Dead Secret, the mystery game I’ve been working on since 2013, will be released on Steam and Oculus on Monday, so I thought I’d write about something I thought about a lot while making it: strategically setting the player’s expectations.

A couple of years ago I was browsing my favorite weirdo bookstore in Japan and found a thin volume called いるのいないの？ (Iruno Inaino?, “Is Anyone There?”).  It is a “horror picture book,” written and drawn in the style of a children’s story, with a hardback binding, simple grammar, and large, full-page pictures.  It’s actually a part of a series of short horror stories all done in this style by a variety of artists.  Despite its looks, this isn’t a book for kids.  I hid it on the top shelf of my bookcase so that my eight-year-old won’t accidentally read it before she’s ready (which is probably not for a few years).  It’s scary.

I suspect that part of the reason Iruno Inaino? is effective is precisely because it’s presented as a children’s book.  The protagonist is a child. The narration is from the child’s perspective, in the child’s voice, and uses an age-appropriate vocabulary.  Without spoiling it for the four or five of you who are going to find it on Amazon, import it, and translate the Japanese, it deals with a fear that any child might feel.  A simple, basic fear that we can identify with even as adults.  The story is effective, but the ultimate effect is achieved by all the elements of the book working together to set your expectations and then defy them.  Children’s books, as a rule, are not supposed to be scary.  Definitely not this scary.  The packaging of Iruno Inaino? is designed to be disarming.  It leaves us vulnerable to the sharp edge of the story’s thin blade.

Horror is about loss of control.  It’s about uncertainty.  A good horror story throws us for a loop and plays with our expectations.  Sometimes this happens because the story has been framed to lead our expectations elsewhere.

I do not think that Event Horizon is a very good film.  It’s absolutely derivative of Alien (a better movie by every standard) and offers very little in the way of compelling story or interesting characters.  About half way through it drops all pretense of plot and goes straight for gore.  And yet, the first time I saw it, I must admit that it scared the shit out of me.  There I was, operating under the theory that good horror stories are, first and foremost, good stories, and wondering how this fairly uninteresting example of space fear managed to get the drop on me. After a second viewing I figured it out: Event Horizon was scary because it caught me off guard.  The advertising I had seen led me to believe that it was a 2001-style Space Explorers movie and not a scary horror film.  I went in expecting Mission to Mars and got Alien instead.  Even as a derivative of Alien, the mechanisms it uses are effective, particularly if you are not expecting them.  The second viewing was a comparatively bland affair because I now knew what to expect (Alien, on the other hand, still scares me).

Packaging and presentation of horror media sets expectations.  And expectations matter a whole lot to the final experience.  Folks who are really good at horror understand this and control the presentation of their works very carefully.

Take, for example, this early trailer for Frictional Games’ SOMA, released almost two years before the game itself came out:

Now, if you’ve played SOMA, you might have noticed something: this sequence doesn’t actually happen in the game.  In fact, the character portrayed here isn’t even the main character.  This entire level segment, the graphics and the script and the physics and the voice acting, all of it was created just for this trailer.  It’s not in the final game, and was never intended to be.  The purpose of this trailer is to show you what SOMA is like without actually spoiling anything.  It’s setting expectations without giving anything away.  All of Frictional’s early trailers are like this.

Here’s another example.  The box art for Catherine is provocative. It looks like an anime pin-up girl and it’s designed to make you believe that Catherine is about some sexy stuff. Catherine is in fact about sex, but not the titillating kind.  In fact, every hint of nudity in Catherine is downright nerve-wracking.  The packaging of this game sets it up as some sexy anime girl game so that it can watch you squirm when it starts posing difficult questions about the ethics of long-term relationships. Sex in Catherine is weaponized, and its goal is to make you feel uncomfortable.  It does this, in part, by misdirecting your expectations and then throwing them off a cliff. The blade gets twisted a lot in this game, and it hurts.

The concept of controlling the presentation and packaging of a game in order to misdirect expectations is fascinating to me.  I can’t really talk about the steps we took in Dead Secret without spoiling it.  I will say that we’ve been very careful to only share builds with streamers that contain the first few minutes of gameplay (like this one).  I guess you’ll just have to play it to see the rest.

[This post contains spoilers about one of the best moments in Siren.  If you haven’t played it yet, please ask yourself what you are doing with your life.]

There is a moment in Siren in which your character, a young teenager, is trying to reach a church to meet her parents. The landscape is full of crazed shibito, some of whom have begun to grow bug-like appendages and extra sets of eyeballs. It is dark and foggy, and to survive the trek you must sneak past these monsters using the title’s signature “sight-jacking.”

The game has trained you well by this point. Siren is a game in which stealth is paramount; if a shibito even so much as hears you it’s all over. A flash of red and quick cut to the enemy’s perspective signals that you fucked up, you failed to sneak well enough, and now you’re about to get killed. By the time you reach the church stage you are a sneaking pro.

Your high schooler crouches behind some bushes as you wait for the bug-thing roaming around a few feet ahead to establish a pattern. Her name is Tomoko. She’s still wearing her red phys ed jumper, as if the cataclysmic event that turned the residents of this rural Japanese town into monsters happened right in the middle of her gym class.

The moment comes and you go for it, crouch-walking just behind the shibito as it pauses to examine a flyer pasted on an ancient farmhouse wall. As you pass you can hear the mutated ex-farmer crying. But then a flash of red, and for just a moment you see yourself through the eyes of the enemy. You didn’t make it. The shibito has turned and discovered your hiding place. Tomoko has no weapons and no means of defense, and this close there’s no chance of escape. That’s it, you fucked up, game over.

And then, it doesn’t happen. The monster doesn’t attack. It saw you, that much is clear, but it doesn’t move. The creature just sits there, a few feet from your 11th-grader, looking at her. Slowly you inch her away and then break out in a run. The monster continues to sob quietly to itself as it watches you go. After a moment it returns to the flyer.

Something is wrong here. The game has trained you that getting caught equates to a quick and grisly demise. You’ve put hours and hours into this game so far, and never has a monster failed to attack. The rules, somehow, have changed.

As Tomoko stumbles forward other monsters see her. There’s a red flash of recognition, but nothing approaches her. Nothing attacks. You don’t know what’s going on, but this opportunity is too good to pass up. You put the girl into a sprint, her red jumper a blur of color across the monochrome landscape, and cover more ground in the next minute than in the previous ten. Soon the church is in sight.

The end of this level is one of the scariest moments Siren has to offer. Sony used it in a television commercial that the Japanese government banned for being too scary. It’s a moment where the game’s developers knife you in the back and then twist the blade.

The reveal is this: when Tomoko reaches the church, she runs to the windows and bangs on them for help. Her parents, sitting inside, turn at the noise and are horrified. Tomoko’s eyes are bleeding. She’s joined the ranks of the terrible shibito without realizing it.

One of the most powerful things about Siren is that it forces you to play as characters that ultimately do not survive. There is a large cast of playable characters, and in the end most of them don’t make it. By constantly switching from person to person, the game denies you the comfort of knowing that the protagonist can never die. It eschews the trope of the untouchable main character by not having a main character.

Alien does this as well. The passive camera refuses to give the viewer a protagonist, and so he must assume that any character might die at any moment (and, for the most part, they do). Alien sets up Dallas, the strong male lead of the crew, as a potential hero and then promptly kills him. There’s no clear hero until everybody but Ripley is dead.

Siren is exceptional in its capacity to replicate the Alien model. How many games can you name in which the protagonist is not obvious? How many that have no obvious protagonist? How many in which some of the playable characters turn out to be enemies of other playable characters? This just doesn’t happen in video games. Games are usually about someone in particular, and death of that person is just a play failure. Oops, you died, restart.

In Siren, characters die and the game progresses. The effect is subtle at first, as individual levels play out like traditional games. There’s a single protagonist and death is a level-restart failure. But the meta game spans time and characters in a way that I’ve not seen often, and the effect becomes powerful: nobody is safe. A level restart is not enough to revive a character. Most will not make it.

The one other title I’ve played that does this is Eternal Darkness. Though very different from Siren, it uses the same multi-character mechanic to keep you on your toes. Characters die and the game progresses, the roster of playable characters continually expands, and early protagonists become antagonists by the end of the game. Eternal Darkness gets away with a few scenes like Eye Bleeding Tomoko because its fundamental game structure allows for playable characters to be killed off. If you make a bad decision you can’t always just restart and try again.

That’s the meta-mechanic here: making decisions carry weight. In a medium where every failure can be undone, horror games must go out of their way to increase the cost of a mistake. This is how good horror games create tension. Atmosphere and environment might suffice for an hour or two, but eventually a horror game must teach its player that their decisions matter. Resident Evil does this by rationing shells and saves. Silent Hill does this putting save points far apart to increase the time cost incurred by restarts. Amnesia obscures the rules related to failure, forcing you to constantly second-guess yourself.

Siren does all these things: it rations resources, has few checkpoints, employs obscured rules, and on top of that it allows characters to die. It is terrifying.

I’m pretty interested in Five Nights at Freddy’s and Slender.  Both are low-budget indie horror games that have, in the last few years, reached incredible levels of viral success.  The two titles are very different: Freddy’s is about a security guard who must survive the night in a department store full of killer animatronic animals, and Slender is a game about finding notes in a forest while avoiding the infamous Slenderman.

Despite the obvious differences both of these titles have enjoyed significant success, particularly with teenagers. The extent of the impact these games have had on American teens wasn’t clear to me until Mike, the Art Director at Robot Invader, told me about his daughter.  Apparently she and her friends like to get together and play a game called “Slenderman.”  The setup, as described to Mike by his daughter, is very similar to the design of Slender, only this game isn’t played with a computer: one kid gets to be Slenderman and she chases the others around outside.  In fact, most of these kids have never played Slender.  The game has wormed its way out of the computer monitor into their consciousness without direct contact.  Freddy’s is even a bigger sensation, apparently.  Here’s a family counselor explaining Five Nights at Freddy’s to worried parents.  Both of these games are simple to play and pretty scary, and have managed to capture a large audience of kids.

Though Slender and Freddy’s are very different games, they share a number of similar traits.  Perhaps it is these traits that make both titles so accessible to younger audiences.

Obfuscated Rules

Both Slender and Freddy’s go out of their way to obscure key parts of their game design.  In Slender, the speed and movement pattern of the Slenderman is unpredictable.  You never see him move, but when you turn your back on him he gets closer.  In Freddy’s, the movement of the monster animals is also unclear, as is the amount of power consumed by each of the actions you can perform.  By making key rules unclear, the authors of these games make it harder for you to treat the experience as a logic puzzle.  There’s no comfort to be found in understanding the rule set and exploiting its weaknesses because the rule set itself is hard to pin down.

Forced Unawareness

Both Slender and Freddy’s force you to look away from the thing that is trying to kill you.  This is a pretty powerful mechanic.  In the case of Freddy’s, you can’t spend all of your time on the monitors or you won’t have enough power for the doors.  In Slender, looking at Slenderman actually hurts you, and you have no choice but to turn away from him and run.  This mechanic induces tension quickly because you must actively deprive yourself of incisive information to win.

Pop-Out Scare Failure Event

Neither Freddy’s nor Slender rely on a series of pop-out scares, but both use them to make failing a level dramatic.  Rather than pop some hideous creature out of a dark corner every few minutes, these titles build tension with the threat of a pop-out scare, which doesn’t actually occur until the player fails and reaches the game over state.  These games are designed for replay, and once you’ve failed once the weight of the impending pop-out scare serves to dramatically increase your level of stress.  This is a very smart way to use pop-outs, I think, because it gives the player no release; rather than employ the common pattern of building tension to a pop-out, then easing off, the pop-out-at-the-end approach only relinquishes its grip when the game is over (or the level completed).  The actual pop-out effect is almost inconsequential.  Tension is built by the impending scare that you know is coming if you fail.

Mettle Tests

The design of Freddy’s and Slender is good, but I think their virality amongst kids has to do with them being tests of mettle.  These games are a safe way to prove your courage, both to yourself and your classmates.  Like Candyman, Bloody Mary, or Hanako-in-the-Toilet, Slender and Freddy’s provide easy-to-reproduce fear challenges that kids can perform without involving adults.  The challenges are equitable and accessible: anybody who has a mirror can try to summon Bloody Mary, and Five Nights at Freddy’s can be played on just about every PC, smartphone, or tablet platform under the sun.  Children use these challenges to test their mettle and then boast about the results to their peers.  This social, competitive aspect of these games is probably what drives them to be as widespread and popular as they are.  But of course, they would have never become tests of teenage fortitude if they had not been well-designed and extremely accessible to begin with.

Slender and Five Nights at Freddy’s are good games, and are worth playing if you want to study interactive horror.  But more interesting is the way that these titles have gripped a segment of the audience not usually considered by horror game designers.  The appeal of these games to teenagers is, to me, the most fascinating part of their success.

When we describe games like Resident Evil to others, we probably say something like this:

In Resident Evil, zombies have taken over a big mansion and you have to stay alive by shooting them and solving puzzles.  You have to conserve ammo and read documents to uncover the mystery behind the mansion and the zombies.

While technically true, this sort of description doesn’t actually describe what you do from moment-to-moment in Resident Evil.  A more accurate description might be:

In Resident Evil you visit a large number of rooms. When you enter a new room, you look for zombies and shoot them.  Then you run around the edges of the room and press the search button to find items and documents and keys and puzzles. By finding items and solving puzzles you can open up doors to new rooms. Eventually you visit all the rooms.

This description isn’t sexy, and it ignores the narrative elements of that game, but it’s accurate.  If you recorded all the button presses used in a Resident Evil play session I’m sure you would find that “search” is one of the most common operations, second only to movement.  Resident Evil, and almost all horror games from that era, spend a lot of energy on the ransacking of the game world.  In these games, missing a hidden item can cause difficulty to increase or even stall progress completely.

Ransacking a room isn’t a very exciting game mechanic, but it is one of the pillars upon which many horror games are built.  The side-effects are numerous: the pace of the game is slowed, players treat the environment as an area to be systematically swept, and item collection becomes synonymous with survival.  Ensuring that every single item has been collected becomes an important, as missing even one could have consequences later.  For example, Silent Hill has items that ruin the ending of the game if missed.  The message is clear: search everything, everywhere, or run the risk of negative repercussions.

On the upside, ransacking forces the player to pay attention to the environment. Resident Evil and its ilk love puzzles that involve opening up a secret area or revealing a previously hidden compartment because these force the player to pay attention.  If you want all the items you’re going to have to open all the doors which means you can’t just go running around shooting zombies haphazardly.  It’s a mechanic that requires focus, and as a result, gives the designer a lot of opportunities to inject story information into the player’s consciousness.  You had better pay attention to the details of that painting and read all of the text in that document if you want to ensure that you’ve found every last nook and cranny there is to find.  Resident Evil in particular doubles down on this approach by making health, ammo, and save items rare commodities that must be collected at all costs.

Hope I found everything.

On the other hand, ransacking can make you play the game like an idiot.  The mechanic encourages you to believe that every item must be found, and so you traverse the corners of every broom closet and alleyway just to ensure that nothing is missed.  It creates anxiety: whenever a ransacking game closes an area off without warning (e.g. in a cutscene), you feel uncomfortable about items you might not have found.  Even worse, collectors going for every item must anticipate upcoming game events that might restrict their access to a segment of the level map.  The problem is exacerbated when the map is large.  The way forward is obvious, but you always take the alternate path, never forget to look under the stairwell, and methodically check every stall in every bathroom.  Each time there is a branch in the map you have to guess which direction is the “real way” and which is likely to dead-end in an item room.  This pattern is especially frustrating when it conflicts with a narrative that wants you to move forward.  The level design in The Last of Us and Alan Wake suffers from this: the story says, “GO! GO! GO!” but the map design is like, “hey, maybe check out this door over here?”

Ransacking encourages the player to think about his environment as a puzzle rather than as a real place.  We’re checking mental checkboxes as we clean rooms out, taking note of rooms we’ll need to return to, maintaining a ledger of tasks completed and tasks outstanding.  Obscure has a character who’s special power is to declare whether all the items in a room have been found, and this power is the most useful of all the characters in the game.  This mode of thinking is not the most conducive to scaring the player.  In fact, it often becomes an annoying chore.  We’re paying attention, but only because we don’t want to be burned later.  Horror games about ransacking and collection must find ways to give the player breathing space, where the pressure of finding every last thing is eliminated and they can enjoy the story for its content.

One mitigation strategy is simply to make items of interest flash or sparkle.  This approach releases the stress of ransacking by making the operation much easier: you can simply look at a space and see if there’s anything to do there or not.  When there’s nothing to see, the player can swiftly move on to something else.  The problem with this method is that it removes the focus element that ransacking provides, and makes it more difficult for the game to grip the player with its narrative.  Fatal Frame and many other games thus employ an uneasy middle ground where some items sparkle but others do not, thereby affording narrative detail for players who wish to consume everything while releasing others from the need to exhaustively search.  Still, when I play these sorts of games, I can’t help but worry that I’ve missed something important, and end up turning the whole place upside down just to be sure.

Resident Evil 4 threw ransacking out the second-floor window by removing the need to find specific items.  There are very few keys or puzzles in that game, narrative sequences tend to be built into the core progression path, and consumable items such as ammunition and health are available in infinite supply from fallen enemies.  There are still item boxes to find hidden away for players who go exploring, but by removing the need to search an environment for items the design ensures that nobody worries that they missed something.

Good thing I checked.

More recently, Amnesia takes an even more dramatic approach by simply eliminating items altogether.  Sure, there’s an occasional key to find, but with no combat and no ammo, very little item management is required.  Lamp oil is rationed to force the player to turn out the lights, but if you do so then oil supply is not a major concern.  In Amnesia, every drawer of every desk can be opened, but there’s almost nothing to find because ransacking is not the mode of thinking that the developers wish you to assume.  The Amnesia school of thought dictates that by removing your ability to think of the environment as a system, you’re forced to think about it as a real place.  That’s probably one of the reasons why that game is so much scarier than most games from the Resident Evil era.

In Dead Secret, the VR murder mystery game I’m building with my team at Robot Invader, exploration of the environment is the primary mode of play.  There is no combat or health, but we do have other sorts of items, and you’ll need to search every room throughly to find them all.  To prevent ransacking from becoming a chore our approach is to put a high concentration of items and documents in each area.  You will not spend your time barraging through a series of rooms, fingers crossed that you haven’t missed anything.  On the contrary, you will take your time in each room to discover the interesting clues it has hidden away, and you can come back later if you want.  Our items do not flash, but sweeping the environment is not difficult because the reticle changes when it passes over items that can be collected.  When experienced in virtual reality, physically looking around your environment becomes one of the main modes of play.

My goal is to get you to focus, to think about the content of the rooms, without forcing you to exhaustively check every single corner.  It’s a careful balance, and hopefully it works.