Movie Ratings Explained: What TMDB, IMDb, and Rotten Tomatoes Actually Measure

Almost everyone uses a rating to decide, and almost nobody knows what the number is measuring. That is a genuine problem, because the three most common scores measure three completely different things and produce numbers that look interchangeable and are not.

This guide covers what each one actually calculates, the specific ways each is misleading, and a practical method for using them that takes about ten seconds per film.

Rotten Tomatoes: a headcount, not a quality score

This is the most misunderstood number in film. The Tomatometer is the percentage of collected critic reviews that were positive. That is all it is. It does not measure how positive, how enthusiastic, or how good the film is.

Each review is sorted into fresh or rotten, a binary. A review that says "a solid, unremarkable film, worth your time" counts identically to a review calling it a masterpiece. So a film where a hundred critics all thought it was a decent 7 out of 10 scores 100%, while a film where sixty critics called it a 10 and forty called it a 4 scores 60%.

The site does publish an average critic score alongside the percentage, usually out of 10, and that number is considerably more informative. Almost nobody looks at it. If you use Rotten Tomatoes at all, use that one.

There is also a separate audience score, which measures something else again and has historically been vulnerable to coordinated review campaigns before a film is widely released.

IMDb: a weighted average of public votes

IMDb ratings are a 1 to 10 average of votes from registered users, with an undisclosed weighting applied to reduce the effect of vote manipulation. The company does not publish the formula, which means the displayed number is not simply the arithmetic mean and cannot be independently reconstructed.

Three known distortions are worth understanding.

• Extreme compression. Almost every film anyone has heard of sits between 5.5 and 8.5. The effective range is about three points wide, so a difference of 0.3 is much more significant than it looks.
• Recency and fandom inflation. Films score highest in the weeks after release, when the people voting are the ones who sought them out. Scores typically settle downward over the following year.
• Demographic skew. The voting population is not representative, and it shows up systematically: action and science fiction score above their critical reception, while romance and musicals score below.

The 1 and 10 votes are also a real factor. A meaningful percentage of votes on any popular film are at the extremes, cast by people expressing an opinion about the film's existence rather than assessing it.

TMDB: the same idea, a smaller and different crowd

TMDB, which supplies the data on this site, uses a straightforward average of user votes on a 1 to 10 scale, with a minimum vote threshold before a score is displayed prominently. The formula is documented and the data is openly available, which is a meaningful advantage.

The voting population is smaller and skews international and more toward people who use the API and contribute metadata. In practice TMDB scores tend to run slightly lower than IMDb on mainstream blockbusters and slightly higher on non-English films, which is a useful corrective.

Every rating displayed on teasertrack comes from TMDB, and every movie page shows the vote count next to it, because the count is the part that determines whether the score means anything.

The three questions to ask about any score

1. How many votes?

This is the single most useful check and it takes two seconds. A score built on a few hundred votes is noise. Below roughly a thousand votes, treat the number as an indication rather than a measurement. Above about ten thousand, it is reasonably stable.

This matters most for older, foreign-language, and independent films, which routinely have excellent scores from small vote pools. That is not a reason to distrust them, it is a reason to read the score as "the people who sought this out liked it," which is different from "you will like it."

2. What genre is it?

Genres have different baselines and comparing across them is meaningless. Horror and comedy systematically score lower than drama, for structural reasons.

The horror case is the clearest example of the problem. Hereditary sits around 7.3, below plenty of films it is obviously better made than, in significant part because a substantial number of viewers rate it low for succeeding at what it set out to do. The Babadook and It Follows have the same pattern.

3. Is the film divisive or merely fine?

An average cannot distinguish between a film everyone thought was a 7 and a film half the audience rated 10 and half rated 4. Both display as 7.0, and they are completely different propositions.

Where a rating distribution is available, look at it. A bimodal distribution, with peaks at both ends, means the film is doing something specific that some people will love. Those are frequently the most interesting films on any list, and the average is the least useful thing about them.

A worked example

Take three films with similar TMDB scores and notice how differently they should be read.

Hereditary

2018 · 7.3/10 · 2h 8m

Whiplash

2014 · 8.4/10 · 1h 47m

Nightcrawler

2014 · 7.7/10 · 1h 58m

Whiplash at 8.4 is a drama with a very large vote count, so the score is stable and comparable to other dramas: it is genuinely near the top of its category. Nightcrawler at 7.7 is a thriller with a large vote count and a deliberately unpleasant protagonist, which suppresses the score relative to its quality. Hereditary at 7.3 is horror, so against the horror baseline it is exceptional rather than merely good.

Read purely as numbers, those three look like a clear ranking. Read with genre and vote count in mind, they are all in roughly the same tier.

The ten-second method

1. Check the vote count first. Under a thousand, treat the score as a hint.
2. Adjust for genre. Add roughly half a point mentally for horror and comedy, subtract a little for action.
3. Ignore the Tomatometer percentage. Use the average critic score instead if you use the site at all.
4. Read two sentences of a review, not the number. A single specific sentence about what a film is like beats any aggregate.
5. Weight your own history heaviest. If you have liked four films by a director, that is better evidence than any score.

When ratings are actually useful

They are good at one thing: filtering out the genuinely bad. A film with a very low score and a large vote count is almost certainly not for you, and that is real information. Ratings work as a floor, not as a ranking.

They are also useful in aggregate across a catalog. Browsing top rated movies will not tell you which single film to watch tonight, but it is a reasonable place to find something you have not heard of that is unlikely to be a waste of time.

Beyond that, a specific recommendation from someone whose taste you understand outperforms every aggregate on this page. That is why our ranked guides explain what each film is like rather than just listing scores: the sentence about the film is doing the work, not the number.