How we work

TV Reviews is an independent directory and research tool for TVs and home entertainment gear. It organizes every product around the exact model number you can buy, because the same TV often carries different model numbers by size, year and region.

The aim is to help you decide, with the numbers and where they came from in view. It is not a lab, and the next sections say exactly what that means.

Two kinds of number appear on product pages, and they are never blended.

• Measured A measured figure was produced by an independent tester with instruments, and we name the tester. Measured values are drawn as solid marks.
• Claimed A claimed figure is what the manufacturer says. Claimed values are drawn as hollow marks.

When both exist, you see both. When neither exists, the page says the figure is not published instead of guessing. If several testers measured the same thing, we show the average and list their names.

Specifications come from manufacturers’ own product and support pages, with a source recorded for the facts that are easy to get wrong. Where a maker’s site cannot be read, we use a press review that prints a specification table and cite that page.

Measurements come from the published lab figures of Tom’s Guide and TechRadar, which both state the instruments and conditions they used. We record which size was tested, because a measurement for one size does not always apply to another.

Reviews and community feedback are paraphrased, attributed and linked to the original, so you can read the source. We do not republish them. A short quotation appears only when an editor adds it deliberately.

Video reviews are links to creators’ public videos. Each one is checked to exist and to be about the product it is attached to.

• No first-hand lab testing yet. We have no measurement lab of our own at the moment. Anything described as measured here was measured by someone else, and we say who. If that changes, those results will be labeled separately.
• No RTINGS data. RTINGS keeps its detailed results behind a paid membership, and we do not use or copy them.
• No invented numbers. A missing specification stays blank, and the page says it is not published.
• No paid rankings. Nobody can buy a position on a list or a score.
• No “we tested” claims. We describe what a named tester found, or what the manufacturer states.

Scores run from 0 to 10 and are calculated by a fixed formula from published specifications and independent measurements. They are not lab results, and every one is labeled as a calculation. The same inputs always produce the same score, and each product page lists the inputs behind each number and whether each was measured, claimed, a specification or a price.

• Movies & dark rooms. Black level carries the most weight. It starts from the display type, since self-lit OLED pixels can switch fully off, and for LCD from the panel type, the backlight design and the maker’s dimming zone count. Independent contrast measurements adjust it. HDR peak brightness, DCI-P3 color coverage and Dolby Vision support are added when known.
• Bright rooms. SDR brightness leads, taken from an independent measurement when one exists and from the maker’s claim otherwise. Claimed figures count for less than measured ones. A measured low reflection result, or a documented anti-glare finish, adds a small bonus.
• Sports. Native refresh rate, pixel response time when measured, viewing angle and daytime brightness. Viewing angle comes from a measurement when there is one and from the panel type when not.
• Gaming. HDMI 2.1 port count, the highest 4K refresh rate, variable refresh rate formats and Auto Low Latency Mode. Measured input lag is included when an independent tester published it.
• Broadcast TV. Whether the TV has a built-in tuner, and which kind. A TV with ATSC 3.0 scores highest, one with no tuner scores lowest. Brightness and viewing angle add to it.
• Streaming. HDR format coverage (Dolby Vision, HDR10+ and HLG), casting support through AirPlay and Google Cast, and the movie score.

Overall blends the use-specific scores, with the largest weights on movies in a dark room, then bright rooms and gaming, then sports, streaming and broadcast. Value compares each TV’s overall score with its checked price, measured against other TVs of a similar size, so a 65-inch set is judged against other 65-inch sets. Only TVs with a real, dated price get a value score.

When a specification is unpublished, the formula either leaves that input out or uses a middle-of-the-road assumption, and the confidence note drops from “Mostly independent measurements” to “Some” or to “Published specs only”. A high score with low confidence deserves more caution than the same score with measurements behind it.

Matches in the Finder and the Builder use the same scores and the same plain rules, so they behave the same way every time. No AI decides a score, a ranking or whether two products are compatible. The optional Ask feature answers from the catalog and says when it does not know.

Every price shows where it was seen and the date it was last checked. Prices change often, so treat them as a guide and confirm at the retailer before you buy. A list price, which is the manufacturer’s suggested price, is labeled as one and is not presented as a street price.

Amazon prices are not shown. Amazon’s affiliate rules make it impractical to display its prices accurately, so product pages link to Amazon without a price, and say “check retailers” when we have no other price.

Where a TV is sold in several sizes, each size has its own model number and price. Price-based lists use the size that fits the list.

Facts that depend on size, such as dimensions, weight, mounting pattern and sometimes panel type, are stored per size. A feature that differs by size or region is shown as a difference, not hidden. The Model Decoder checks a number against the catalog first and only then falls back to naming patterns, and it says when it cannot read a number.

Some links to retailers are affiliate links, which can earn a commission when you buy. That never changes what we cover or how we rank. The disclosure page explains it in full.

If a specification, price, model number or measurement looks wrong, tell us and, if you can, include a link to the page that shows the right figure. Corrections are made when they are verified against a source.