In any case, I wrote a thing about the whole 'subversion' thing, 'cause I felt like it. Take it as you will.
---
People get caught up on examples when thinking about horror, but examples only help you make tropes. And tropes aren’t scary because we’re used to them.
So start from the beginning. Think about the basic conceit of your game, and where the horror will have to come from.
At the most basic, fear is uncertainty.
And so there are two tenets to good horror and nothing else is needed:
SUBVERT EXPECTATIONS
and
PACE THE TENSION
I’ll use examples to illustrate stuff, but focus on the points rather than the examples, they’re just illustrations.
---
Most people expect horror to be like a heartbeat. Drop, spike, even, drop, spike, even. In other words, jump scares. The problem with this, is it tires people out far faster. They do it automatically: they have to either get used to it, or they get weary and quit.
Most horror is actually like a sinewave. Pacing smoothly between high and low tension, with occasional subversion. This is less likely to tire people out as it gives them downtime and comfort to offset the surges of tension, but many people are very accustomed to it, as it gets predictable and it’s...what’s long been done, basically.
What you want to aim for is an arrhythmic pattern, still keeping both above in mind. Shock is necessary, pacing is necessary, but everything in moderation. Let your player rest, but never let them sleep, so to speak.
A random pattern can work in this regard, but it’s much better to plan it out yourself. Random patterns may let you down and artificially kill the tension and you’ll never know why because everyone’s experience will be different.
So: What you need to do is subvert expectations, and then let them get comfortable with the new subversion, subvert that, and so on down. Your aim is to reach BREAKDOWN before they grow too weary or bored to continue.
Breakdown is the state where their expectations have fallen apart completely. You could do anything at this point, and they won’t know how to react. This is where you can do random patterns, within reason. You still don’t want them to get too weary or comfortable.
Now in any given game, there might be multiple overlaid patterns going on. But they tend to be clearly separated from each other when you look at the systems.
---
An example of the heartbeat is in Amnesia, with the puzzle rooms.
You go into an enclosed room or set of rooms often to solve a puzzle, and almost always have to back out the way you came. There is always a monster there after it’s over. It’s an obvious spike that becomes more obvious and less effective as the game goes on.
X is a monster after. O is no monster.
OXXXXXXXXXXXXXXXXXXXXXX
Boring, and right quick.
Here’s how I would have done it: imagine there are 20 rooms in a row.
OXXXOOOOXOXXXOOXOXXO
Breaking it down into chunks:
O XXX OOOO XO XXX OOX OXXO
First O is teaching the player the game’s systems. They’re still vulnerable at this point and anticipation will do your work for you.
XXX is three in a row, shocking them each time in slighly different ways. They begin to fear by anticipation, but the first O room will leave them some uncertainty, and by the third room they’ll sort of think they’ve got it figured out.
OOOO is where you start to mess with them. Make the ambiance creepier, but leave them alone. Having more Os in a row than Xs will help make them complacent by the end, but they’ll never be absolutely sure, especially since XXXOOOO also subconsciously tells them that there is no pattern.
XO is when you throw them off the edge. Attack them to ruin their complacency, and subvert their expectation of a chain of enemies.
XXX immediately turns that new expectation back on them. At first it will shock them again, but they’ll feel stronger by the third.
OOX is where you just ruin everything for them. XXX will remind them of the first XXXOOOO pattern. They’ll be expecting Os after the last few, and you’ll give them a couple, but fewer instead of more. Then hit them when they’re utterly unsure. Breakdown.
Beyond that point, just mix it up. Doesn’t matter too much. If you can get them over the cliff they won’t know what to expect at all. And then you can do whatever you like if you still have them. Just gotta keep the pacing moving.
---
A good example of effective arrhythmia is in the White Chamber.
When you step into a room and there’s a horror event going on, the style of everything is slightly (or severely) off. So you get used to that idea.
Then, at one point, you enter a central room, and all the paths but one are blocked by a giant dark fence and the walls are all messed up, and...well, you know. There’s only one way out, and you don’t want to take it because you’re expecting the worst.
But when you finally muster the courage to go in there, it’s normal. You solve a puzzle, your brain goes ‘directed me to the puzzle, no harm’. And then when you come back out again the fence is gone. All is well, you are comforted. You go into the next room and it’s all gone to hell when you’re at your most vulnerable. Fun fun!
---
Basically, just don’t get caught up in thinking about what other people have done. Horror gets caught up in shock and spectacle, but surprises are not scares. The subtler and smoother your tension is, the easier you can grab people. If you sidestep tropes entirely, people won't know what to expect and will breakdown far quicker.
Here’s a few things that can be done in any game at all that don’t involve ‘how do i get the player into a dark room’:
- Change a thing. Change a background object, move something, something small and subtle. Most people won’t notice it, but their brain might. And if they do, it’ll freak them out all the same.
- Have solutions to problems in-game that stop working. Pills that stop healing you, guns that run out of bullets, etc. I don’t mean make stuff break and make people hunt for it, make it an easy enough fix that they get tense without getting frustrated. And this differs in method from game to game, but let the player be the one who runs out. They'll make themselves tense as they run low, and even more when they know they have to use their last X. Don't take it from them unless it's vitally important, as they may feel cheated.
- Have time be an omnipresent threat. Have sounds in the distance that sometimes sound closer. Have clocks on the walls. Make things take time to find. Time in horror always feels short, which is how it should be, even if you have all the time in the world.
- Move doors. Or make doors go to the wrong places. It’s amazingly effective and can surge tension in an instant, but only use it once. As soon as you expect it, it’s useless.
Here’s a few things to rarely or never do:
- Take control from the player. Horror is at its best when it’s happening to you. If they can’t do anything, they’ll feel less invested in it. (It’s a bad trope that this can make the player feel helpless. It can, but it backfires incredibly easily.)
- Expect the player to try something multiple times. Seen some games do this, and while it might work to have a door not work the first time and then open when you desperately check it again, you can’t control your player’s level of desperation so it can mess with your tension curve, or just make them quit.
- Have things that make no sense at all. Comedy will kill your mood and we laugh at absurdity. Have things be right, but wrong. Subversion again. Inversion just doesn’t work in something like this. Instead of being attacked by giant ants, being attacked by tons of ordinary but incredibly aggressive ants. Or the like.
- Make them run from something they can see clearly. It just doesn’t work for the same reason why we can’t identify with someone being chased in a movie.
- Take time to kill the player. Don’t make them redo 20 minutes of work because they missed a switch and were guaranteed to be dead anyway. Don’t drag out death scenes. Actually...
- Kill the player. You should only die if you really screw up. Death is catharsis in a horror game. The fear of having to redo things isn’t sufficient to keep the tension from cratering when you kill the player.