of chapter three.

The first few chapters of The Archer’s Flight had covered Deica’s teenage years as she secretly learned archery from her grandfather, up to a point when a dragon threatened her village. In chapter three, the villagers beseeched a local hero for help; Deica’s grandfather and the other village elders were meeting with him the next day.

I was looking forward to the suggestions for this chapter, because the situation offered a good dramatic conflict and a rich set of options. Deica was proficient with the bow and could help against the dragon, but the centaurs wouldn’t normally let a girl (much less a “deformed” one) take part in a battle like this. Would she plead with her grandfather to stand up for her? Crash the meeting and show the centaurs how well she could shoot? Secretly tag along when they went hunting the dragon? I had already thought about some different ways the centaurs might react and was planning ahead to the finale, when Deica would join the battle against the dragon and (presumably) fire the decisive shot.

Some of the players did make the kind of suggestions I’d expected (“If I can show them what I can do, they’ll let me come,” posted one). But then one suggested that Deica disguise herself as a boy and run off to the big city. In the “voting phase,” the players voted for this option.

The choice was completely appropriate: Deica was adventurous, she felt alienated from the village, and due to a childhood experience, she was afraid of dragons. But it was also completely unexpected. Suddenly I had to write the next chapter about Deica’s travels to the big city, and I had no idea what the big city was like, what she might do there, or how I could possibly tie this back to the huge loose end of the dragon attacking her village. All my plans for the story were going awry.

And I loved it - because this was what I’d experienced in role-playing games: that no one knew where the story was going, but we were going there together. I happily cast aside my plans and started making new ones.

So, chapter four became the beginning of Deica’s travels, and the story continued. On her way to the big city Deica got a job with a minotaur merchant; when they arrived she was almost lynched because of her strange two-legged body, and eventually she was “adopted” by a double-bodied man who called her a legendary “Single” and tried to make her the figurehead ruler of his country. At each point, the players chose Deica’s action (e.g., the decision point in chapter ten was whether to travel with the minotaur merchant or the double-bodied man), and I revised the plot to fit those decisions. Finally, Deica escaped the double-bodied people; on her own with a horse and money for the first time, she resolved to return to her village to solve the mystery of her form. Here, I put the story on hiatus.

This way of telling stories has a family resemblance to other forms of interactive storytelling, such as shared-author stories or role-playing games. But it also has other interesting features that became apparent as The Archer’s Flight progressed:

Open-ended interactivity. Players could (and did) take the story beyond what the author had imagined. I had set out to write a fairly straightforward fantasy about a battle with a dragon, and the players’ choices turned it into a (more interesting, I think) fantastic journey.

Flexibility. The story had a sliding scale of interactivity - for any chapter, a player could just read the story, could vote without making suggestions, or could post one or multiple suggestions. Players used this flexibility often: for example, for most of the chapters there were five to ten players posting, but often it was not the same ones each chapter. Also, there were typically three to four times as many votes as posts, and many times more page views than votes (see figure 32.1).

Social interaction. The format provided rich soil for social interaction: players could agree or disagree with each other, build on suggestions, or try to sway votes. For reasons beyond the scope of this space to discuss, there wasn’t as much social interaction in The Archer’s Flight as in other storygames later told on the same web site, but there was some back-and-forth discussion of options.

A strong narrative. The Archer’s Flight was guided from beginning to end by one author, and the ultimate result was a coherent story told in a single voice. That story was published as a novella in May 2005.

Through storygames like The Archer’s Flight, storygaming has proved compelling and its popularity has grown among both gamers and writers. The City of IF web site now has over 1,000 visitors every day playing dozens of storygames in fantasy, science fiction, mystery, humor, and other genres.


Keavney, Mark (2005). The Archer’s Flight, City of IF (May 2005).