About the Society
Straight answers to weekly fantasy football questions, built from public data with a model we explain in full below.
What we do
The Fantasy Society publishes weekly rankings and projections, start/sit calls, a waiver wire and trade values for NFL quarterbacks, running backs, wide receivers and tight ends. Every page is free to read and needs no account.
Our numbers come from a model, not from gut feel or anonymous experts. The model is deliberately simple so that you can check its reasoning, and every call on the site shows the numbers behind it. We would rather be transparent and occasionally wrong than mysterious.
Where the data comes from
- Statistics, rosters, depth charts, snap counts and schedules from nflverse, an open-source project, licensed CC BY 4.0.
- Injury reports from FTN Data, distributed by nflverse and licensed CC BY-SA 4.0.
We download fresh data once a day, late morning US Eastern time, and rebuild the site from it. The current data was last updated on September 24, 2026. We are not affiliated with or endorsed by the NFL, its teams or nflverse.
How the numbers work
All projections are in half-PPR by default. PPR and standard projections are made the same way from each player's points in that format, using nflverse's standard fantasy scoring. Kickers and defenses use standard scoring, the same in every format.
1. A baseline for every player
We start with each player's fantasy points per game this season, weighting recent games more heavily: each game counts 85% as much as the one after it. Early in the year a couple of games say little, so if the player played at least four games last season, we blend in last season's average as if it were three extra games. As this season goes on, its own games take over.
Quarterbacks who are not first on their team's depth chart are projected close to zero, because backups rarely play.
2. This week's matchup
For every defense we measure the fantasy points it has allowed per game to each position, compared with the league average, counting this season twice as much as last season. We move the baseline by 60% of that difference and never by more than 15% either way, because a few weeks of defensive data are noisy. A defense that has allowed 20% more points than average to receivers lifts a receiver's projection by 12%.
3. Injuries and byes
Using the official injury report for the current week: players ruled out, on injured reserve or on a bye get no projection. A doubtful player keeps 30% of his projection and a questionable player keeps 90%. We do not read news or practice reports beyond those designations.
4. Start or sit calls
We rank each position two ways: by a player's usual level (the baseline) and by this week's projection. A start is a player outside the usual top tier whom a good matchup lifts into starting range with no injury tag. A sit is a usual starter whom a tough matchup or an injury tag drops several spots. Starting range is the top 12 quarterbacks, 24 running backs, 30 receivers and 12 tight ends, roughly a 12-team league's lineups. Head-to-head pages simply pick the higher projection and say so when the gap is under a point.
5. Waiver wire
Free data has no league roster percentages, so we estimate who is available: we leave out players who ranked as fantasy starters by points per game last season and this year's first-round rookies. Among the rest, we rank players by their rest-of-season baseline, their most recent game and a rising share of snaps. Always check your own league.
6. Trade values
A player's trade value is how many points he should score over a freely available replacement for the rest of the regular season. Replacement level is the 13th quarterback, 30th running back, 36th receiver and 13th tight end, roughly the best player on waivers in a 12-team, one-quarterback league. We multiply the per-game gap by the games his team has left (cutting that for players who are out this week or on injured reserve) and scale the result so the most valuable player is 100.
What the model can't see
It doesn't know about coaching changes, trades announced after the last data update, weather, betting lines or a player's role changing mid-week. Early in the season it leans on last year. Treat every number as a strong starting point for your own decision, not the final word.
Corrections and feedback
Data errors usually fix themselves with the next daily update; model problems we'll fix and note here.
The site is for information and entertainment. It is not betting advice; see our terms.