Four kinds of cruelty.

Moral Incentives and Story Structure

October 12, 2011 - Features / Story Structure / Tragic Ends

There are a lot of ways you can classify the structure of a story, and many of them have been applied to games in one way or another. One that caused some discussion recently is based on a lecture by Kurt Vonnegut in which he describes stories in terms of the fortune of the protagonist over time:

Inspired by this lecture, Paul Sztajer at Throw the Looking Glass wrote a post arguing that game narratives have a tendency to cut the Cinderella story in half by never inflicting ill fortune on the protagonist. By not letting bad things happen to the player, games limit the kinds of stories they can tell.

 So why is this? Part of the answer, perhaps, is that players like to feel powerful, and get angry when power is taken away from them. As Sztajer writes:

Both these problems are symptomatic of a larger issue: that designers don’t like to punish the player. Games are centred around the idea of rewarding the player for playing well, and there is therefore a feeling that if you make the player character worse, you’re punishing the player and they’ll stop playing.

It’s not really an answer I buy, however. Some of the most emotionally engaging moments I’ve ever experienced in games have been moments where power was taken away from me. Based on what I keep hearing about this Aeris character, I’m not the only one. That said, I think there’s something to this idea of people being concerned about punishment and rewards for playing well.

The thing is, I hear players complain about a game punishing them a lot, but not in the same sense that Sztajer is talking about. A good example is the boss fights in Deus Ex: Human Revolution, which were widely considered painful and unfair. This isn’t just about punishing the player. My Twitter’s been ablaze for weeks with people who are just delighted about the brutal, ego-deflating boss battles in Dark SoulsDeus Ex players didn’t complain because they felt punished; they complained because the game encouraged stealthy, nonviolent modes of play outside of boss fights, then forced them into a situation for which their characters weren’t equipped. This is why the battles felt unfair.

This is what really makes players angry: when you set up a system of incentives, then jump without warning into a situation that doesn’t follow that system at all. As I understand it, Dark Souls is punishing from the word go. The game advertises what it is, and players don’t make it to the boss battles if that kind of masochism isn’t their thing.

The same goes for stories. People weren’t furious at the ending of Fallout 3 because their characters died. I mean, they were furious for a lot of reasons. But one of them is that they went from a story that consistently rewarded good behavior to one that killed you for being good, with no transition whatsoever. Contrast with Dragon Age: Origins, which had a similar ending but foreshadowed the everloving hell out of it. That game set up a world where being good often involved sacrifice, if only of your party members’ respect, so the ending didn’t come out of left field. Whatever people thought of Dragon Age, there was no comparable outcry about the fairness of the ending.

I had some of these thoughts in mind following an email exchange with commenter Ari, who really ought to have a blog I can link to in situations like this, yet inexplicably does not. We were talking about character virtues and how designers can track them to make richer stories. After all, moral virtue plays a big part in traditional storytelling. It’s not just that good stuff and bad stuff happens to Cinderella, but that this all happens in contrast to Cinderella’s moral blamelessness. That’s what creates the pathos in her story: no matter how awful things get, she stays a good person. Tragedies, on the other hand, work when a character’s flaws catch up to him. Martyrdom stories work when the character is clearly too good for the world around her.

Therefore, one of the ways you can classify stories is by how their worlds work as moral incentive systems. Vonnegut graphs good and bad fortune against a time axis, and that tells you something about a story. Another graph you can make is good and bad fortune versus good and bad morality on the part of the protagonist: whether he’s rewarded or punished for good behavior.

So let’s say you have this Fortune axis, with good fortune at the top and bad fortune at the bottom. And you have this Morality axis, with evil at the left and good at the right. You can draw all kinds of curves in here that correspond to different moral incentive systems. Here’s one of the simplest:

If there’s a simple linear relationship between morality and fortune, then the better a protagonist acts, the more he’s rewarded for it. If he’s bad, he gets punished. This is the incentive system of the simple morality tales you tell children, where being bad gets you in trouble and you don’t get to go to the party until you apologize. It’s not the basis for a lot of deeply compelling stories, but it’s a starting point.

The inverse of this is a story world where being good gets you punished and being bad gets you rewarded. This is not as common. The only consistent real-world examples I can think of are Lars von Trier’s sadistic melodramas, like Dogville, Breaking the Waves, or Dancer in the Dark, where, for no particular reason, a morally pure woman gets absolutely destroyed for two hours. Some of the goofier versions of fantasy noir, like Sin City, can also fall under this category. Stories set in this moral universe waver between painfully bleak and stupidly hilarious.

Here’s a slightly more complicated curve: a parabola that peaks in the middle of the morality scale. In this world, extreme goodness and extreme badness are punished equally, and only the morally indifferent are rewarded. This is the moral universe of a lot of simple satires, such as South Park, which embody the argument that the morally sanctimonious are just as bad as the damned.

And finally, here’s a morality/fortune graph that should look pretty familiar to gamers. In the inverse parabola, very good and very bad characters are rewarded, while those in the middle are punished. This, or a variation on it, is obviously the moral universe of most games with a morality system. You get bonuses for consistent good or bad behavior, but the ethically ambivalent miss out. Mind you, if you’re bad, you might get a “bad” ending. But a cutscene tacked on to the ending will never have much impact on how a player reads a game, if it’s inconsistent with what they’ve been told so far. For the same reason that good players of Fallout 3 can’t make a martyrdom ending work with the stories they’re trying to tell, evil players of Bioshock aren’t going to feel that bad about getting yelled at by Dr. Tenenbaum.

Outside of games, my intuition is that this moral universe is pretty hard to find. The protagonist and antagonist tend to embody opposing points on the morality line, so for a fixed-narrative story to embody this universe consistently, both the protagonist and the antagonist would have to win at the end. And that would be odd.

In reality, however, very few fixed-narrative stories actually embody a consistent moral universe for their entire duration. I mentioned before that Cinderella stays good all the way through. So why is it that her fortune varies so drastically? The reason is that the incentive system keeps switching up throughout her story. It’s really a series of graphs, like this:

In the beginning of the story, Cinderella’s good behavior is punished (her mother dies, she’s tormented by her stepfamily, and so on), while the bad behavior of her stepsisters is rewarded (they get free labor). Then the fairy godmother shows up, and the moral incentives reverse: Cinderella gets to go to the ball, and the stepsisters get snubbed. Then comes midnight and another reversal: her carriage turns into a pumpkin, and the stepsisters convince the Prince that she’s nobody. Finally, the Prince shows up at their humble home, and fortunes reverse again: Cinderella gets to be a princess, and something horrific happens to the stepsisters, level of horrificness depending on what era you’re telling the story in.

The important thing about all of these reversals is that the story announces them with a magical or unusual event. Cinderella’s moral universe doesn’t just change out of nowhere: a friggin’ fairy shows up and casts a spell. Then the clock strikes midnight. Then a powerful person appears at a hovel. Events like these, and the emphasis they are given, tell us that the nature of the story world is about to change. The ending of Fallout 3 follows a moment of thrilling, robot-marching victory for your side, and is announced by a bland conversation that is given no more emphasis than any other conversation in the game. You’re blindsided. The ending of Dragon Age: Origins, on the other hand, is preceded by a momentous offer of magical intervention. This signals the change in the story world in a traditional way.

In stories without magic, these changes are signaled in other ways, but they’re still signaled. Don Corleone gets shot. A messenger from Corinth shows up at Oedipus’s palace. Mercurio and Tybalt die. All major events that signal a sudden shift in the moral universe of the story and, subsequently, either a change in the protagonists’ behavior or in how their behavior is rewarded. And all these events are given special treatment in the storytelling.

I think this is an important lesson for narrative game design. It’s not that you can’t suddenly change the incentive system of a game halfway through. In fact, I think it would be interesting if more games did this; that’s one way out of Sztajer’s Half-Cinderella problem. The Half-Cinderella arises when a player notices the incentive system they’re playing in and figures out how to optimize it. After all, optimizing reward systems is what players do when left to their own devices. Changing up the incentive system could, in theory, make for a more varied and emotionally gripping experience, whether we’re talking about gameplay incentives or moral ones. But if you don’t signal the change, you’re just screwing with the player.

Share Button

Related Posts

  • Dragon Age 2: The Champion of KirkwallDragon Age 2: The Champion of Kirkwall It took me a long time, but I finally finished my first playthrough of Dragon Age 2. This is the first of a couple posts I'll be writing on it. To get the overall opinion stuff out of […]
  • Four Types of Videogame TragedyFour Types of Videogame Tragedy In my last post on tragedy, The Wrong Ending, I presented what I saw as an essential problem of tragedy in videogames: an ending where things go badly is often seen by players as […]
  • Pieces of StoryPieces of Story Telling good stories in games is hard. Interactivity screws up a lot of the things we associate with good storytelling, such as pacing, control of the audience’s knowledge, and and […]

› tags: bioshock / cinderella / dark souls / deus ex: human revolution / dragon age: origins / fallout 3 / lars von trier / mass effect / morality / player manipulation / vonnegut /


  1. […] Hollis, who if you are following you should be, has another great post up on her blog. In it, she puts forth some thoughts on a discussion of the incentives and plot […]

  2. Ari says:

    ‘Inexplicably’? Line, I beg to differ! I’m not arguing the fact that I don’t have a blog, because it’s true, I don’t: I’m arguing the fact that there’s no explanation for me not having one. In a nutshell, I’ve been writing my game ideas down since waaay before ‘blog’ became a household word. Turning over 15 years/over 75 Word documents/over 360 pages worth of notes into a blog is no small-time endeavor! Add to that the fact that I’m constantly revising old ideas, and even adding contradicting ideas from time to time, and I’m sure you’ll be able to appreciate that It’s gonna take me a while, even if it were near the top of my to-do list, which right now, what with dealing with a promotion at work and an upcoming move, it isn’t.

    Now then, with that out of the way…
    Seeing these familiar-looking charts gave me the idea that the character’s behavior is one aspect of the equation, and the most obvious. Slightly less so is the world in which said character exists in: what sort of behaviors does the world itself reward? Where does it sit on the sliding scale of Idealism vs. Cynicism? Harmony between the character’s morality and the morality favored by the world usually results in positive fortune for that character; dissonance, naturally, leads to ill fortune. But I’d argue that it’s in the dissonance where the most interesting stories are found.
    In cases like these, also factor in the following: what sort of story do you want to tell? Do you want the readers/players to cheer for the hero, or to jeer? Does the hero change the world? Does the world change the hero? Both? Neither?

    I’m not sure if all/any of the above could even be coded and put into a game, but it definitely allows more possibilities, more detail in storytelling, especially in video games, where “fight more, get stronger” seems to be the only path characters can take.

  3. Ari says:

    I like the fact that your graphs plot fortune & morality, in contrast to Vonnegut’s graphs, which plot fortune over time. In your version, one can see at a glance how the world views not only morality vs. immorality, but also partisanship vs. balance.
    Ask any D&D player or DragonLance reader: there’s more than just good and evil. There’s also neutrality, which very few video games even acknowledge, let alone reward.

  4. […] back to the Fortune and Morality curves of Line Hollis’ post, I could understand my standing in the game (measure of the resistance to my actions) by […]

  5. Line Hollis says:

    Hey, congrats on the promotion! I’m just busting your chops, I just want to be able to link to you. Agreed that stories where the protagonist is dissonant with the moral universe are a great unexplored area. This is where classical tragedies happen, for example, and as Dan Cox wrote in his response (which I’ll be linking to later), no one’s really figured out how to pull off tragedy in games yet. I think we were discussing that a bit in the same email thread; can you adjust the story in response to the protagonist’s flaws, maybe adjust the function/morality curve on the fly? It seems possible, but it’s yet another thing that requires thinking about story as a simulation, rather than an object with variables.

  6. Ari says:

    Thanks! I’ll have something linkable eventually…can’t even give an ETA right now. Too many things up in the air.
    As for tragedies in game…I know that games can have tragedy embedded into their narrative. Prior to the gods ascending Kratos to fill the position left by Ares, (the first) God of War can be seen as a tragedy. Also, I’ve never played Mafia 2, but I highly recommend reading

    What I’m trying to accomplish is having a tragedy in a game *while having the gameplay support that*. It’s harder than it sounds, but I know I’m on to something…

    And adjusting the morality curve on the fly…damn, that sounds more complicated still. I like your idea of changing the incentive system at a predefined point in the game (after properly warning the player, of course), and I believe it’d be much easier to implement.

    How would Fallout 3 have played if you could play someone living in DC before the bombs fell? Karma all of a sudden might have more weight and meaning to it when society is still intact.

    I also love your comment about players optimizing systems: as a DM, I know it’s what they do best! But in a video game with real choice and consequence, ‘optimizing’ may take on a different meaning. If I play a game that allows for a tragic outcome, then I should be able to optimize for tragedy. Player characters (in RPGs) become more powerful over time, but what if this power were just the set-up for the fall the character takes at the end? If a tragic outcome isn’t the only possibility, then the player is in control of the character’s very destiny. NOW you’re playing with power!

  7. ~hellfire~ says:


    I disagree that games haven’t been able to do tragedy. Deus Ex: Invisible War(a game most people hated) is a great example of one that does. I remember at one point I found out that I killed someone earlier in the game, thinking my reasons were good. This enraged and frustrated me: I had been playing the game trying not to kill anyone, but felt in this one case it had been justified. Then discovered it wasn’t, and that I’d been used as someone else’s tool. It’s a small moment in the game, but a tragic one.

    There are lots of other moments like that. I mean, every single ending the game offers is pretty unsatisfying; they’re still well done and meaningful, but they don’t make the player feel particularly good about his or her choices. Which is brilliant.

    What I do think is that games don’t offer up tragedy often enough or well enough. I think tragedy is really effective in games when it’s not just a function of a character’s traits or flaws, but also the player’s. Depriving the player of knowing her action’s consequences is one way to do that, as in the case I mentioned: I had the power to kill people, and actively misused that power.

    • Dan Cox says:

      I think I agree with you, hellfire. (Something I often find myself saying.) However, the game is not really a tragedy in the classical sense unless the antagonist can win. This is why I said, at the time, I had never seen any game do it. Since that post, two games have been brought to my attention, The Graveyard and The Path, as rebuttals to that.

      It could be argued that that most of the outcomes from exploring the forests for the women in The Path are tragic endings. (You, as the player, put them in that situation though. The central conflict is harder to parse.) Having knowledge of the ideal unrealized reality (actually following the path) during any additional playing session though means that the game can be a comedy. The only way to “end” that game though is to have all the women meet their “wolves”, so it could be categorized as a tragedy in the consideration of the overarching meta-experience of playing the game.

  8. Ari says:

    A couple of disagreements here @ hellfire & Dan Cox:
    Tragic elements in a game aren’t the same as having the entire game be a tragedy (again, in the classical sense). It would be interesting to define what makes a game tragic. Also, the antagonist doesn’t always have to win so much as the protagonist has to be defeated, usually by his or her own flaws. Who was the antagonist in Oedipus? When he blinded himself, did that mean that the Sphinx won? And wasn’t Iago eventually found out and punished at the end of Othello?

    I like the idea of a tragic fall being the result of the player’s flaws, but all players are essentially flawed in the sense that they don’t know everything that the character knows. Despite being the connection into the virtual world of the game, the player and the character are always two different beings. At best, you can direct and control your character, but you can’t fully become him/her/it.

    It might be interesting to have a game in which you do have knowledge of the consequences, yet you choose to act in a way that leads to tragedy.

  9. […] increasing becoming the case, Ari hits on exactly what I was going to be writing about today in the following comment: “It would be interesting to define what makes a game […]

  10. Ari says:

    Another difference between classic tragedy and video game tragedy: in a classic tragedy, I’m not responsible for the character’s fall. In a tragic video game, I am, at least to a great degree. It’s not my flaw so much as the character’s flaw that leads to the tragic end, but I did encourage its use.
    I also ask myself, what’s the designer’s expected reaction from the players at the end? Do we expect schadenfreude? Pathos? A feeling of responsibility for the character’s predicament? Something else?
    If John Q Player just chuckles evilly as he screws up the character’s life, then it’s not that different than letting your Sims die, and it lacks the power and the purpose of classic tragedy.
    Despite the divide between player and character, despite the fact that we can’t fully expect players to feel true pathos for the character any more than we’d expect them to feel brave for fighting virtual monsters or to feel real love for their virtual paramour (although I do admit a certain attraction for Tali’Zorah), any evocation of emotion through art (which games are, dammit!) will be less powerful than that same emotion through real life, just as art isn’t real life itself, but an approximation of it. Still, I always advocate the broadening of video games’ emotional expressiveness, so every bit helps.

  11. […] Cox of Digital Ephemera, with the help of commenter Ari, has taken some rough ideas in my post on Moral Incentives and Story Structure and made something terrific out of them. In a series of posts  at his blog, Dan and Ari have been […]

Leave a Reply

Your email address will not be published. Required fields are marked *