Markiplier's Indie Horror Picks and a Design

Here's something I keep thinking about: the most interesting horror games right now aren't on the App Store. They're not on Steam's front page either. They're sitting on itch.io, built by solo developers and tiny teams, priced at zero or next to it, waiting for someone with a large enough audience to pull them into the light.

That pipeline — itch.io dev uploads a weird little game, Markiplier plays it, suddenly thousands of people wishlist it — is doing more for small-scale game discovery than most press outlets combined. I cover a platform that gets dismissed constantly for not being "real gaming," and yet the distribution model everyone in indie dev is quietly envying right now looks a lot like what mobile figured out years ago: low barrier to entry, direct creator-to-audience relationship, massive global reach. The gatekeepers are optional.

Markiplier's latest three-game horror session spotlights O.B.O.L., Wet Candy, and Mortem — three itch.io titles at very different stages of their design lives. Watching him move through all three is, unintentionally, a pretty clean illustration of what separates games that have something to say from games that have a checklist to complete.

O.B.O.L. — Demo That Actually Earns Its Wishlist

O.B.O.L. (available on itch.io at mreliptik.itch.io/obol) is a demo, and Markiplier clocks it immediately as sitting somewhere between Inscryption and Buckshot Roulette — two games that proved the tabletop-horror subgenre has real legs. The setup is Eldritch corporate bureaucracy: you're signing a blood contract with a Lovecraftian employer, presided over by a Scottish sea creature with the cadence of a pub regular who moonlights as an ancient evil.

The mechanic is deceptively simple. A coin flips. You bet caps. Your opponent, the creature, reads the "future" — announcing how many of each outcome will appear in the next several throws. The information is real, but it's deliberately incomplete, and figuring out how to use it against the house is where the game lives.

Markiplier spends a solid chunk of time genuinely working the math, getting it wrong, loudly acknowledging he got it wrong ("Everyone ignore the fact that I was very dumb in that one"), and then rallying. That cycle — misread the information, recalibrate, press advantage — is exactly what good mechanical horror does. The dread isn't a jump scare. It's the slowly tightening resource pressure. When he finally strings together a winning sequence by stacking items and exploiting a passed-out opponent: "YES. OH MY GOD, I'M THE BEST." That's earned.

The environmental lore teased between rounds — diary entries about drilling schedules and food rations, imagery that Markiplier reads as potentially Frostpunk-adjacent — suggests the full game has a world with actual weight behind it. "I love the atmosphere," he says. "I love the underwater... This hearkens back to when all we had was just a few coins and a few cards." There's something there. The demo knows it's a demo. It shows you enough to want more without pretending to be finished. That's harder to pull off than it sounds.

Wet Candy — Discomfort That Earns Its Ending

Wet Candy, from developer Twist and Scream (twistandscream.itch.io/wet-candy), does something structurally interesting: it wraps a horror story in the most mundane possible premise — a kid wants candy from a pool — and then steadily corrupts every assumption that premise carries.

The surface reading is simple. The mechanical reading — swimming deeper, following a directional indicator past increasingly wrong things, bodies in the water, a parent on a bridge screaming about their missing daughter — points somewhere darker. Markiplier lands on a conflict-and-loss reading by the end, noting it "seems to be some kind of Vietnam War story, probably based a little more true than one would like." That reading comes from the visual layering: military-adjacent imagery, the body-strewn water, the parent's grief on the bridge, the suggestion of something that was taken and cannot be returned. Whether that's the intended interpretation is an open question — itch.io horror games rarely come with director's commentary — but the fact that it supports a reading that coherent means the design is doing something. The mechanics and the mood are pulling in the same direction, even when the story stays oblique.

"It's always one of these where I get to the end of it," Markiplier says, "and I'm like, I really don't feel I should — LIKE THAT WAS GREAT. HAD A BLAST. OH, MAN." That uncomfortable whiplash is kind of the point.

Mortem — The Store That Has Nothing To Do With Anything

Mortem (thewindyproject.itch.io/mortem) is where the session gets genuinely useful as design criticism, because Markiplier doesn't just say it's bad — he explains exactly why, and then builds out an alternative version on the fly.

The premise is familiar: watch your uncle's corner store for a night while something ominous happens in the city. You restock shelves, mop floors (in a way that suggests the developer has a theoretical understanding of mopping but has perhaps never held a mop), serve two customers, encounter some odd NPCs, find a body, stuff it in a wardrobe, and then the game kind of just... ends. The horror arrives disconnected from everything that came before it.

"The entire horror part of it was disconnected entirely from the main setting, which was the sweet shop," Markiplier says. "There was nothing that had to do with the finale there that had anything to do with the store... It's checkboxes. It's loosely associated checkboxes to be like, this fills out enough time to qualify as a horror game."

He's right, and the checkboxes are the tell. The store tasks don't generate dread. The NPCs don't build toward anything. The old man who drops an ominous hint about knowing your uncle walks out without consequence. Every interaction is load-bearing in the sense that it takes time, but none of it is load-bearing in the sense that it matters. That's the structural failure.

What's interesting is that Markiplier doesn't just sit in the critique — he riffs his way into a better version in real time. What if the uncle had died? What if you weren't keeping the store open but cleaning it out, and the horror was discovering what your uncle was actually involved in? What if the store was a front, the old man was a partner, and the game made you the person who has to decide what to do with that information? The bones he's describing — inherited guilt, a corrupted domestic space, a city that knows something you don't — are good bones. They're just not the bones Mortem has.

The thing about itch.io as a release platform is that it's genuinely frictionless in both directions. It's incredibly easy to ship a game. It's also incredibly easy to ship a game before it knows what it is. There's no release schedule pressure from a publisher, no QA pipeline, no external voice asking "but what is this about?" — and sometimes that freedom produces something like O.B.O.L., where the constraints of a demo format actually focus the design. Other times it produces Mortem, where the absence of that pressure means nobody asked the hard question before the upload button got pressed.

That's not a knock on itch.io. It's a knock on skipping the part where you figure out what your game is trying to do. The platform is neutral. The question of whether the work is ready is not.

The gap between O.B.O.L. and Mortem isn't talent — both games have atmosphere, have interesting aesthetic instincts, have a reason to exist. The gap is whether the mechanic and the story are working on the same problem. O.B.O.L. puts you in a cosmic power negotiation and makes the coin flip feel like your life is on it. Mortem puts you in a corner store and makes the mopping feel like homework.

A player can feel the difference in the first five minutes. Markiplier's audience — millions of people, most of them on exactly the kind of accessible platform that itch.io represents — absolutely did. 🎮

— Jordan Mercer