Part 1
Roger Ebert used to say the star rating system is relative, not absolute. Asking whether a movie is good is really asking whether it is good compared to what you expected from it. That is a useful warning, because it means a five-star review of a competent thriller and a five-star review of a film like Everything Everywhere All at Once are not measuring the same thing, even though the page looks identical.
I have sat through plenty of films that were fine while they played and gone completely quiet a week later. The gap between fine and great is not really about plot. It shows up in a few places: what the story is actually about, whether the craft disappears into the experience, and whether the thing survives contact with your own memory.
Story is the skeleton, not the whole body
A tight plot keeps you watching, but plot alone rarely separates great from good. No Country for Old Men, 122 minutes, is one of the tightest thrillers the Coen brothers ever built, and it won Best Picture at the 2008 ceremony largely because the plot machinery is in service of an actual argument: that violence in this world has stopped making moral sense to the people meant to police it. Take away that argument and you have a competent chase movie with a scary villain.
Ask what the film believes. If you can answer in a sentence, and it earned that belief through events rather than a speech, that is a strong sign. If the only answer you can give is a plot summary, the film may still be enjoyable, but it has not earned the word great yet.
A thin plot can still carry a great film
The reverse is also true. Jordan Peele's Get Out runs on a plot that, described flatly, sounds like a modest horror premise: a young man meets his girlfriend's parents and something is off. The film cost roughly $4.5 million to make and returned nearly $260 million worldwide, and it won Peele an Original Screenplay Oscar, the first Black writer to win in that category. What makes it great is not the mechanics of the premise. It is that every scene is doing at least two things at once, a horror beat and a piece of social observation, and neither one is decoration for the other.
Craft is the part you are not supposed to notice
A scene worth rewatching muted
Pick a scene you found tense and rewatch it with the sound off. In Denis Villeneuve's Arrival, 116 minutes, the linguist scenes get their dread almost entirely from the score and the framing of the alien ship's interior, not from anything a character says out loud. Great films use craft to make you feel something before you understand why, and you can test that claim directly by removing the dialogue and seeing what is left standing.
[[AD]]
A quick craft checklist
Does the camera ever move for a reason you can feel rather than just to follow the action? Is the score telling you what to feel, or letting you find it on your own? When the film cuts, does it cut on an idea, or only on a physical action? None of these questions have a single correct answer, but films that consistently choose the harder, more expressive option tend to be the ones that stay with you longer than the runtime.
[[VIDEO]]
Performance lives in the reaction shot
The monologue gets the clip reel, but the strongest acting usually happens between lines. Ke Huy Quan won a supporting-actor Oscar for Everything Everywhere All at Once largely on the strength of quiet moments, not the film's louder multiverse sequences. Watch a face when nobody is speaking. A performance that stays alive in silence does more work than one that only switches on for the big dialogue.
Javier Bardem's Anton Chigurh in No Country for Old Men is the same idea from the villain's side. He barely raises his voice across the entire film, and the stillness is scarier than shouting would have been. A performance that announces its own menace is doing your job for you as a viewer. One that withholds it makes you lean forward instead.
The lingering test
Here is the simplest measure I actually use. A week later, what do I still think about? Everything Everywhere All at Once took seven of its eleven Oscar nominations, including Best Picture, Best Director, and acting wins for Michelle Yeoh, Ke Huy Quan, and Jamie Lee Curtis, and the image I still carry from it is a single quiet conversation between two rocks on a cliff, not any of the fight choreography. That is the residue Ebert's four-star scale was trying to point at: not spectacle, but a thing that keeps working on you after the credits.
Box office and star ratings measure something else
None of this lines up neatly with box office or aggregate scores, and it should not. A film's worldwide gross tells you how many people bought a ticket, not what stayed with them afterward. Arrival made a respectable but unremarkable amount at the box office next to that year's blockbusters, and it is still the film from 2016 I hear people bring up unprompted years later. The lingering test and the opening-weekend numbers are answering two completely different questions.
