How the Sence Fan Index works
Where the comments come from and how to read every number on this site.
Where the data comes from
The comments behind this site come from the official Facebook pages of all 138 FBS football programs, covering the 2025–26 season from August 23, 2025 to January 19, 2026. Fan pages, subreddits, and burner accounts are out of scope; this is what fans say in each program’s own front yard.
“Classified by Sence” means each comment was read by Sence’s classification models, which assign exactly one sentiment label and any number of tags from the system taxonomy Sence runs for sports customers. The figures on this site are those labels, aggregated; every comment in the period is counted, with no sampling and no extrapolation.
Comments quoted on this site are reproduced verbatim, truncated for length where needed. Commenter names are screened out before anything is stored or displayed, and no comment is linked back to the person who wrote it.
Two things on the site are not comment data. Schedules and results come from the season schedule; the coordinates behind the map are the published location of the stadium each team played its home games in, taken from that venue’s Wikipedia article, on state boundaries from the US Census Bureau via us-atlas.
How to read the numbers
Rates on this site follow Sence’s Sence Score convention: comments matching a tag per 1,000 comments analyzed, to one decimal. Rates rather than raw counts, so a large fanbase can’t top a board on volume alone.
A tag group’s total counts distinct comments; a comment tagged both Parlay and Moneyline counts once toward Sports Betting. Subtag figures count tag applications, so the subtags under a group can add up to more than the group’s total. Wherever subtags are itemized, that caveat is restated.
Fan Score is positive comments per 1,000 analyzed. It measures the rate of positivity in a comment section, not quality, loyalty, or anything grander.
Not every team is ranked on every metric. Teams below 1,500 analyzed comments aren’t ranked on sentiment metrics; tag-group rankings require 2,000 comments and at least 25 comments in the group; subtag rankings require 2,000 comments and at least 15 applications of the subtag. A team below a floor is excluded from that board entirely (shown as “low volume,” never ranked last), and every rank states its cohort (“82nd of 133”), which is why cohort sizes vary from metric to metric. Week-over-week mood swings carry their own floor: the Biggest Win and Deepest Despair boards only count a swing when both weeks cleared 50 comments and the later one contained a game.
We report abuse rates because they’re the part of the comment section nobody screenshots. They are why moderation exists, and they’re measured the same way as everything else here: per 1,000 comments, no extrapolation.
What people call these things
The site’s own words are toxicity, harm, abuse, negative, and explicit, because those are the classifier’s words. Most people asking these questions use different ones. Here is the honest mapping between the two.
- Angriest, most negative
- The negative board: negative comments per 1,000, across the full season.The most negative fanbases in college football
- Most miserable
- The season superlative for the bleakest comment section in the data, ranked on the same negative rate.Most Miserable
- Most profane, swears the most
- The explicit-language board: explicit comments per 1,000.The most explicit fanbases in college football · Watch Your Profanity
- Sorest losers, most bitter
- Two season superlatives, both read off how a fanbase talks in the weeks it lost.Sorest Losers · Most Bitter
- Nastiest, meanest
- The combined harm and identity-based abuse board. The two tag groups are summed, so it is a rate for comparing fanbases and never a count of comments, and what it measures is what the classifier flagged, not intent: nobody has to mean it for a comment to count.The most toxic fanbases in college football
Most annoying, most hated and most obnoxious are judgments other fanbases make about a fanbase. This data reads only what a fanbase says on its own page, so the Sence Fan Index does not rank them.
These rankings come from counting classified comments across a full season, not from a poll or an editor’s list.
Questions we get asked
What does "per 1,000 comments" mean?
A rate is comments matching that tag per 1,000 total comments for that team (the Sence Score convention), so a high-volume fanbase cannot top a board on volume alone. Rates are rounded to one decimal. The toxicity figure sums two tag groups, so it is a rate for comparing fanbases and never a count of comments.
What is the Fan Score?
Positive comments per 1,000 total comments, as classified by Sence. Higher means a happier comment section.
Why do subtag counts add up to more than the group total?
A tag-group total counts distinct comments that carried any tag in the group. Subtag figures count tag applications, and one comment can carry several tags, so subtags can sum past the group total.
Why are some teams not ranked on a metric?
Teams below the volume floors (1,500 comments for sentiment metrics; 2,000 comments and 25 group tags for tag-group metrics; 2,000 comments and 15 subtag applications for subtag metrics) are excluded from that ranking rather than ranked last, because their rates are too noisy to rank a school on.
Using these numbers
Quote us. Figures from this site may be quoted and cited in reporting, research, and commentary, and the share cards may be reproduced, on two conditions: credit the figure to “Sence Fan Index, by Sence,” and link to the page on this site the figure or card came from. For a team’s own number that is the team page; for a ranking it is that board.
That permission covers quotation and citation. It is not a license to republish the underlying aggregates in bulk, and there is no bulk download of them: the pages, the boards, and the share cards are the published form of this data. Nothing here stops us from offering a formal license later; this is what is offered today.
Two things we ask when you quote a rate. Keep the per 1,000 wording, because a rate restated as a raw count of comments is a different and wrong claim. And the per-axis identity-based abuse boards are deliberately unranked: they report each fanbase’s rate without naming a leader, and we ask that reuse does the same.
The press page has the credit line, a top ten off the harm and abuse board, and an index of every share card on the site.
Publisher, dates, and corrections
The Sence Fan Index is published by Sence, which built the classification models behind every figure on it.
The data covers the 2025–26 season, August 23, 2025 to January 19, 2026.
If a number here looks wrong, tell us and we will check it against the export and correct the page, or say plainly why it stands. Contact Sence.
What is Sence
Sence reads the comment sections no team of humans could keep up with: every comment, classified for sentiment, storylines, and abuse in real time, so the people running a page know exactly what their community is saying and can act on the worst of it before anyone else sees it. It built every number on this site, and it works the same way on any community. Including yours.