What a price means

A player is worth what he gives you over the player you could have had for free. If every team starts two running backs, the thirtieth-best running back is roughly what waivers will hand you for nothing. A back who scores fifteen points a game when the free one scores nine is worth those six points — not fifteen.

That gap is the whole model. Points above the replacement player, converted to dollars, scaled so the most valuable asset in the league sits at $10,000. Everything else on the site — movers, waivers, trades, your roster — is a view on that one number.

Projections are built from the bottom up, not scraped from anyone's rankings. For each team we estimate how many plays it runs, how often it passes, and how many points it scores. Then we split that between the players on it: target share, carry share, red-zone work, snap share. Then we apply each player's own efficiency.

The inputs are public: play-by-play, weekly stats, snap counts, Next Gen Stats, depth charts, injury reports, betting lines, and team ratings. Preseason expectations come from consensus rankings and average draft position.

Before any football is played, the board leans heavily on the market — where players are actually being drafted. A bottom-up model with no games behind it produces confident nonsense, and a board that opens with a shock at every position is not useful.

That anchor fades fast. By eight games in, what a player has actually done this season dominates, and draft position barely matters. This is deliberate: the market is a good prior and a poor conclusion.

A star expected to miss four weeks is still a star. Prices are built on what a player does in the games he plays, then discounted for how many he is likely to miss. An injury softens a price; it does not delete it. Only a season-ending injury takes a player to nothing.

Scarcity is different in every league, so the price is too. In a superflex league quarterbacks are genuinely scarce and their prices rise. Add a point per reception for tight ends and the tight ends reprice. On the Exchange, open My league and the whole board recomputes as you change the settings — because the replacement level moves, not because we apply a fudge factor.

Prices carry a small multiplier from what the crowd is doing — net adds and drops across Sleeper — and from betting lines moving on a team. It is capped at 12% in either direction. Hype is real and it is worth something, but it cannot be allowed to swamp what a player actually does. A player pinned at the cap is itself a signal: that is what overbought looks like.

Sentiment applies only to the live season. It cannot be backfilled, because nobody archives what the crowd was doing in 2019 — so historical price charts are model-only, and the site says so where you see them.

The rating compares our price to what the crowd actually pays, using FantasyCalc's trade-based values. Where we are higher, the market is underpaying and it reads as a buy. The comparison is done within position, because our model prices quarterbacks below the crowd by design in one-quarterback leagues — comparing across positions would turn that design choice into a fake sell signal on every quarterback.

Only players the consensus actually covers, and only above $1,000, get a rating. Below that, a percentage gap is noise: a $300 player at $900 is “200% undervalued” and means nothing next to a $9,000 asset. Unrated is the honest answer, so we say it.

It does not read the news. A coach's Wednesday press conference, a holdout, a trade rumour, a beat writer's tweet — none of it reaches the model until it shows up as snaps, targets, or an injury designation.

It cannot tell a sprain from a torn ACL. Injury reports say “knee”, so season-ending injuries are confirmed by hand rather than guessed at.

It is weakest on rookies and on players whose role just changed, which is exactly where fantasy is decided. A backup quarterback who becomes a starter is priced at nothing until the snaps prove it.

And it reprices once a week, on Tuesday. That is a deliberate choice, not a limitation: a market that moves every minute invites you to trade on noise.

Tuesday morning, after the week's games are final. Every player gets a new price and a stored one-line reason for the move, which is what you see on Movers and next to each row on the board. Nothing on this site updates in real time.