INDEPENDENT MINDS. EXCELLENT TELEVISION.THE PREVIEW EDITION / SEPTEMBER 2026
episodecompassMy queue 0
SOURCE NOTEBOOK

TV Shows & Streaming Database — Deep Research Brief, Part 06: Ratings, Reviews, Charts, and Viewership Data

99 preserved extracts. Research planning document; requested work is not evidence that the work was completed.

← Research library
99 extracts
  1. NOTE 3929

    TV Shows & Streaming Database — Deep Research Brief, Part 06: Ratings, Reviews, Charts, and Viewership Data

    Conduct exhaustive research on how TV series are rated by users and critics, how those scores are computed and manipulated, and how viewership, charts, and demand data are measured and published worldwide. The goal is to give the website a transparent, fraud-resistant rating methodology at show, season, and episode level; a clear understanding of which critic and chart data can be licensed or cited; a documented set of review-bombing and manipulation case studies; and a viewership charts hub that presents each source with correct methodology notes and limitations.

  2. NOTE 3930

    1. How to research this part

    Treat every methodology detail, threshold, figure, and date here as a lead to verify against the source's own methodology page, help center, or press release. Methodologies change without notice.

  3. NOTE 3931

    1. How to research this part

    Minimum counts: at least 15 rating systems documented; at least 20 measurement/chart sources across at least 8 countries; at least 30 review-bombing or manipulation cases (at least 15 TV-specific); at least 52 weeks of chart history for Nielsen and Netflix Top 10 analyzed.

  4. NOTE 3932

    1. How to research this part

    Prefer primary sources: IMDb Help Center, Rotten Tomatoes "About" and methodology pages, Metacritic "How we create the Metascore," Nielsen press releases and methodology docs, Netflix Top 10 site and "What We Watched" reports, BARB/OzTAM/Numeris/Médiamétrie/AGF/Video Research published methodology.

  5. NOTE 3933

    1. How to research this part

    Record "as of" dates for every score snapshot and chart value. Save the methodology version in effect.

  6. NOTE 3934

    1. How to research this part

    Do not scrape rating sites or charts in violation of terms; use licensed feeds, official downloads (e.g., Netflix Top 10 TSV files, verify availability), and IMDb non-commercial datasets only within their license.

  7. NOTE 3935

    1. How to research this part

    Never present third-party viewership estimates as exact; label panel estimates, census data, self-reported platform data, and demand proxies separately.

  8. NOTE 3936

    2.1 IMDb

    Weighted rating: IMDb states it applies undisclosed filters to reduce manipulation rather than a simple mean (verify current wording). The historical Top 250 formula WR = (v/(v+m))·R + (m/(v+m))·C (verify whether still published and current m values for Top 250 TV).

  9. NOTE 3937

    2.1 IMDb

    Episode ratings, season averages, demographic breakdowns (removed or changed? verify), rating distribution histograms, "unusual voting activity" alert notice (verify when introduced and example titles).

  10. NOTE 3938

    2.1 IMDb

    Questions: minimum vote counts for Top 250 TV; how episode ratings appear in datasets (title.ratings.tsv, title.episode.tsv) and license limits for commercial use.

  11. NOTE 3939

    2.2 Rotten Tomatoes

    Tomatometer: percentage of approved critics with positive reviews; season-level scores for TV (not series-level aggregate, verify); minimum review counts for a score and Certified Fresh TV thresholds (verify, e.g., 20+ reviews historically for a season).

  12. NOTE 3940

    2.2 Rotten Tomatoes

    Popcornmeter (renamed from Audience Score in 2024, verify); Verified Hot badge; verified-audience rules (ticket purchase via Fandango for film; how TV "verified" works or doesn't, verify).

  13. NOTE 3941

    2.2 Rotten Tomatoes

    Changes to watch: 2019 removal of pre-release audience comments and "want to see" scores (after Captain Marvel controversy, verify); Top Critics definitions; critic eligibility criteria changes (verify dates).

  14. NOTE 3942

    2.3 Metacritic

    Metascore: weighted average of critic scores normalized to 0–100, weights undisclosed; season-level TV scores; color bands (verify); minimum critic counts for TV (verify).

  15. NOTE 3943

    2.3 Metacritic

    User scores: 0–10; 2020–2023 changes such as waiting period before user reviews post-release (36 hours for games, verify whether TV too).

  16. NOTE 3944

    2.4 Community trackers

    Letterboxd (film-focused; verify any TV support), Serializd (season/episode ratings, half stars), Trakt (10-point hearts/ratings, community percent), Simkl, TMDB (user score %), TVmaze (verify), MyAnimeList (weighted score with minimum members, verify formula), AniList (mean vs. weighted average score), MyDramaList, Kitsu, Taste.io, Likewise (verify).

  17. NOTE 3945

    2.4 Community trackers

    Produce: comparison table — scale, granularity (show/season/episode), displayed metric, weighting, thresholds, publicly documented anti-fraud, API access, license for display.

  18. NOTE 3946

    3. Statistical methods

    Bayesian averages: prior mean C, prior weight m; show how m should vary by level (show vs. episode) and by genre population. Worked example: an episode with 12 votes averaging 9.6 vs. one with 4,000 votes averaging 9.1, with m = 50 and C = 7.4.

  19. NOTE 3947

    3. Statistical methods

    Wilson score lower bound for positive/negative ratios; credible intervals on 5-star distributions (Dirichlet prior).

  20. NOTE 3948

    3. Statistical methods

    Rater-bias normalization (per-user mean centering, z-scores), and effects on sparse data.

  21. NOTE 3949

    3. Statistical methods

    Recency/trending: exponential decay of activity (check-ins, ratings, list adds) with half-lives; compare Hacker News-style gravity and time-bucketed z-scores.

  22. NOTE 3950

    3. Statistical methods

    Season and series aggregates: episode-weighted vs. vote-weighted; how to handle finales with inflated votes.

  23. NOTE 3951

    3. Statistical methods

    Questions: which methods do IMDb, MAL, AniList, and Trakt use publicly; what thresholds prevent tiny-sample Top lists.

  24. NOTE 3952

    3.1 Episode heatmaps

    Research existing visualizations (SeriesHeat, "Graph TV" style sites, Reddit data-viz posts, verify names) and IMDb dataset-based tools.

  25. NOTE 3953

    3.1 Episode heatmaps

    Design questions: color scale (sequential with accessible contrast), handling missing and low-vote cells, seasons of different lengths, specials, anime absolute numbering.

  26. NOTE 3954

    3.1 Episode heatmaps

    "When does it get good": define metric (first sustained rise above show median for N episodes); worked example on a show commonly described as having a slow first season (lead: Parks and Recreation, The Good Place, Star Trek: The Next Generation; verify).

  27. NOTE 3955

    3.1 Episode heatmaps

    Case leads for dramatic dips: Game of Thrones S8, Dexter S8, How I Met Your Mother finale, The Walking Dead mid-seasons (verify vote patterns).

  28. NOTE 3956

    4.1 Case studies (verify every date, platform, and scale)

    The Rings of Power (Prime Video, 2022): Amazon delayed user reviews 72 hours; audience scores dropped amid racially motivated campaigns (verify).

  29. NOTE 3957

    4.1 Case studies (verify every date, platform, and scale)

    She-Hulk: Attorney at Law (Disney+, 2022): IMDb notice about unusual voting; score gaps (verify).

  30. NOTE 3958

    4.1 Case studies (verify every date, platform, and scale)

    Ms. Marvel (2022) IMDb demographic voting disparities (verify).

  31. NOTE 3959

    4.1 Case studies (verify every date, platform, and scale)

    The Last of Us episode 3 "Long, Long Time" (HBO, 2023) review bombing on IMDb (verify).

  32. NOTE 3960

    4.1 Case studies (verify every date, platform, and scale)

    The Acolyte (Disney+, 2024): Popcornmeter far below Tomatometer; cancellation debate (verify).

  33. NOTE 3961

    4.1 Case studies (verify every date, platform, and scale)

    The Witcher: Blood Origin and The Witcher S3/S4 recasting backlash (verify).

  34. NOTE 3962

    4.1 Case studies (verify every date, platform, and scale)

    Velma (HBO Max, 2023) extreme low IMDb score (verify).

  35. NOTE 3963

    4.1 Case studies (verify every date, platform, and scale)

    Obi-Wan Kenobi and harassment of actor Moses Ingram (2022, verify).

  36. NOTE 3964

    4.1 Case studies (verify every date, platform, and scale)

    Cowboy Bebop live-action (Netflix, 2021) (verify).

  37. NOTE 3965

    4.1 Case studies (verify every date, platform, and scale)

    Positive brigading/fandom inflation: K-drama and anime fandom vote campaigns on IMDb/MAL; Arcane, Attack on Titan finale rating swings on MAL (verify).

  38. NOTE 3966

    4.1 Case studies (verify every date, platform, and scale)

    Pre-release bombing: shows rated before premiere (research IMDb policy on pre-release voting for series).

  39. NOTE 3967

    4.1 Case studies (verify every date, platform, and scale)

    Non-TV context for methods: Captain Marvel (RT 2019), Brie Larson-related campaigns; Steam review-bomb labeling (2019) as a design precedent (verify).

  40. NOTE 3968

    4.2 Detection signals to research

    Velocity spikes vs. baseline; new-account share; accounts with single-title activity; 1/10 and 10/10 polarization; IP/device clustering (privacy-limited); off-site referral surges (Reddit, X, 4chan, Discord); coordinated text similarity in reviews; country/timezone anomalies; rating without watch progress.

  41. NOTE 3969

    4.2 Detection signals to research

    Platform responses: IMDb filters and notices; RT verified ratings; Amazon review delays; Steam off-topic review mode.

  42. NOTE 3970

    4.2 Detection signals to research

    Produce: fraud-signal catalog with each signal's false-positive risk and privacy implications (note where a lawyer should review data retention).

  43. NOTE 3971

    4.3 Paid and bot campaigns

    Research documented cases of paid ratings or fake reviews in entertainment (e.g., Rotten Tomatoes critic-payment story around Ophelia/Bunker 15, reported 2023 by Vulture, verify) and FTC rule on fake reviews (final rule 2024, verify) as it applies to review platforms; note lawyer review.

  44. NOTE 3972

    5.1 Nielsen

    Streaming Top 10: minutes viewed on TV sets, weekly, ~4-week lag (verify); includes which platforms; exclusion of mobile/computer viewing (verify).

  45. NOTE 3973

    5.1 Nielsen

    The Gauge: monthly share of TV usage by broadcast, cable, streaming, and by distributor (Media Distributor Gauge). YouTube's leading share (verify recent %).

  46. NOTE 3974

    5.1 Nielsen

    Big Data + Panel: accreditation by MRC (Jan 2025, verify) and transition as currency.

  47. NOTE 3975

    5.1 Nielsen

    ARTEY year-end report: 2025 edition naming Bluey the top streaming title (verify minutes).

  48. NOTE 3976

    5.1 Nielsen

    Questions: how Nielsen credits multi-platform titles (e.g., Grey's Anatomy on Netflix + Hulu combined?), how it treats YouTube content, what data can be cited vs. licensed.

  49. NOTE 3977

    5.2 Platform self-reported data

    Netflix Top 10 (weekly; views = hours ÷ runtime since June 2023, verify); Most Popular lists (91-day window); country top 10s in ~90 countries (verify count).

  50. NOTE 3978

    5.2 Platform self-reported data

    Netflix engagement report "What We Watched": launched Dec 2023, semiannual; moved to once a year July 2026 (verify).

  51. NOTE 3979

    5.2 Platform self-reported data

    Other platforms: Prime Video, Disney+, HBO Max, Paramount+, Peacock, Apple TV in-app top 10 rows (algorithmic, not published numbers); occasional press-release claims ("most-watched premiere") — record definitions (e.g., "global viewers in first 3 days," multi-platform).

  52. NOTE 3980

    5.2 Platform self-reported data

    Produce: claim-definition table for at least 40 platform viewership press claims 2024–2026.

  53. NOTE 3981

    5.3 Third-party

    Luminate Streaming Viewership (launched 2023, verify methodology—ACR smart-TV data).

  54. NOTE 3982

    5.3 Third-party

    Samba TV (ACR households, first-weekend), Comscore, iSpot/605 (verify), Antenna (subscriptions, not viewing), Parrot Analytics Demand Expressions (demand proxy from social/search/piracy signals), Reelgood and JustWatch popularity charts (user-activity on their apps), Whip Media/TV Time charts (discontinued with TV Time shutdown July 15, 2026, verify), Trakt trending.

  55. NOTE 3983

    6. International measurement

    UK — BARB: consolidated 7- and 28-day viewing, four-screen dashboard, SVOD measurement via router meters (verify), BARB Top 50 weekly.

  56. NOTE 3984

    6. International measurement

    Australia — OzTAM and VOZ (Virtual Australia total TV) with Nielsen/Regional TAM; VOZ national streaming reports (verify brand changes).

  57. NOTE 3985

    6. International measurement

    Canada — Numeris (PPM-based, VAM cross-platform, verify).

  58. NOTE 3986

    6. International measurement

    France — Médiamétrie (Médiamat, 4-screen, streaming SVOD measurement from 2024, verify).

  59. NOTE 3987

    6. International measurement

    Germany — AGF Videoforschung (with GfK/NIQ panel, verify), including streaming measurement.

  60. NOTE 3988

    6. International measurement

    Japan — Video Research (real-time and time-shift ratings, TVer data, verify).

  61. NOTE 3989

    6. International measurement

    Also research: South Korea — Nielsen Korea and TNMS ratings; FlixPatrol-style aggregations (verify terms); India — BARC; Brazil — Kantar IBOPE Media; Spain — Kantar/GECA; Italy — Auditel; Nordics — Mediavision/Kantar (verify).

  62. NOTE 3990

    6. International measurement

    For each: owner/governance (JIC model), panel size, what counts (live, VOSDAL, time-shift, BVOD, SVOD), reporting lag, public vs. paid data, citation rules.

  63. NOTE 3991

    7. Presenting charts responsibly

    Chart card fields: source, metric definition (minutes, hours, views, households, demand multiplier), window, country, device scope, publication date, methodology link, known limitations, comparability flag.

  64. NOTE 3992

    7. Presenting charts responsibly

    Rules: never mix minutes and views in one ranking; never compare Netflix "views" to Nielsen minutes without conversion caveats; show platform self-reporting labels.

  65. NOTE 3993

    7. Presenting charts responsibly

    Worked example: a single show's week in Netflix Top 10 (views), Nielsen Streaming Top 10 (minutes, 4 weeks later), and Parrot demand (multiplier) presented side by side with correct notes.

  66. NOTE 3994

    7. Presenting charts responsibly

    Research how Variety, Deadline, and What's on Netflix present these, and common misreadings.

  67. NOTE 3995

    7. Presenting charts responsibly

    Produce: chart style guide and a methodology note template (150 words max each).

  68. NOTE 3996

    8. Licensing critic scores and reviews

    Rotten Tomatoes (Fandango-owned; verify ownership after Versant spin-off—Fandango/RT status in Comcast/Versant, verify) data licensing options and logo usage rules.

  69. NOTE 3997

    8. Licensing critic scores and reviews

    Metacritic (Fandom-owned since 2022 acquisition from Paramount/Red Ventures chain, verify) licensing.

  70. NOTE 3998

    8. Licensing critic scores and reviews

    Critic review linking and short excerpt quoting limits; fair use not a safe harbor—note lawyer review.

  71. NOTE 3999

    8. Licensing critic scores and reviews

    Alternatives: build own critic aggregation by linking to reviews with headlines only; partner APIs; OpenCritic-style (games) model as precedent.

  72. NOTE 4000

    8. Licensing critic scores and reviews

    Questions: costs, attribution, display limits, cache duration, whether scores can be shown on SEO pages.

  73. NOTE 4001

    9. Proposed site rating methodology

    Scales: 10-point (half stars on 5) for shows, seasons, episodes; optional thumbs for quick episode reactions.

  74. NOTE 4002

    9. Proposed site rating methodology

    Eligibility: episode ratings only after marking watched; season ratings after ≥ 50% episodes watched (test threshold); show rating after one full season or ≥ 5 episodes; no pre-release ratings (anticipation "want" signal kept separate).

  75. NOTE 4003

    9. Proposed site rating methodology

    Weighting: verified-progress weight (scrobbled > manual check-in > bulk-marked), account age and history, trust score; cap weight per account.

  76. NOTE 4004

    9. Proposed site rating methodology

    Aggregation: Bayesian average with level-specific priors; display raw distribution and vote count; confidence badge below threshold.

  77. NOTE 4005

    9. Proposed site rating methodology

    Anti-fraud: velocity freezes with public "under review" notice; delayed publication of reviews for new titles (e.g., 24–72 hours, test); moderation of harassment; spoiler tagging.

  78. NOTE 4006

    9. Proposed site rating methodology

    Critics vs. audience: keep separate; never blend.

  79. NOTE 4007

    9. Proposed site rating methodology

    Transparency: public methodology page with changelog; annual audit.

  80. NOTE 4008

    9. Proposed site rating methodology

    Produce: formula spec, test dataset of 1,000 synthetic ratings including bombing scenarios, and expected outputs.

  81. NOTE 4009

    10. Timeline 2019 – 2026

    Log: RT audience changes (2019), Popcornmeter rename (2024), Netflix Top 10 launch (2021) and views metric (2023), engagement reports (Dec 2023; annual July 2026), Nielsen Streaming Top 10 launch (2020, verify), Big Data + Panel accreditation (2025), Luminate streaming launch (2023), Metacritic ownership changes (2022–2023), major review-bombing incidents (2022–2024), TV Time/Whip Media chart shutdown (2026), FTC fake-review rule (2024) — all (verify).

  82. NOTE 4010

    11. Deliverables for Part 06

    Rating-system comparison database (15+ systems) with methodology versions and sources.

  83. NOTE 4011

    11. Deliverables for Part 06

    Statistical methods guide with worked examples and code-ready formulas.

  84. NOTE 4012

    11. Deliverables for Part 06

    Episode heatmap design spec and "when does it get good" metric definition.

  85. NOTE 4013

    11. Deliverables for Part 06

    Review-bombing/manipulation case database (30+ cases).

  86. NOTE 4014

    11. Deliverables for Part 06

    Fraud-signal catalog with privacy notes.

  87. NOTE 4015

    11. Deliverables for Part 06

    Viewership source directory (20+ sources, 8+ countries): owner, method, scope, lag, access, citation rules.

  88. NOTE 4016

    11. Deliverables for Part 06

    Platform viewership-claim definition table (40+ claims).

  89. NOTE 4017

    11. Deliverables for Part 06

    Chart presentation style guide and methodology note templates.

  90. NOTE 4018

    11. Deliverables for Part 06

    Critic-score licensing options memo (with points for lawyer review).

  91. NOTE 4019

    11. Deliverables for Part 06

    Proposed site rating methodology spec plus synthetic test dataset.

  92. NOTE 4020

    11. Deliverables for Part 06

    Glossary of at least 50 terms (Bayesian average, VOSDAL, ACR, minutes viewed, views, JIC, consolidated, etc.).

  93. NOTE 4021

    11. Deliverables for Part 06

    Annotated source list of at least 70 sources with access dates.

  94. NOTE 4022

    Research standards

    State the research date; record the methodology version and access date for every score and chart.

  95. NOTE 4023

    Research standards

    Cite methodology pages and official releases directly; use trade press only for context.

  96. NOTE 4024

    Research standards

    Label every number as census, panel estimate, platform self-reported, or demand proxy.

  97. NOTE 4025

    Research standards

    Do not claim a review-bombing incident is coordinated without evidence; describe what was observed and who said what.

  98. NOTE 4026

    Research standards

    Protect user privacy in fraud detection; document data minimization.

  99. NOTE 4027

    Research standards

    Do not scrape scores; document licensed or permitted access paths and note where a lawyer should review.