A rank is an average of ranks. We do not model players; we measure what four independent sources think and publish the consensus, with each source’s own number beside it.
Four sources, equal weight. Two are markets, one is a solved price, one is a panel. They disagree often enough to be worth averaging and agree often enough to be worth trusting. Every row on the rankings page shows all four, so when the board says a player is #12 you can see that one site had him #6 and another #19, and decide for yourself.
Ranks, not values. The sites publish values on incompatible scales — one tops out at 9,999, another has no ceiling, one has no values at all. Averaging those would mean choosing a scale and pretending the others meant the same thing. Averaging ranks needs no conversion, and a rank is the thing a fantasy manager actually argues about.
KeepTradeCut — a crowd-sourced market: thousands of “keep, trade, or cut” votes a day, published as values for 1QB, superflex and TE-premium, dynasty and redraft.
FantasyCalc — values solved from real trades across public leagues, published per format. Its TE-premium parameter currently returns the same board as its base feed.
Dynasty Daddy — a blended market value for 1QB and superflex, dynasty and redraft. No TE-premium feed.
Flock Fantasy — the average of eight analysts’ hand-made ranks for superflex, 1QB and both redraft shapes. No TE-premium feed.
Seven steps, every hour, for each of the eight boards.
Every hour we fetch all four sites. A feed whose payload has not changed is skipped, so a quiet hour costs nothing and moves nothing.
Each site’s list becomes a dense overall order and a position order for each of the eight boards. Names are matched to one player record; an id that disagrees with the name and position is not trusted.
A player’s board rank comes from the mean of the sites’ overall ranks. A site that does not rank a player simply does not vote; nothing is imputed.
Sort by that average. Ties break on the average position rank, then on how many sites voted. Overall rank is the row number.
Position rank is your slot within your position on that sorted list — derived, never averaged — so WR12 can never sit above WR11.
Tiers are natural breaks in the board’s hidden value curve, held steady between refreshes so a one-place wobble does not relabel a player.
If anything moved, the board version ticks and the rankings page you have open swaps rows in place within a minute.
Four players, four sites. Average the overall ranks, sort, number the rows. That is the whole algorithm.
| Board rank | Player | KTC | FantasyCalc | DynastyDaddy | Flock | Average |
|---|---|---|---|---|---|---|
| #1 | Player A | 1 | 3 | 2 | 2 | 2 |
| #2 | Player B | 3 | 1 | 1 | 3 | 2 |
| #3 | Player C | 2 | 2 | 3 | 1 | 2 |
| #4 | Player D | 4 | 5 | — | — | 4.5 |
Player D is ranked by two sites and gets a rank from those two. A row with a single source is flagged thin on the board and is dropped altogether when that one site has him past #250 — a lone deep-bench opinion is noise, not a ranking.
Your league maps to the nearest of these from its roster and scoring settings. PPR versus half and 4- versus 6-point passing touchdowns do not change the order: none of the four sites publish rankings that vary on them.
| Board | Players | Picks | Thin | Version | Re-blended |
|---|---|---|---|---|---|
| Dynasty · 1QB | 410 | 31 | 4 | 195 | 34 min ago |
| Dynasty · 1QB · TE Premium | 410 | 30 | 3 | 194 | 34 min ago |
| Dynasty · Superflex | 414 | 31 | 7 | 194 | 34 min ago |
| Dynasty · Superflex · TE Premium | 414 | 30 | 6 | 195 | 34 min ago |
| Redraft · 1QB | 309 | 0 | 22 | 193 | 34 min ago |
| Redraft · 1QB · TE Premium | 311 | 0 | 24 | 195 | 34 min ago |
| Redraft · Superflex | 313 | 0 | 32 | 175 | 34 min ago |
| Redraft · Superflex · TE Premium | 313 | 0 | 32 | 175 | 34 min ago |
TE premium is priced by KeepTradeCut alone today: FantasyCalc’s premium feed returns its base board and the other two have none. On the premium boards those sites contribute their base rank, so the tight-end bump you see is one market’s opinion, not four.
Draft picks live on the dynasty boards as their own rows — “2027 Early 1st” — blended from the three sites that price picks, for the next two draft classes. Redraft boards carry none.
Three things on a row are not averages, and each is explained here rather than assumed.
Tiers are breaks in a curve, not fixed bands. Behind every board sits a hidden value curve fitted from the sites’ own values; tiers are placed where that curve genuinely steps, with the first tier capped small and no middle tier allowed to swallow the board. Boundaries hold steady between refreshes so a one-place wobble does not relabel a player.
Movement is real, not decorative: the board is snapshotted every hour and the 24-hour and 7-day arrows compare a player’s rank to the snapshot nearest that far back. We currently hold 200 hourly snapshots, the oldest from 8 d ago.
The hidden value. Adding assets together — a two-for-one, a roster total, a waiver score — needs a number, and ranks cannot be added. So a value on a 0–100 curve exists on the server for exactly those jobs and is never sent to a page. The trade calculator returns a verdict and a percentage gap; it will not show you a total to argue with, and neither will anything else here. A consolidation premium is applied when one side gives up the single best asset, because that is how real leagues price two-for-ones.
The honest list.
A crowd market lags on breaking news; a solved price needs trades to move; a panel is eight people. Averaging softens each one’s error but cannot see what none of them saw. The per-site ranks on every row are there so you can tell which kind of disagreement you are looking at.
Two of the four feeds identify players only by name. We match on name, team and position, refuse an id that disagrees with the name, and drop inactive namesakes — but a rookie who shares a name with a veteran is the failure mode, and we audit for it rather than promise it away.
A player ranked by one site is shown with a thin marker and a single vote. He is on the board because hiding him would hide the fact that a market prices him; treat the rank accordingly.
Eight boards cover what the sources publish. A league with unusual scoring — heavy TE premium, two QB slots plus superflex, points per first down — maps to the nearest board and the order will not move for the difference.
Rankings shown here are derived from ranks published by KeepTradeCut, FantasyCalc, Dynasty Daddy and Flock Fantasy, fetched from their public feeds. Their values are used only to fit the hidden curve and are not reproduced. Each source is credited on every row that uses it.