How the Celluloid Score Works: Four Inputs, One Verdict

Desk-researched August 2026. Celluloid Score is our own composite — not a Rotten Tomatoes Tomatometer, not a Metacritic Metascore, not an IMDb average. Every number on a title page is labelled. If an input is missing, we omit it; we never store a zero to mean “unknown.”

Should you use Celluloid Score in 2026?

Yes — as a decision layer, not as a replacement for reading a review. Celluloid Score averages up to four labelled inputs on a 0–100 scale, then tells you who the film is not for. Skip it if you only want a binary Fresh/Rotten badge; Rotten Tomatoes still owns that job.

The rest of this guide is the arithmetic, the Hindi exception, a worked example from published title pages, and the cases where the number is the wrong tool.

What is the Celluloid Score and how does it work?

Celluloid Score is a 0–100 composite published only after a title has screened and we have at least two real inputs. It averages the available Celluloid Editorial Rating, Metascore, Letterboxd (×20) and IMDb (×10). Audience sentiment is shown on the page and excluded from the average.

A higher number means broader agreement that the title is worth your time — across an editor, a critic aggregator, and two public rating traditions. It is not a poll of “all critics,” and it is not a claim that we sat in every screening.

Input	Scale we store	How it enters the average	In the composite?
Celluloid Editorial Rating	0–100	As-is	Yes
Metascore	0–100	As-is	Yes, when Metacritic has one
Letterboxd	0–5 stars	×20	Yes, when present
IMDb	0–10	×10	Yes, when present
Audience snapshot	0–100	Display only	No

Is Celluloid Score legit, or just another aggregator?

It is a first-party editorial composite with named provenance — legit as a method, not a substitute for a studio’s own data. Kaiser Khan is the accountable editor. The formula is public on this page and on the scoring policy. We do not sell rankings.

What it is not: a Tomatometer, a Metascore, or an IMDb average wearing our logo. Those products remain cited as inputs, with their own scales, on every scored title page. Supplier scores are never marked up as our aggregateRating — that is a manual-action risk, and we refuse it.

The honest caveat: we have not invented a fifth “secret” source. If a page shows three inputs, the average is of those three. If a regional film has no Metascore, we do not pretend Metacritic covered it.

How is it calculated? A worked example

Take the inputs published on our Oppenheimer page (checked 12 August 2026): editorial 93, Metascore 90, Letterboxd 4.2, IMDb 8.3. Convert the last two — 4.2 × 20 = 84, 8.3 × 10 = 83 — then average four values: (93 + 90 + 84 + 83) ÷ 4 = 87.5, which rounds to 88.

That is the non-Hindi path: equal weights, missing inputs omitted, then round to a whole number. The audience snapshot on that page (91%) is visible and not in the 88.

A second check, Barbie (same date): editorial 88, Metascore 80, Letterboxd 3.9 (78), IMDb 6.9 (69) averages to 79. The gap between a pink-event smash and a more divided IMDb crowd is the point of showing every input — the composite is not a hiding place.

Unreleased films never carry a Celluloid Score. “Coming Soon” is a fact about the calendar, not a 0%.

Why exclude the audience snapshot?

Because coverage and sample provenance vary too much by market to rank on. A US theatrical audience score and a regional platform score are not the same measurement. We still show the snapshot so you can see critic/audience divergence; we refuse to let an inconsistent sample move the ranking.

When critics and audiences disagree by a wide margin, that sentence is often more useful than the composite. Title pages spell the gap in plain language. The composite stays the four-input average so one noisy channel cannot dominate.

Why average four inputs instead of picking Rotten Tomatoes?

Because any single score can mislead, and RT’s job is a different question. A review-bombed — or review-bought — film can warp IMDb while Letterboxd and Metascore hold. Averaging independent measurements rewards films that win more than one constituency.

Rotten Tomatoes answers “what share of critics were net-positive?” Celluloid Score answers “how do an editor, a critic aggregator, and two public diaries land on the same 0–100 scale — and who should skip this?” Those are complementary, not rivals. We link the scoring policy rather than implying we replaced RT.

How do Hindi and Bollywood titles get scored?

They use the same four inputs, then a reliability pass: Letterboxd and Metascore count more, IMDb counts less when it runs hot, and a small haircut applies so local inflation alone cannot mint a Celluloid Pick. Festival and arthouse Indian films with strong international Metascores are largely unaffected.

Industry reporting has documented paid influencer reviews and coordinated IMDb campaigns around Hindi theatrical releases; IMDb has previously flagged unusual voting on Indian titles. We do not claim we can see every brigade. We claim the composite will not treat a sky-high IMDb number as decisive when Letterboxd and Metascore disagree.

Weights on that path, as implemented: Letterboxd 1.4, Metascore 1.35, editorial 0.85, IMDb 0.45, then multiply the weighted average by 0.96. Details and the “why there is no score yet” empty state: scoring policy and the India hub.

What counts as a good Celluloid Score?

90+ is a Celluloid Pick; 60–89 is Recommended; below 60 is Not Recommended. Those are our labels, not Rotten Tomatoes’ Certified Fresh. A 61% and an 89% are both “Recommended” — read the inputs before you treat the tier as a substitute for taste.

Tier	Range	What it means	Who it is for
Celluloid Pick	90–100	Exceptional across the inputs we have	Viewers who want a high-consensus bet
Recommended	60–89	Generally worth watching	Most people, if the genre fits
Not Recommended	0–59	Weak reception overall	Only if you have a specific reason to watch anyway

One important rule: absence of a score is not a zero. Review Pending means we have not published a composite yet. It does not mean the film failed.

Celluloid Score vs Rotten Tomatoes vs Metacritic vs IMDb — which should you use?

Use Celluloid Score to decide; use RT/Metacritic/IMDb to see the raw constituencies. Start here for a verdict and a not-for line, then open the title page’s input row if you distrust the blend. No tool wins every film.

	Celluloid Score	Rotten Tomatoes	Metacritic	IMDb
Question it answers	Is it worth watching, and for whom?	What share of critics were positive?	Weighted critic average	Broad user average
Scale	0–100 composite	% + audience %	0–100 Metascore	0–10
Says “skip it”?	Yes — not-for is required on scored pages	Rarely	No	No
Provenance on our pages	Named inputs + method URL	Not our product	Cited as an input	Cited as an input
Markup we emit	Own reviews only, never supplier aggregateRating	—	Visible text, not our aggregate	Visible text, not our aggregate

When is Celluloid Score NOT the right tool?

Skip the composite when you need a single-constituency read, a live box-office number, or a film that has not screened. You can still use the title page for cast, runtime and where-to-watch — just do not treat a missing score as a verdict.

• You only care what “Top Critics” thought → Metacritic or RT, not our blend
• You want opening-weekend gross or Indonesian admissions → those are different facts; we publish admissions where we have them, never as a substitute score
• The film is unreleased → no Celluloid Score, by design
• You already know you hate the genre → a 92% horror Pick is still the wrong Friday night
• You need a phone number for a streamer → we are not customer support

Who is Celluloid Score NOT for?

It is the wrong ranking if you want one critic’s voice with no maths, or if you want a score you can buy. You can still read the review. Being honest about this is what makes the number trustworthy.

• You want only Kaiser Khan’s 0–100 with no Metascore or IMDb in the room — that number is on the page as Editorial Rating; the composite is the blend
• You want rankings for sale to studios — we do not sell them; paid placements never move Celluloid Score
• You need first-hand “I was at the junket” colour we did not have — we will not invent it
• You treat any 90+ as a moral obligation to watch — Picks are high-consensus, not homework

The verdict: should you use Celluloid Score?

Yes — as the first number on the page, with the inputs still visible, when you want a decision rather than a listing. Search the title, read the Quick Verdict and the not-for line, glance at the four inputs, and only then hit play. Who should skip the composite entirely: anyone who needs a binary Fresh badge or a film that has not opened.

Its strengths are the public formula, the Hindi reliability pass, honest empty states, and Markdown fact sheets at /movies/{slug}.md for machines that should not have to parse our HTML. The people who get a bad recommendation from Celluloid Score almost always ignored the not-for line or treated Review Pending as a zero — two mistakes this guide exists to prevent.

After you have a verdict, compare two titles, browse Celluloid Picks, or read how we handle mistakes on corrections.

Does a high Celluloid Score mean I will like it?

No. It means independent measurements agreed the title is well regarded. Taste still sits with you. A 91% slow cinema Pick is a good bet for that audience and a bad Friday-night popcorn bet — which is why scored pages name who should skip it.

If the not-for line includes you, trust that line over the badge.

Can studios pay for a higher Celluloid Score?

No. Screeners, junkets and hosted travel do not buy a score. Material support is disclosed when relevant. Rankings are not for sale. See the editorial policy.

Where do I see the raw inputs?

On every scored movie and show page, in the score breakdown, and in the Markdown fact sheet (/movies/{slug}.md or /shows/{slug}.md). If an input is missing, the cell is empty — not a greyed-out zero.

How often do scores change?

When an input we publish changes and we verify it. We update the visible “Last reviewed” date and sitemap lastmod only when the page actually changed. A lastmod that lies is worse than one that is old.