You know that feeling when the credits roll and you realize you have absolutely no idea what you just watched.
Not because you weren't paying attention. Because the film was actively working against you.
Right. And Daniel wrote in with a whole thing about exactly that. He wants five movies with some of the most extraordinarily confusing, convoluted, difficult-to-follow plots ever put on screen. And he's specific about what he means. Not just intellectually demanding films. Not movies that leave one or two questions dangling. He's talking about films where a substantial proportion of viewers struggled to understand what was happening. Who the characters were. When events were occurring. What was real. How the plotlines connected. What had actually happened by the time the credits rolled.
And he wants the mechanisms. What specifically makes each one hard to follow. Nonlinear chronology, multiple timelines, unreliable narrators, identity shifts, dreams, time travel, causal loops, parallel realities, withheld information, enormous casts, excessive interconnected developments.
Then for each film, he wants us to actually explain what happens. In reasonably straightforward language. Highlight the scenes that are notorious for confusing audiences. Discuss whether understanding requires a second viewing or external diagrams or extensive online discussion. And where possible, evidence that the confusion was widespread rather than just our subjective assessment.
He also wants the movies to be good. That's the key constraint. Preference for otherwise accomplished films where the complexity became an obstacle to enjoying them.
And then rank five films in order of increasing narrative bewilderment. Number five is merely extremely difficult to follow. Number one is where keeping track becomes an undertaking in its own right.
There's a lot to unpack here.
So let's start with the paradox that runs through this entire list. Every film we're going to talk about is critically acclaimed. Primer won the Sundance Grand Jury Prize. Mulholland Drive topped the BBC's poll of 177 critics as the greatest film of the twenty-first century. Synecdoche, New York was Roger Ebert's film of the decade. Last Year at Marienbad won the Golden Lion at Venice. These aren't failures. They're canon.
Confusion and prestige coexist. That's the thing that makes this interesting. We tend to assume that if a film is confusing, it's because something went wrong. Bad writing, bad editing, bad structure. But every film on this list is a critical darling. The confusion is often deliberate. It's achieved through sophisticated craft.
But deliberate obscurity and accidental confusion are different things. Shane Carruth and David Lynch both explicitly refused to simplify or explain their films. Carruth said the machine and the characters' experience are inherently complicated, so it needed to be that way in order for the audience to be where Abe and Aaron are. Lynch said of Inland Empire, it's about a woman in trouble, and it's a mystery, and that's all I want to say about it.
Then you have Cloud Atlas, where the Wachowskis and Tom Tykwer attempted accessibility. Six interlocking eras, an ensemble cast playing multiple roles across periods, and they still lost viewers.
And that's the fascinating part. They weren't trying to be cryptic. They wanted people to follow it. The themes are universal. The emotional throughlines are clear. But the structure itself became a barrier. So intent isn't the deciding factor. You can try to be accessible and still end up with something audiences can't track.
Which gives us our ranking tool. The second viewing test. Memento and Mulholland Drive reward a second viewing with clarity. You can piece together the puzzle. Primer and Inland Empire arguably cannot be resolved by rewatch alone. Primer needs external diagrams and three-dimensional timelines. Inland Empire resists any single reading. That's the axis we're using.
And I want to be clear about what we mean by "reward." We're not saying these films are bad because they're confusing. We're saying the confusion operates differently in each case. Some films are puzzles with solutions. Some are experiences that resist solution entirely. And that distinction matters for how we rank them.
Start at number five with a film that's extremely difficult to follow but generally solvable. Memento.
Two thousand, Christopher Nolan. Two timelines running simultaneously. Black-and-white sequences in chronological order. Color sequences in reverse-chronological order. They meet at the end.
Leonard Shelby has anterograde amnesia. He cannot form new memories. He tattoos clues on his body and takes Polaroids to track his wife's killer. The confusion is mechanical. You have to mentally reverse the color sequences while tracking the forward-moving black-and-white ones. That's a cognitive load most films never ask of you.
And it's not just the structure. It's the way the structure mirrors Leonard's condition. He can't form new memories, so he lives in a perpetual present tense. He doesn't know what came before. He doesn't know what he's already done. The film puts you in that same position. You're experiencing events without the context that would make them meaningful.
And the film withholds the crucial revelation until the final reverse-chronological scene. That Leonard has already killed his wife's attacker. That he's manipulating himself into a cycle of false revenge. Because he can't remember that he's already done it.
Let me walk through that ending, because it's the key to the whole film. In the final black-and-white sequence, which is actually the chronological beginning, we see Leonard in a motel room. He's talking to someone on the phone. And he reveals that his wife survived the attack. She died later, from an insulin overdose. And Leonard doesn't know if he gave her the insulin or if she gave it to herself. He's blocked it out. So the quest for revenge that drives the entire film is built on a false premise. He's already killed the man who attacked his wife. He's been repeating the cycle over and over, using his condition to erase his own actions.
The structure is the point. Nolan is putting you inside Leonard's experience. You don't know what came before. You can't trust your own memory of the film because the film is deliberately scrambling it. By the time you reach the end, which is actually the beginning, you understand why Leonard does what he does. He's not solving a mystery. He's constructing one.
National Film Registry selection in twenty seventeen. Budget somewhere between five and nine million. Box office forty point one million. One hundred thirteen minutes. Two Oscar nominations, Original Screenplay and Film Editing.
And it's solvable. That's why it ranks lowest. Watch it twice and the structure clicks. The puzzle has a solution. Nolan is difficult but he's not cruel.
Number four. Cloud Atlas. Twenty twelve. Lana and Lilly Wachowski and Tom Tykwer.
Six distinct eras. Eighteen forty-nine, nineteen thirty-six, nineteen seventy-three, twenty twelve, twenty-one forty-four, twenty-three twenty-one. An ensemble cast. Tom Hanks, Halle Berry, Jim Broadbent, Hugo Weaving, and others playing multiple roles across periods. Connected by a reincarnation theme and a comet-shaped birthmark.
The confusion here is different. It's not about reversing timelines. It's about tracking which actor is playing whom in which era. And how the nested narratives connect. A journal read in one era. A letter in another. A film in another. You're constantly recalibrating who these people are and what century you're in.
Let me give a concrete example. Tom Hanks plays a doctor in the eighteen forty-nine segment, a physicist in nineteen thirty-six, a journalist in nineteen seventy-three, a gangster author in twenty twelve, and a post-apocalyptic tribesman in twenty-three twenty-one. Same actor, six different characters, six different time periods. And the film cuts between them rapidly. You'll be watching a scene in the nineteenth century, and suddenly you're in a futuristic Korea, and then you're in London in the nineteen thirties. The transitions are jarring on purpose.
The structure is a matryoshka doll. The eighteen forty-nine journal is read in nineteen thirty-six. The nineteen thirty-six letters are read in nineteen seventy-three. The nineteen seventy-three story becomes a film watched in twenty twelve. Each era contains the previous one as a document. And the actors recur across all of them, which means you're constantly seeing Tom Hanks and thinking, wait, is he the same character? Is this reincarnation? Is this thematic casting?
And the film doesn't answer that question directly. It leaves it ambiguous. Are these literal reincarnations? Are they thematic echoes? Is the birthmark a marker of soul continuity or just a visual motif? The film suggests all of these possibilities without committing to any of them.
One hundred seventy-two minutes. Budget somewhere between one hundred and one hundred forty-six point seven million. Box office one hundred thirty point five million. Polarized critics. It appeared on both Best and Worst film lists.
The Wachowskis and Tykwer were trying to make something accessible. The themes are legible. Freedom, oppression, love across time. But the execution required viewers to hold six narratives in their heads simultaneously. And most people can't do that on a first viewing.
Number three. Synecdoche, New York. Two thousand eight. Charlie Kaufman.
Reality and fiction collapse. Doppelgängers. Decades compressed into a single narrative. Philip Seymour Hoffman plays Caden Cotard, a theater director who receives a MacArthur grant and builds a life-size replica of New York inside a warehouse. He hires actors to play people from his life. Then he hires actors to play those actors.
The timeline spans decades but compresses them without clear markers. Characters age, die, and are replaced by actors playing them. The boundary between Caden's real life and his theatrical production dissolves entirely. By the end, you're not sure if anything we've seen is happening outside the warehouse.
Roger Ebert named it the best film of the decade. Entertainment Weekly's Owen Gleiberman gave it a D plus and wrote, I gave up making heads or tails of Synecdoche, New York. The LA Times called it hard-to-follow. Rotten Tomatoes sixty-nine percent. Metacritic sixty-seven.
Budget twenty million. Box office four point five million. It appeared on one hundred one Best of 2008 lists, twenty of them at number one. Ranked twentieth-greatest film of the twenty-first century in the BBC poll.
The scene where Caden hires an actor to play himself, then hires another actor to play that actor. It's infinite regress. You're watching a man watch a version of himself watch a version of himself. And the film doesn't tell you where the nesting stops.
There's also the matter of time. The film spans decades, but you don't always know how much time has passed. Characters have conversations that seem to take minutes but actually span years. Caden's daughter grows up, becomes a painter, becomes a performance artist, and then becomes a woman who speaks only in a strange accent. And you're never sure if this is real or if it's part of the play.
The film is about a man trying to make sense of his life by turning it into art. But the art consumes the life. The production becomes more real than the reality it's based on. And by the end, Caden is directing a play about a man directing a play about a man directing a play. It's turtles all the way down.
The top two films are where comprehension becomes contested. Where even dedicated viewers disagree about what happened.
Number two. Mulholland Drive. Two thousand one. David Lynch.
Dream logic and a mid-film identity collapse. Naomi Watts and Laura Harring play dual roles. Four-fifths of the way through, the narrative shifts. Betty is actually Diane. Rita is Camilla. The first section was a dying fantasy.
The confusion is retroactive. You have to reinterpret everything you've seen. The first ninety minutes are revealed to be the dream of a woman who has hired a hitman to kill her lover. Every scene you watched, every relationship you thought you understood, gets recontextualized. And Lynch doesn't explain it. He just shows you the blue box, the narrative collapses, and suddenly you're in a different film.
Let me walk through the pivot point, because it's one of the most famous narrative shifts in modern cinema. Betty and Rita go to a theater called Club Silencio. A man on stage says, no hay banda, there is no band. Everything is an illusion. Then a woman sings a Spanish version of Roy Orbison's "Crying." She collapses, but the singing continues. The performance was a recording. The illusion is revealed. And then Betty finds a blue box in her purse. She opens it. And the film cuts to a completely different reality. Betty is now Diane. She's a failed actress. Rita is Camilla, a successful actress who has rejected Diane romantically. And the first ninety minutes were Diane's dream, a wish-fulfillment fantasy where she was the talented newcomer and Camilla was the helpless amnesiac who needed her.
Lynch supplied ten DVD clues, including, where is Aunt Ruth? The Boston Globe's Steve Greenlee argues Aunt Ruth is the true protagonist. The producer and editor, Mary Sweeney, said, everyone is informing it with their own psychological package. It means different things to different people. I can't lend legitimacy to any theory.
That's the thing. The film resists a single reading. Named best film of the twenty-first century in the BBC poll. Number two on the New York Times list. Premiered at Cannes in May two thousand one. The site Mulholland-Drive dot net hosts twenty-nine competing theories. Twenty-nine. That's not a film with a solution. That's a film with a community of people who can't agree on what they watched.
And it rewards a second viewing. That's why it's number two and not number one. You can piece together the dream logic. You can see the clues Lynch planted. But you'll never be certain you're right.
Number one. Primer. Two thousand four. Shane Carruth.
Time travel with nested causal loops. Budget seven thousand dollars. Shot in five weeks. Carruth wrote, directed, produced, edited, scored, and starred. Won the Sundance Grand Jury Prize in two thousand four.
The film depicts only a fraction of actual events, out of chronological order. At least three Aarons and two Abes coexist in the final timeline. Carruth deliberately obfuscated the plot. He said, this machine and Abe and Aaron's experience are inherently complicated so it needed to be that way in order for the audience to be where Abe and Aaron are.
Mike D'Angelo in Esquire wrote, anybody who claims he fully understands what's going on in Primer after seeing it just once is either a savant or a liar. Chuck Klosterman said, I've watched it seven or eight times and I still don't know what happened. The A.V. Club's Scott Tobias wrote, were it not for the presence of voiceover narration, the film would be close to incomprehensible.
Seventy-seven minute runtime. Rotten Tomatoes seventy-two percent. Metacritic sixty-eight. Box office eight hundred forty-one thousand, nine hundred twenty-six dollars.
The film requires six-page explanations and three-dimensional timeline diagrams. That's not a joke. The fan community has produced those. You need external scaffolding to understand what you watched. And even then, there's disagreement about what the diagrams should look like.
Let me try to explain the basic premise, because it's actually simpler than the film makes it seem. Two engineers, Abe and Aaron, accidentally invent a time machine. It's a box that lets you travel back to a moment when the box was previously turned on. So you can't go back further than the box's own history. Abe and Aaron start using it to make money, then to fix personal mistakes. But they start creating paradoxes. Multiple versions of themselves. Timelines that overwrite each other. And the film doesn't explain which version is which. It just shows you fragments of a much larger story.
The second viewing test applied. Memento and Mulholland Drive reward a second viewing with clarity. Primer and Inland Empire arguably cannot be resolved by rewatch alone. Primer needs external diagrams. Inland Empire resists any single reading. That's why Primer takes the top spot.
The role of external explanations is the tell. For Primer, six-page written explanations and three-dimensional timeline diagrams. For Mulholland Drive, twenty-nine competing theories catalogued. For Inland Empire, ScreenRant and other outlets have produced ending-explained articles that themselves acknowledge the impossibility of a definitive reading. When the explainers admit they can't explain it, you know you're in unusual territory.
And here's a fun fact that puts Primer in perspective. The film cost seven thousand dollars. That's less than most people spend on a used car. Carruth shot it on sixteen-millimeter film in his spare time. He used friends' houses as locations. He composed the score himself. And he won Sundance. That's not just an independent film success story. That's a reminder that narrative complexity doesn't require a massive budget. It requires a filmmaker willing to trust that the audience will do the work.
Hilbert has something to say about this.
Hilbert: I ran Mulholland Drive fourteen times in one week.
Fourteen.
Hilbert: Repertory cinema. Lynch retrospective. Late eighties. We had a print that had been around, and the distributor wanted it back, so we squeezed the whole run into seven days. Two screenings a night. I was the projectionist, so I sat through all of them.
And you understood it more each time?
Hilbert: Less. By the seventh screening I started noticing things I was sure weren't there the first time. A lamp that changed position. A phone that rang before it was picked up. I became convinced the film was gaslighting me personally.
That's a real thing, actually. The film is designed to make you doubt your own perception.
Hilbert: Everyone talks about the dream logic. The identity collapse. Nobody talks about the sound design. That's what makes it incomprehensible. The ambient hum. The distorted dialogue. The way the audio drops out at key moments. You literally cannot hear what's happening half the time. The film is confusing because the sound is working against you.
That's a dimension we haven't touched. The sensory experience of confusion.
Hilbert: I kept a notebook during those fourteen screenings. Wrote down every time I noticed something new. By the end I had forty-seven entries.
Forty-seven.
Hilbert: I still have it. Never opened it again. Afraid it'll have changed.
The notebook is the artifact. The record of a man trying to impose order on something that refuses to resolve.
Hilbert: I was good at that job. Threading a projector, timing the reels, keeping the print clean. Nobody writes about that part. But the sound design on that film, the way it messes with you, that's the thing I remember. The picture is one thing. The sound is what gets under your skin.
So is there a point at which narrative complexity stops being a feature and becomes a failure? Every film on this list is acclaimed. But every one also alienated a substantial portion of its audience. Does that matter?
The confusion didn't kill these films. It became part of their identity. Primer and Mulholland Drive have spawned decades of fan theories, diagrams, explainer articles. Inland Empire was remastered by Janus Films in twenty twenty-two. Synecdoche appeared on one hundred one Best of 2008 lists. The difficulty is the legacy.
And there's something else. These films don't just confuse you. They invite you to participate. They assume you'll watch them more than once. They assume you'll talk about them. They assume you'll argue about them. The confusion is a form of engagement. It's the opposite of a film you forget the moment it's over.
The films that confuse us most are often the ones we can't stop thinking about. Maybe that's the point. Maybe the confusion is the hook. You walk out of the theater not knowing what you saw, and you can't let it go.
Hilbert's notebook. Forty-seven entries. He never opened it again because he was afraid it would have changed. That's what these films do. They keep changing after you've seen them.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop.
If you enjoyed this episode, leave a review on your podcast platform of choice. It helps other listeners find the show.
We'll be back soon.
See you tomorrow.
