A breakdown of the different types of rating figures found for US Television. We show the total viewer figure and the mosrt commonly used Nielsen A18-49 chart position.

Rating: Ratings are essentially percentages, measuring the portion of a given group — be it households, adults 18-49 or women 25-54 — watching a given show. Adults 18-49 is the primary demographic by which ad rates are set for entertainment programming, so it's the most commonly reported (one point in that demo equals 1.28 million people). So a 2.0 rating for The Masked Singer means that 2 percent of people in that age range, roughly 2.56 million people, watched the show.

Share: The percentage of a given group who are watching TV at that time and are tuned into a given program. Wednesday's Masked Singer had a 10 share in adults 18-49 (10 percent of adults under 50, who had their TVs on at that hour, watched it). It's typically written as "rating/share," so 2.0/10 for The Masked Singer.

Total viewers: Pretty self-explanatory — the average number of people watching a program in any given minute while it airs.

Overnight metered market ratings: These are the first ratings released each morning — or they were, anyway, until Oct. 3. Nielsen is planning to include out-of-home viewing in these numbers from now on (the first day of the new system didn't go well), which means they'll be released around midday now. Metered market ratings only take measurements from 44 markets (56 previously) for households and 25 markets for adults 18-49, so they're best considered as a first draft on how programming performed rather than definitive. They had been useful for gauging live events since they measure programs instead of just time periods.

Live-plus-same-day: The ratings that get reported each day, first as "fast nationals" in the morning and then as final numbers in the afternoon. They include both live viewing from the previous night and delayed viewing until 3 a.m. local time. Fast nationals are generally pretty accurate for entertainment programs, with occasional small adjustments in the finals.

Live-plus-3: Same-day ratings with three additional days of DVR and on-demand viewing added in. The majority of delayed viewing that Nielsen measures happens in this timeframe, with most shows growing their audiences by a good amount.

Live-plus-7: The same as live-plus-3, extended to a full week. In the 2018-19 season, two dozen series at least doubled their 18-49 ratings after seven days.

C3 and C7 ratings: Arguably the most important ratings numbers that the public doesn't usually see. These ratings track the number of viewers who actually watch commercials — which is why Nielsen ratings exist in the first place — over three or seven days. They play a big role in setting rates for advertisers buying commercial time. The occasional glimpses at C3 and C7 ratings in recent years have suggested they're higher than same-day numbers but a good distance short of live-plus-3 and live-plus-7 numbers.

Live-plus-35: An even longer-tail measurement that takes into account viewing that happens up to five weeks after a show airs. It's not a huge piece of the viewing pie, but it's not tiny, either.

Multiplatform ratings: Things can get a bit fuzzy here, as multiplatform ratings can include streaming and digital viewing via a network's app or third-party service like Hulu, plus on-air replays. The digital audience is growing — some shows get more viewers there than from their on-air showings — but no company in the business willingly offers up definitive streaming or digital viewership. It's only included as part of a whole. (It is possible to subtract, say live-plus-7 ratings from a multiplatform total to get a rough estimate of how many people watch something via nontraditional platforms).

Furthermore, each network has its own way of calculating cross-platform viewing, and timeframes can get murky. HBO touted a massive audience of 44 million viewers for the final season of Game of Thrones, but that included up six weeks of streaming and replays of the season premiere, five weeks of episode two and so on.

Streaming ratings: Are not really a thing. Nielsen does measure the audience for streaming shows, but Netflix and other platforms have disputed the ratings service's numbers as they don't take into account viewing on other devices.

Netflix has reported some viewership figures in recent quarterly earnings reports, but they're not really analogous to Nielsen ratings. Netflix considers a piece of content as having been "viewed" when a member account watches at least 70 percent of one episode of a series or 70 percent of a feature film. It also counts subscribers around the world rather than just the domestic viewers that Nielsen measures. The numbers can be useful in comparing one Netflix show to another, but the service has thus far only publicly released highlights, not a full tally.

For live events that include a streaming option, networks or other providers will often cite an "average minute audience" for a live stream. That's the closest thing to Nielsen's average total viewers statistic.

Social ratings: Nielsen measures social engagement around TV shows, counting the number of posts about a given episode and the reach of the conversation. As with all ratings, higher is better, but heavy social conversation and high on-air ratings don't necessarily go hand in hand.

Third-party measurements: A number of companies measure things like out-of-home viewing or binge viewing, but they can rely on users to opt in to sharing data, which can lead to a less representative sample.