The film nobody quite agreed on
Few films in recent memory have divided audiences quite as cleanly as It Comes at Night. Horror fans who showed up expecting a creature feature — lured by a marketing campaign that suggested something prowling in the dark woods — left baffled and often furious. A smaller, quieter audience recognized something else entirely: a serious-minded psychological thriller about what fear does to people who are otherwise decent, and how quickly decency buckles under pressure. Both reactions make sense. Understanding which camp you'll fall into before you press play is, frankly, useful consumer information.
Writer-director Trey Edward Shults was thirty years old when he made this, his second feature, and the confidence on display is both impressive and slightly unnerving in itself. This is not the work of someone hedging bets.
What it's about
The premise is deliberately spare. An unnamed plague has ravaged the world. A family — father, mother, teenage son — has retreated to a boarded-up house in the woods and established a set of rigid rules for survival. Their isolation is broken when another small family arrives, also fleeing, also desperate. The two groups attempt to coexist. Things go wrong.
That's essentially it, and Shults trusts it to be enough. There are no grand explanations of the catastrophe, no news broadcasts helpfully narrating the collapse, no third-act reveal of what the illness actually is. The world outside the house exists only as threat and rumor.
Craft
The filmmaking here is genuinely accomplished. Cinematographer Drew Daniels does extraordinary work in the confined spaces of the house, where the oppressive production design — dark wood, low ceilings, a single red door that becomes the film's most potent symbol — conspires to make even daylight feel contaminated. The film is mostly dark, and not lazily so; Daniels finds real texture in shadow, making each rare burst of natural light feel both beautiful and dangerous.
Shults uses the 4:3 aspect ratio for dream sequences — the frame literally narrowing around the protagonist as his nightmares intrude — a device that sounds gimmicky described in print but works surprisingly well in practice. These moments are genuinely unsettling, and they do real thematic work, blurring the line between threat perceived and threat real in ways that pay off structurally.
Joel Edgerton anchors the film as Paul, the patriarch. It's a quiet, internal performance that resists easy sympathy — Paul is competent, controlled, and increasingly frightening in his logic. Edgerton never plays him as a villain, which is exactly right; he plays him as a man who has decided, in advance and in good conscience, that his family comes before everyone else's, and is now living with the consequences of that decision made concrete. It's the kind of performance that grows in retrospect. Kelvin Harrison Jr., then still relatively unknown, is equally impressive as the teenage son Travis, and carries most of the film's emotional weight through his eyes alone.
The pacing is slow — deliberately, defiantly slow — and Shults earns most of it. Each scene of the two families navigating cautious domestic normalcy is thick with the knowledge that it can't hold. A shared dinner, a child's laughter, a quiet moment of genuine warmth: all of it lands with a particular weight because the film has so efficiently established that trust is a resource being drawn down, not replenished.
What works and what doesn't
What works: the atmosphere is immaculate, the performances are strong across the board, and the film's central argument — that paranoia is not irrational in extremis, and that's precisely what makes it tragic — is executed with real intelligence. There's a sequence roughly two-thirds of the way through, involving a door that was supposed to be locked, that is one of the better-constructed scenes of ambient dread in recent horror-adjacent cinema. No jump scares, no score sting; just the slow, horrible crystallization of suspicion into something that can't be walked back.
What doesn't work, or works less well: the film's refusal to provide any emotional release — any catharsis, any revelation — tips over at times from rigorous restraint into a kind of punishing withholding. The ending will strike some viewers as the only honest conclusion possible and others as a refusal to do the final work of drama. I lean toward the former reading, but I understand the frustration. The marketing-versus-product mismatch also feels worth naming, not as an artistic criticism of Shults but as a practical one: the film was sold dishonestly to an audience that was always going to feel cheated, and that manufactured backlash obscured genuine discussion of what the film was actually attempting.
The dream sequences, while effectively staged, are slightly overdone by the third or fourth iteration. The film makes its point about Travis's psychological deterioration clearly enough through Kelvin Harrison Jr.'s waking performance; the recurring nightmare imagery starts to feel like underlining.
Who should watch it
It Comes at Night is for viewers who can tolerate — or actively prefer — horror that works through implication, mood, and character rather than incident. If you found The Witch or A Quiet Place too slow or too cerebral, this will likely test your patience harder. If those films worked on you, this one almost certainly will too. It asks something of its audience, and not everyone will feel the exchange is fair. But it's the work of a filmmaker with a genuine vision and the technical ability to execute it, and that remains rarer than it should be.
