The Game Theory of Story Endings

Do happy endings really make you as happy if you see them coming a mile away? When we watch a trashy action flick or a fluffy romantic comedy, aren’t the conflicts less interesting because we know it’ll all end happily ever after? Someone has to bite the bullet and write a sad ending to give plausibility to the threat of unhappiness. It’s disincentivized because sad endings are more challenging and risk upsetting the audience, but someone has to do it.

Steven E. Landsburg muses about this in The Armchair Economist:

I am intrigued by the market for movie endings. Movie-goers want two things in an ending: They want it to be happy and they want it to be unpredictable. There is some optimal frequency of sad endings that maintains the right level of suspense. Yet the market might fail to provide enough sad endings.

An individual director who films a sad ending risks short-term losses, as word gets around that the movie is “unsatisfying.” It is true that there are long-term gains, as viewers are kept off their guard for future movies. Unfortunately, most of those gains may be captured by other directors, because movie-goers remember only that the murderer does sometimes catch up with the heroine in the basement, and do not remember that it happens only in movies with particular directors. Under these circumstances, no individual director may be willing to incur costs for his rivals’ benefit.

A solution is for directors to display their names prominently, so that viewers know when a movie was made by someone unpredictable. Viewers, however, may find it in their interests to retaliate by covering their eyes when the director’s name is shown.

If you can be associated more strongly with unpredictability, you reap more benefits. You’re also more strongly associated with the unhappy ending, which might turn audiences away.

One way to ease the blow of an unexpected sad ending is to make deaths triumphant, defiant, or heroic. Think of how Spock died in The Wrath of Khan (No, I’m not going to give a spoiler alert for a 30 year old movie). Sure, people die in Star Trek all the time – when Kirk, Spock, and fresh-faced, red-shirted Ensign Jimmy beam down to explore a planet for life, we all know one of them isn’t going to make it back. But to kill a main character is more significant. And it was done in a touching way. They got the unpredictability without upsetting their audience.

I genuinely respect Joss Whedon for his willingness to throw curve balls like this in his story lines. He’s developed a reputation for having sympathetic characters die, leave, or change sides – often without warning. Rather than watching Buffy, Firefly and Serenity thinking “So, how is it all going to work out this time?” we’re forced to think “Is it going to work out this time?”

TV Tropes has a name for all this – Anyone Can Die:

This is where no one is exempt from being killed, including the main characters (maybe even the hero). The Sacrificial Lamb is often used to establish the writer’s Anyone Can Die cred early on. However, if the Lamb’s death is a one-off with no follow-up, it’s just Killed Off for Real. To really be Anyone Can Die, the work must include multiple deaths, happening at different points in the story. Bonus points if the death is unnecessary and devoid of Heroic Sacrifice.

In game theory situations, reputation plays a large role. TV Tropes mentions building a ‘Anyone Can Die’ cred, which can be achieved through repeated interactions. In a TV series or multiple films by the same director, you get a feel for whether the good guys always prevail. But even within a single story, early and repeated signaling can make the remainder of the plot more intense. When a major character is killed off without it being a Heroic Sacrifice, that’s a powerful signal that anything can happen. The musical Into the Woods will always have a special place in my heart for mastering this dynamic.

But there’s another route. Historical dramas can increase society’s perception of “sadness plausibility” without anyone taking a hit for being a downer. Nobody’s going to feel unsatisfied that Titanic, The Great Escape, or Butch Cassidy and the Sundance Kid have sad endings. (Or if they do, they can take it up with reality for writing a depressing script. It’s not easy to keep those separate in our brains; we just get the overall sense that sometimes stories have sad endings. And that perception helps us enjoy all the other movies we watch.

Funding for which arts?

Jesse’s recent post, about how words like “art” don’t have universally agreed-upon, precisely-defined meanings, reminded me of a book I just read by one of my favorite bloggers, economist Tyler Cowen at Marginal Revolution. He’s gotten a lot of media and internet attention in the past couple of months for his e-book The Great Stagnation, but the one I’m referring to is an older volume called Good and Plenty: The Creative Successes of American Arts Funding.

Good and Plenty makes the case that although the U.S. government ostensibly spends little on arts funding, compared to Europe for example, we also have huge de facto, indirect subsidies of art through copyright law and tax policies that encourage private giving. Those indirect subsidies are actually much more conducive to a thriving arts sector than Europe’s centralized funding, Cowen argues.

One of the many things I like about Cowen is his cheeky but serious way of challenging the ostensible reasons people have for their beliefs, or societies have for their institutions or laws. (This is something his friend and colleague Robin Hanson does even more persistently, by the way, over at another of my favorite blogs, Overcoming Bias.) The form of the argument is basically: You claim that the reason you do Y is because of X. But if you really believed X were the case, you would also do Z, which you don’t.

So in addition to the question of how we fund art, Cowen also asks why we’ve chosen to define “art” (for funding purposes, at least) to include certain things but not others. For example:

  • What about fashion? Clothing design exhibits beauty and creativity just like sculpture, and a beautiful piece of clothing arguably has higher positive externalities than a beautiful sculpture (because the owner wears it in public, other people get to appreciate its beauty, unlike a privately owned sculpture which sits at home).
  • What about sports? Many sports showcase stunning displays of grace and power, just like dance. And sports games, and seasons, have a lot in common with drama: protagonists, antagonists, triumph and defeat. “The drama in sports, of course, is real rather than staged, but presumably this should contribute to its aesthetic merit,” Cowen says.
  • What about toys? “Most young children are deeply concerned with matters aesthetic,” Cowen says. “They are fascinated by colors, shapes, sounds, and textures… The toy presents the child with an aesthetic package, so to speak, which the child either loves or rejects. When children they love a toy, they are truly passionate about enjoying its aesthetic qualities.”

Yet our government doesn’t subsidize fashion, or toys, or sports (at least not for artistic reasons; local governments may fund stadiums and so on , but that’s for other reasons, like economic development). So that would seem to cast doubt on the reason people generally give to justify funding the arts — that they provide people with enjoyable and moving aesthetic experiences — since those reasons should also suffice to justify subsidizing things like sports and fashion and toys, which we don’t.

Of course, the big differences between the traditional arts and Cowen’s examples are that the former have (1) more cachet, historically, and (2) less commercial viability. But we still have to explain why we should spend the money to prop up the traditional arts when there are other sources of aesthetic enjoyment that don’t require any propping.

%d bloggers like this: