Understanding movie ratings and scores

Every streaming catalogue attaches a number to a film. The numbers come from different sources, measure different things, and are routinely compared as though they were interchangeable. Knowing what each actually represents makes them useful rather than decorative.

Audience scores and what they measure

The most common number is an average of ratings submitted by viewers.

What it measures is not quality but satisfaction relative to expectation. A film that delivers exactly what its audience came for scores well. A film that is accomplished but not what the audience expected scores badly. This is why genre films with devoted followings frequently outscore acclaimed work.

Two biases are worth knowing about. Self-selection: only people who chose to watch a film rate it, and people choose films they expect to like, so scores skew upward. And enthusiasm asymmetry: strong feeling motivates rating far more than mild approval, which pushes distributions toward the extremes and hollows out the middle.

The practical read is that an audience score tells you how well a film satisfied the people who sought it out. That is genuinely useful information, and it is not the same as whether you will like it.

Critic scores and what they measure

Critic aggregates work differently. A smaller number of professional reviews are converted to a scale and averaged, sometimes as a percentage of favourable reviews rather than an average of scores.

This distinction matters more than people realise. A percentage-favourable figure tells you how many critics liked a film at all, not how much. A film every critic found mildly acceptable can score higher than one that half found extraordinary and half disliked — which is close to useless if what you want is something extraordinary.

Critics also watch far more films than most viewers, which shifts their reference point. Novelty is weighted more heavily, familiarity penalised more harshly. For a viewer who has not seen four hundred films this year, that calibration does not transfer cleanly.

Why the two disagree

Systematic gaps between critic and audience scores are common and informative.

Audience above critics typically indicates a film serving its genre audience precisely while offering little to someone outside it. Sequels, franchise entries and broad comedies show this pattern.

Critics above audience typically indicates a film doing something unfamiliar or demanding, which the audience that showed up was not prepared for. Slow pacing, ambiguous endings and unconventional structure produce this reliably.

Neither gap means anyone is wrong. The gap itself is the signal: it tells you the film is divisive, and often in a way that predicts whether you specifically will get on with it.

Sample size matters more than the score

A number attached to a film with forty ratings and one attached to a film with forty thousand are not comparable, and catalogues rarely make the distinction visible.

Small samples are volatile and easily skewed — by a devoted niche, by people connected to the production, occasionally by coordinated effort. A very high score on a film almost nobody has rated is weak evidence about anything.

Where a service shows the count, it deserves as much attention as the score. Where it does not, treat obscure titles' scores with more scepticism than well-known ones.

Certification is a different thing entirely

Age certification — U, UA, A in India, and their equivalents elsewhere — is frequently displayed alongside quality scores and measures something completely unrelated.

Certification describes suitability for age groups according to a board's criteria. It says nothing about whether a film is good. A certificate also does not transfer between countries: the same film can carry markedly different ratings across territories, reflecting different norms about violence, language and sexual content rather than anything about the film.

The one genuinely useful thing a certificate tells you is what kind of content to expect, which matters when watching with family and rarely otherwise.

What ratings are actually good for

Given all of that, a few uses where they hold up.

Filtering the bottom. Very low scores across both critics and audiences are reasonably reliable. Films that almost nobody liked usually have identifiable problems.

Confirming a hunch. If you are already inclined toward a film, a score tells you whether that instinct is widely shared, which is more useful than starting from the number.

Reading the gap. As above, disagreement between critics and audiences is more informative than either figure alone.

Comparing within a genre. Scores are more meaningful between similar films than across different kinds of film, because the audience and the expectations are closer.

What they are bad for

Ranking across genres. A comedy and a drama scoring the same number are not equivalently good; they satisfied different expectations.

Predicting your response to anything unusual. The more distinctive a film, the less an average predicts an individual reaction. Averages describe the middle of a distribution, and distinctive films have distributions with no middle.

Older films. Scores for films from decades ago are compiled from people who chose to seek them out — a self-selected and forgiving audience. Historical scores skew high for this reason.

The number we show, and what it is

For transparency: the rating shown on this site is an aggregate audience score supplied with our availability data. It is a reasonable general signal and carries every limitation described above.

We show it because it is genuinely useful for filtering, and because a catalogue with no signal at all is harder to navigate. We do not weight it heavily in how films are ordered, because ordering a free catalogue by score would bury exactly the unusual films that make browsing worthwhile.

A more useful habit

Ratings are best used to narrow rather than to decide.

Use them to remove obviously poor options, then choose from what remains on the basis of what the film actually is — subject, director, running time, when it was made. Those tell you more about whether an evening will suit you than a single averaged number ever can.

That is a slower method than sorting by score, and it consistently produces better evenings.

Why scores drift over time

A film's score is not fixed at release, and the direction it moves is fairly predictable.

Immediately after release, ratings come from the people most motivated to see it — fans of the director, the genre, the source material. That group is favourably disposed, so opening scores skew high.

As a wider and less invested audience arrives over the following months, scores usually settle downward. Then, for some films, they recover years later, as the only people still seeking out a decade-old film are those with a specific reason to.

The consequence is that a score has an implicit date attached that nobody displays. A number on a film released last month and a number on one released in 2010 were produced by structurally different audiences, and the second is generally the more stable figure.

The one number nobody publishes

The most useful metric would be completion rate: what proportion of people who started a film finished it.

Services have this data and do not share it. It measures something ratings cannot — whether a film holds you — and it is immune to most of the biases above, since abandoning a film is an honest act requiring no motivation to report.

Its absence is worth noting mainly to keep the published numbers in perspective. The figures a catalogue shows you are the ones convenient to collect, not the ones that would answer your question best.

A note on review bombing

Occasionally a film's audience score collapses for reasons unconnected to the film — a controversy about a cast member, a political reading of the content, or organised campaigning by a group with a grievance.

This is visible if you look for it. The distribution goes bimodal, with a large spike at the lowest possible score and another at the highest, and very little in between. A genuine mixed reception produces a spread; a campaign produces two towers.

Where a service shows the distribution rather than only the average, a glance at the shape is worth more than the number. Where it does not, a score wildly out of line with critical reception and with the film's apparent reputation is worth treating as suspect rather than informative.

The reverse also happens, less often: coordinated positive rating around a release. Both distort the same figure in the same way, and both are more common on recent films than older ones, which is another reason older scores are the steadier reading.

The simplest useful test

If you take one thing from this: look at how many people rated a film before looking at what they rated it. A modest score from many thousands of people is far more informative than an excellent score from forty, and catalogues display the two identically.