Until Dawn is a horror game about a bunch of dead-eyed teenagers in their mid-20s, none of whom have full control of their necks. The game is set in a dark, secluded ski lodge in the mountains in the dead of winter, where they’ve all gathered to celebrate the one-year anniversary of a cruel prank that resulted in the disappearance and/or death of two of their friends because why not.
Even though it’s emulating the style of horror movies, it’s inexplicably split into “episodes,” each with its own “Previously on Until Dawn” sequence to recap the stuff you just did 30 minutes ago. According to the episode count at least, I’m still only about halfway through the game.
Normally you’d experience an artistic work to its conclusion before you’d be arrogant enough to start critiquing it, but I’m not going to do that for two reasons:
- Even B-grade horror movies scare the hell out of me, so I can only play the game in short, tense bursts where my heart’s racing and I’m not particularly enjoying it. By the time I actually finish the game, it’s probably not even going to be relevant anymore and I might as well be writing a navel-gazing analysis of the ludonarrative complexities of Night Trap.
- I feel like I’ve already seen everything that interests me about the game.
What interests me is the way games are developing a unique language of storytelling. I particularly like trying to pick apart horror movies, because they have a built-in tension between active and passive storytelling that they share with narrative-driven games. (It probably helps that horror movies are usually so easy to pick apart, because they’re usually so direct in what they’re trying to do and trying to say).
So this is, if anything, a “first impressions” instead of a review of the game. So far, the game hasn’t blown me away with its originality or any particularly brilliant achievement, but what it does have going for it is that it’s completely accessible and surprisingly compelling.
It’s weird to be An Old Person (in video game terms) whose first exposure to horror games was Uninvited‘s black-and-white, MacPaint-drawn rooms, but still be dismissive of Until Dawn as an artistic achievement. Just look at it! It’s got recognizable actors like Hayden Panettiere and Peter Stormare painstakingly motion captured and rendered down to their pores, walking around extremely detailed environments with dramatic lighting. Plus each of the eight characters is controllable at some point in the game, which would be a ton of animation work even before all the narrative branching were taken into account.
So it’s frustrating to think of all that work being undone by dozens of small, seemingly insignificant details. Like how the characters seem stiff and overly fidgety unless they’re in a canned cutscene, at which point they’re clearly walking around on a sound stage. Or the eyes that never focus quite right, or the necks that don’t move quite naturally. As well done as it all is, it still ends up feeling like a bunch of robots wearing rubber masks of the actors.
That feels to me like a problem of technology, though. (I can already hear the groans and see the eye-rolling of character animators and modelers reading my dismissive “Press button to make character look human” take on it). The environments don’t get off as easy, because that seems like a problem of design. And one that hasn’t been solved by any narrative-driven game I can think of.
The first iteration of Until Dawn was apparently focused on directing a light source with a Playstation Move controller, and that’s very evident in the final game. There’s dramatic and atmospheric lighting throughout, with a flashlight or lantern cutting through the darkness, and it all makes for a very distinctive look. (One criticism I will make is that for every distinctive environment like the ski lodge or cable car station, there’s another generic one that’s lifted directly from the book of Early 21st Century Horror Movie Locations).
But the environments are detailed while by necessity having few objects that you can actually interact with. So important items are marked with a glint of blue light. Which means that a stray reflection off snow, or a strong specular highlight on a doorknob or vase or something, looks like an object of interest, and you keep getting taken out of the moment trying to figure out how to get to it and activate it.
On top of that are all the problems of level design that aren’t at all unique to Until Dawn: areas that feel like narrow corridors from point A to point B, rooms where it’s not obvious how far you can travel until you run into an invisible wall, spaces that give the illusion of being freely explorable but actually only have one or two areas of interest.
All of these problems are common, because they’re all fundamentally the result of having multiple design goals that are completely at odds with each other. Everything you do to encourage exploration and decision-making are going to kill your game’s pacing, and vice versa. Too few objects to interact with, and the environment feels barren and video game-like; too many, and you’re wasting time looking at incidental things that have no bearing on the plot, draining all the urgency out of the moment. The more you make a level “intuitive,” where the “right” way to go is the one that just feels right with no obvious clues, the less the player feels as if he’s actively exploring a space and making decisions.
It works the same way that the uncanny valley does for characters, since counter-intuitively, making things more realistic or more subtle just makes the problem worse. Playing Until Dawn has frequently reminded me of Gone Home, since they both have you wandering around dark spaces looking for things to pick up and turn over in your hands to get the next bit of environmental storytelling. Gone Home‘s objects and environments are obviously much less detailed and realistic than Until Dawn’s — whether out of intentional design or simply the fact that it’s a much smaller team and smaller budget — but it still has a better sense of place. It’s entirely likely that I’ve already spent more time in Until Dawn‘s ski lodge than the entire running time of Gone Home, but the latter’s house is the one that feels like a real place. I can still remember the layout of that house and where stuff happened, while I have no clear picture of how the ski lodge’s rooms even fit together.
It occurs to me now that “uncanny valley” is inherently optimistic; it just assumes that the problem will go away if we keep pushing forward. I’m starting to become skeptical. I’m sure that there’ll come a point in the future where it’s feasible and even practical to motion-capture an entire performance. Production on that type of game will become just like it is currently for linear media, and the software will be advanced enough to seamlessly blend between pre-recorded and procedurally generated movement in real time. In fact, after seeing how far “intelligent assistants” have come on cell phones in the past few years, I no longer think it’s unrealistic to expect CG actors to be able to understand natural language and respond intelligently.
But all that assumes that making things more realistic will solve all the problems, when we’ve seen time and again that interactive entertainment is a medium that rewards artifice and punishes realism. It’s Understanding Comics material: our brains are constantly looking for tiny, nitpicking details that will make something realistic seem “off,” while at the same time eagerly filling in the blanks on less detailed things to make them seem more recognizable and human.
To bring it down out of the clouds and back to a specific example from Until Dawn: most of the “teenage” characters are going for a completely naturalistic performance in both their voice delivery and motion capture, which ends up with lots of “ums” and “ahs” and overly-casual poses that just seem weird in comparison to everything else. It inevitably feels like a mannequin playing a recording of a real person instead of a real person.
Peter Stormare’s character, on the other hand, is played completely batshit crazy. He’s chewing the scenery so hard that he tears right through the fourth wall. It doesn’t feel at all real, but his character is still somehow the most compelling. Even though the lines he’s given and the questions he asks aren’t all that interesting, in my opinion. Most of the teenagers feel like disposable ciphers in comparison.
Warning: Explicit Language
So I think the “problem” with going for something hyper-realistic isn’t actually in rendering or art direction, but in game design and narrative design. When I said that Gone Home does a better job of establishing a sense of place, I don’t think it’s because of its relative low fidelity, or even due its careful and thoughtful level design (although both contributed to it). I think the main reason is that the game’s pacing allowed you to explore the space at your own leisure. You aren’t just dumped into a space and left on your own — it’s clear that a good bit of thought went into gating the sections of the house in a believable way and making sure that the revelations of the story could play out non-linearly and still make sense — but there’s nothing pressuring you towards the next story development except for your own interest and curiosity.
My biggest criticism of Gone Home is still the same as it was when I first played it: there’s absolutely no sense of player agency in the entire narrative. Everything interesting has already happened by the time the player’s game starts. And by the end of the game, it even seems to be mocking the player for wanting to participate in the experience.
Until Dawn is basically at the other end of the spectrum, desperate to remind the player how much they’re shaping the entire experience around player choices. I was going to say that it fetishizes player choice, and it does so almost literally: instead of fetishes, there are vaguely native American-ish totems lying around everywhere that dispense prophetic visions. (When each one gets discovered, the camera zooms inside, which makes for a hilarious image of a teenager picking up this weird totem and immediately slamming her face into it).
In case you’re unfamiliar with the concept of branching narratives, the game helpfully talks about the butterfly effect, then drives the metaphor home with an elaborate sequence where you fly over the veins in a butterfly wing. Each crucial junction point in the game is punctuated with an animation of butterflies. There’s a screen listing each of the story threads, which lets you page through your choices and see how they build on top of each other.
When I played Telltale’s Walking Dead series, I said that that game’s notifications of branching points (“So-and-so will remember that”) were a pleasant surprise. They seemed jarring, artificial, and clumsy at first, but in practice turned out to work like musical stingers. If we accept non-diegetic music in movies and don’t freak out that there’s suddenly a full orchestra in the shower with Janet Leigh, why dismiss non-diegetic notifications of in-story developments as being too “gamey?”
There are plenty of aspects of cinematic “language” that would be weird if we hadn’t spent a century being trained to accept them without a second thought. Cuts and montages are the most apparent, but even the way filmmakers compose shots is so deliberately unnatural that when we’re shown a scene framed the way a person would actually see it, it’s unsettling. In games, though, the tendency has been to reject all the game-like elements almost as if we should be ashamed of something so clumsy and primitive. Health meters have to be explained in-world as displays generated by your hazard suit, assuming they’re not eliminated altogether. Even Mario games have to explain that there’s a Lakitu following you around with a camera. Instead of developing a new language for games, it seems as if there’s a desire to hide the fact that they’re games as much as possible.
The obvious problem with Telltale’s approach is that they haven’t done anything with it. It was promising at first as an intriguing warning of story developments to come; after so much of it, it feels as empty as a jump scare. It’s just “a thing that these games do,” as if they’re not as interested in actual innovation in storytelling as they are in branding.
Until Dawn‘s notifications make Telltale’s seem restrained by comparison, but I think they work better as a result. When a story branch occurs or a clue is found, you’re shown exactly what happened, given a good idea of what it means, and you’re explicitly shown the junction points that led up to it.
That’s not to say that the choices are particularly interesting. So far, they’ve been all over the place — actual considered decisions are extremely rare. Most take the form of split second binary choices that don’t give enough information to judge against each other, e.g. “take the quickest route or the safest one?” Others let you slightly steer a conversation in one direction or another, which supposedly affects your relationships with the other characters. Others just go the Saw route and have you deciding between one horrible thing or another.
(When I bought the game, I’d forgotten that modern horror movies have been overtaken by stuff that don’t interest me at all or actively repels me, like all the torture porn franchises or the found-footage craze. I’d been expecting something more like 80s slasher movies or the Scream series. I’d still like to see more done with the Final Destination movies, because I think they’re relentlessly clever and it’d be interesting to see if it worked at all when made interactive).
As often as the game reminds me that characters can die as a result of my choices, I rarely feel like I’m making informed choices. But I’m not sure that that’s a failure of the game, because I don’t believe the game is trying to present a narrative built off of your player’s informed choices. I believe its ambitions are a lot more modest and straightforward. I believe it just wants to be a pastiche of horror movies, but with a simple layer of interactivity: instead of yelling at the screen “don’t go into that room!” you get to decide whether the character goes into that room or keeps going down the hallway.
In other words: it aspires to be a movie with some moments of interactivity, instead of a story-driven game that’s presented cinematically. The reason I believe that’s the case is because it uses the language of horror movies throughout, even at the expense of the game.
Sometimes, it works fine: the sequence in which Sam is exploring the lodge alone seems to be what the game was designed for, the standard scene from any slasher movie translated shot-by-shot into a video game.
Occasionally, it works so well that it seems too ingenious to be completely intentional: having multiple controllable characters is nothing new, but it turns out to be a perfect way to re-introduce cinematic edits into a video game narrative. Usually, games have to take place in some version of real time, and you’re either relinquishing complete control of the pacing to the player, or making the player feel like everything’s on rails and she’s in a shooting gallery. Thirty Flights of Loving is all about using cinematic cuts, flashbacks, and flash forwards in a first-person game, but it’s frankly tough to tell how much of the experiment could be translated to feature length. Until Dawn has no reservations about cutting away right as something interesting happens, but it doesn’t feel like missing time, or like the player’s had control ripped away from her, because she’s immediately given another part of the story to work on.
Most of the time, it just seems to be doing its own thing with the player input as something of an afterthought. You’re occasionally given something to open or push or flip over to read the back, but it doesn’t do much for engagement or immersion since you’re just following prompts. Same with several of the arbitrary binary choices: I can’t reliably predict what’ll happen if I choose hide instead of run, but I’m going to try it anyway. A lot of the quick-time button-press sequences, on the other hand, work surprisingly well. I despise QTE sequences on a philosophical level, but in this case, they add tension throughout — the usual “something is going to jump out and kill these fools,” but with the added stress of knowing that you could be asked to participate at any moment. Of course, these are even more random, unpredictable, and have no regard for agency: there are several sequences where Mike does a whole sequence of acrobatics in a cutscene. Or decides to shoot something, and even though I think it’d be a big mistake, I’ve got no option except to pull the trigger.
(To be fair: there are a few moments where the player’s choice not to do something is used for dramatic effect, and those are pretty well done).
But Until Dawn also insists on using some horror movie tricks that just definitively do not work in a video game, and it’s infuriating. The worst offender so far is the painfully long sequence of Mike and Jessica making their way through wooded paths up to a cabin. The game cuts away — over and over again — to show that there’s a strange person in the woods stalking them. We get just about every possible variation that’s been used in movies before: the Predator style POV shot. The shot where the characters walk off-frame but the camera stays behind to show the stalker waiting in the woods. There’s a particularly asinine jump scare one where the stalker is suddenly visible in a set of binoculars right as Mike stops using them. They’re infuriatingly tone-deaf, because they act as if what works in a movie will work in a video game with absolutely no thought given to player agency.
It’s entirely possible for a player to know more than a character, and to get tension out of that. In fact, Until Dawn does an adequate job of it later on, starting off the sequence I mentioned earlier where Sam is walking through the house alone. The player knows for a fact that there’s a killer in the house, but really, that’s something the player’s known since scene one. In that case, having the audience know more than the characters works exactly the same way it would in a horror movie.
The sequence of Mike and Jessica walking through the woods cuts away so often and so clumsily that it goes past “frustrating” all the way to “insulting.” If you show me a POV shot where a weird dude is looking directly at the characters I’m controlling, and then immediately turn the joystick back over to me, of course my first inclination is going to be to walk directly to where the guy is standing and ask him what’s going on. If you show me a flash of a bad guy in a set of binoculars, of course I’m going to immediately try and use the binoculars again.
The fact that I think it works with Sam’s sequence but completely fails with Mike and Jessica’s may seem like a contradiction until you consider what role the player has in Until Dawn. For me, at least, I’m never playing as Mike or as Sam. I’m floating in limbo somewhere between the director of a horror movie and the movie’s audience. Maybe I’m a production assistant?
During Sam’s sequence, the character’s goal and the player’s goal are aligned: we both want to find out what’s going on. So even though I know more than she does, following the trail is the best course of action because I know something interesting will happen when I get there. During Mike & Jessica’s sequence, their goal is to get to some absurdly distant cabin to have sex. I have absolutely nothing to gain from their having sex. I’m more interested in when this story is going to finally commit to being a horror movie and make something happen, already. So introducing the threat and then repeatedly showing it and pulling it away isn’t cleverly manipulating the tension between what the audience knows and what the character knows. It’s just showing me the thing I want to do — bring on the confrontation! — and then yanking it away from me for no discernible reason.
The other day I caught the tail end of a conversation/skillfully-defused argument where a bunch of people were trying to call out Patrick Klepek of Kotaku for writing “Emily, Who Is The Worst, Deserves to Die in Until Dawn“. As far as I can tell (Twitter makes it difficult to eavesdrop on other people’s conversations these days), people were accusing the article, if not Klepek himself, for being somehow complicit in the horror genre’s long history of misogyny. Or maybe it was because Emily and her boyfriend are the only characters who aren’t 100% white in the game? Like I said: tough to tell exactly what the complaint was.
Regardless, objecting to lack of empathy for a character in a horror movie doesn’t just miss the point; I believe it’s even more ghoulish than the alternative. Obviously, there are volumes of material looking at the “problematic” aspects of the horror genre, and its treatment — both intentional and subconscious — of women, ethnic minorities, and gay and transgender people. But to suggest that the audience is expected to feel genuine empathy for any of the characters in a slasher movie is either a pointlessly broad rejection of the entire genre, or is seriously messed up.
Knowing more than the characters do is an implicit part of watching a horror movie, since from the start of the film, you know they’re probably going to die horribly. If you’re feeling sympathetic towards the characters and getting attached to them, I’ve got to wonder why you started watching the movie in the first place. “I love seeing three-dimensional, fully realized humans be murdered and/or horribly traumatized!” The distance and lack of empathy is what makes the movies work at all.
The stakes in a horror movie (no vampire puns intended) aren’t something horrible happening to the characters, but something happening to you. You’re going to get startled by the jump scare or the face suddenly appearing in the bathroom mirror. You’re the one who has to feel tense knowing the character’s walking into danger. You’re the one who’s going to have to see something gross and disgusting. Until Dawn isn’t subtle about manipulating this: it explicitly asks you what bothers you, then shows you exactly that a few scenes later. Consider it “enthusiastic listening.”
Manipulating the audience as much as the characters is something that translates particularly well to video games, because the audience is even more invested in what’s happening. Not in the characters, necessarily, but in their story and in the things they’ll have to see. And as much as I hate QTEs, they add that layer of being invested in what I’ll be expected to do.
So far, even with its faults, Until Dawn pulls it off better than any game I can remember since the first Silent Hill or Eternal Darkness. It doesn’t require breaking the fourth wall, but it does require being aware of the fourth wall and how to use it. It may be as simple as the fact that the game is structured not so that you feel in control of what happens, but that you feel responsible for it.
Whatever the case, what’s increasingly clear to me is how much of the experience of interactive narratives depends on artifice. It rewards explicitly exposing the mechanics instead of subtlety. It benefits from deliberate design instead of hyper-realism. So much of the marketing of games — which has inevitably taken over the design, at least outside of indie games — is focused on selling the idea of player empowerment. You’re in control! You’re making all the decisions! This fantastic world has been built entirely for you! That should’ve been setting off all our bullshit alarms, even before GamerGate happened and made it explicitly obvious what a shitty goal that was for a medium that’s striving to be artistic expression.
It’s definitely true for horror games, but I think it’s true for all story-driven games: players aren’t giving you their money so that they can be in control; they’re giving you their money so that they can have fun being artfully manipulated.