Victory Conditions

So I've been replaying Final Fantasy 6 recently, and I'll spare you the “OMG I FORGOT HOW GOOD THIS GAME IS” talk, but something occurred to me.

For those of you that don't know (and even those of you that do), there is a moment, after what is more or less the apocalypse, where you wake up in control of single character, only to discover that you and your adoptive grandfather have been stuck on an island for over a year since the catastrophe, and Celes (your character) has only just now woken up. There used to be a few other people on the island, but they've since died of either disease or sheer ennui. As far as you know at this point, Celes and Cid are the only human beings left alive at this point, and Cid has taken deathly ill.

You're presented with a minigame, in which Celes catches fish to feed to Cid, in order to nurse him back to health. Fast fish add to his gauge, slow fish subtract, and it slowly drains over time. If it gets up to 10, he recovers, and if it gets down to 0, he dies.

The weird thing is, you can't lose the game at this point. Ultimately, it makes no substantial difference if he lives or dies. Either way, a short series of events transpires that end up with you on a raft, trying to make for the mainland, in order to discover if there's anyone else out there.

Whenever I played this game as a kid, I would always save Cid, since I wanted to beat every possible challenge set before me, and it really wasn't that difficult. I had never even seen what happens when you let him die.

That is, until this most recent time. When you come in to feed Cid, and you've let him die, Celes checks on him, discovers that he's dead, and runs from the room in tears (accompanied by her theme from the opera earlier). She walks to the northernmost point of the island, and throws herself off the cliffs, figuring that with everyone else gone, she has nothing left. She nevertheless survives the fall, only to find the raft that Cid was secretly working on before he took ill. Why he didn't mention it before escapes me, but there you have it.

From the connection you have to the music, to the little save point-ish sparkles they use for her tears as she falls, it's easily one of the most emotionally powerful moments to ever come out of a Super Nintendo (perhaps only second to Lucca telling Robo at the end of Chrono Trigger: “Don't pretend you don't care when you're really sad”).

The scene if you save Cid, however, is kind of an anti-climax. He... feels better then tells you about the raft. He stays on the island, and you can visit him any time you like after you get an airship. Kind of... well, boring.

And yet, because I was presented with success/failure conditions and the task was easy enough, I never saw this different, decidedly superior path for the story to take. This has gotten me thinking: when the most dramatically interesting path for a story to take involves a character failing at the task set before him/her for no reason other than their own personal failings, what's a game designer to do?

I realize this is a bit of a soapbox issue for me: If I refuse to take control away from the player at any point, how can I craft any kind of emotionally engaging situation that requires the player's character to act in some way?

Malgayne, as I was talking with him about it, really hit the nail on the head, I think. If I think of myself as being Celes, then clearly I have to do everything within my power to save Cid, but if I think of myself as a 3rd party, guiding the story, then it frees me up have something happen because I want to see how it would play out.

In a game like Mass Effect or Fallout, there could never be a moment in which the villain gets away because your character falters at the crucial moment, and lacks the resolve to pull the trigger. Or, more appropriately, if this ever happens, it must either happen in a cutscene, during which you have no control, or events must occur that make your hesitation functionally irrelevant. If it significantly changed the course of the game from every point forward, the workload would be, at this point, prohibitively high. I would like to think this will not always be the case, but that's where we are right now.

And even if you were given the choice to have your character hesitate, and not shoot the villain at the crucal moment, all Keyzer Soze style, how many of you would honestly take that chance? The “Game Over” screen has conditioned gamers everywhere that failure to accomplish your originally intended goal means that you stop playing the game. JRPGs and that whole genre has, by this point, devolved into a story (occasionally a very engaging one) that you must pass a series of largely unrelated challenges to experience. If you fail a challenge, you don't learn any more of the story until you beat said challenge, so the impulse is always to try as hard as possible to beat every fight put in front of you. If I was legitimately afraid that I would go game over, I would have my character kill the villain right then and there, out of fear I would have to replay some part of the game because I failed the challenge.

But it's not as simple as “well, give players the option to make their characters fail on purpose without ending the game”. Take the example of feeding Cid the fish: I didn't want to save Cid because I cared about his character that much, or because I really identified with Celes and her struggle to find an anchor in the cruel new world she had found herself in. I wanted to save Cid because the game had given me a challenge, and dammit, I'm going to win, even if it involves nothing other than walking up to the shoreline and mashing the A button to pick up fish. Frankly, if a game tasks me with something, I want to be able to succeed at it, but I have a lot of difficulty imagining how you could design a challenge for which the success condition involves your character throwing herself off a cliff in despair.

I think the real problem here is that when I say “game”, people think “success/failure conditions”. It seems like the industry would benefit from moving more in the direction of something like Daniel Benmergui (his website can be foundhere), who has always been fairly clear that in most of his games, while there may be some conditions that can be described as more positive for the characters involved, there's rarely one set of circumstances characterized as a goal for the player (other than perhaps experiencing all possible permutations of events).

But most of these projects are small in scope, and deal with the player's control over a small number of variables. I would say “I hope that this take on games evolves into something with the relative size and scope of something like Mass Effect of Final Fantasy 6”, but we all know it will eventually. It only took 24 years for Mario to make it up to the first Mario Galaxy, so let's check back around 2030, right?


Frank said...

When I played FF6 for the first time (I was around 13), I either didn't know or didn't bother reading about the difference between fast and slow fish, so I fed Cid slow fish and he died, and I was like "well that sucks" and played through the rest of the game. It wasn't till my second playthrough that I figured out how to save Cid, and I kept him alive every game since, not because of a metagaming "must win" impulse (and you know my urge to powergame is fearsome), but because I already knew the emotional impact of him dying and preferred the world where he was alive, even if it had no other effect on the game.

Multiple playthroughs and my initial failure maximized the emotional impact of the scenario for me. I suppose it was dependent on me "identifying" with Celes, not wanting to imagine her feeling terribly alone, but how can you have a compelling story without making the player feel something like that?

As for games not allowing player choice to affect the story because it's prohibitively expensive--well, this is just me talking as a programmer, AI-ist, and game player, but it really shouldn't be that bad guys. In every motion picture, some one goes through and checks for consistently constantly, from page to page and shoot to shoot. And in the Sims game, some one goes through and defines the motivation and effect of every action a Sim can take. Combine the two and see what happens. That amount of effort seems relatively minor compared to building and animating and voicing tens of 3D characters. In terms of gamelay, mainstream games like Mass Effect have progressed almost NIL from some combination of Zork and Quake/Jedi Knight/Daggerfall--13 years since 3D came into its own, and we have nothing to show for it but graphics and physics and enough storage for voice acting. I'd wager it's simply because few studios have made the effort. Chris Crawford, eccentric rogue game designer, has something working at storytron.com. You should check it out.

William said...

I'm not sure I agree with you that it's that simple.

Let's start with the assumption that a game has a well defined, well plotted, emotionally satisfying narrative (a really tall order, given the industry, I admit). Just defining what everybody thinks and feels at every point, and writing some kind of algorithm to define how various emotional states interact with each other doesn't seem like it would produce satisfying emotional resolutions to stories.

It would, if you could do a really good job of it, produce people acting like they would in real life (or at least how real people would act if placed in a fantastical situation), but I wonder how satisfying that would be to watch.

Real life is filled with drama, I admit, but I feel like the drama comes from when ordinary people act in an extraordinary fashion, and I bet it depends on your view of human nature and how deterministic your brain chemistry is, but I'm tempted to say that modeling these kinds of things is, at best, very difficult, and maybe impossible.

That's not to say that I think that creating something that can (quasi-)randomly generate emotionally satisfying stories is impossible, but I think that it needs to be approached from the angle of what makes drama and catharsis. While there's a lot of ink on the subject from respectable sources, it needs to be both more coherently codified and actively acknowledged by the game industry.

I think it's clear that true random generation needs a few breakthroughs (though not necessarily technological ones) before it becomes feasible, I think that what's going to come first is some kind of breakthrough that allows the number of possible endings to explode exponentially without similarly increasing the workload.

I'd love to see a game like Mass Effect, where you have an intensely customizable avatar, and the game gives you enough choices to slowly build a fairly comprehensive profile of the kind of character you are playing, and then makes slight tweaks to the ending to ensure that your character is successful at saving the universe because of exactly the kind of person they are.

Would it really be that bad to just assume that the player will play a psychologically consistent character, and to take choices away from them if they've proven themselves to act a certain way? I'm not sure about that one yet.

Frank said...

You story snob--people like comedies and action adventures too. Do you think the formula for climax-denoument/story closure is that involved? Catharsis is too strong a word for the end of a typical sitcom episode.

Well when I'm set financially or socially, let's give it a crack.

Alison T-dove said...

This is a cool post. I'm pretty sure if I faced that challenge in a game I'd "be Celes" and have to save my brother--but your point that it makes an anti-climactic bit in the story line is very true. I wonder why it's in there at all?