Methodology
How our rankings are calculated.
Every value on the board starts from three independent sources. The pipeline puts them on a common scale and blends them into one number. Walk the eight stages below in order, or jump straight to the one you want.
Stage 01 of 08
Where the numbers come from
Three independent sources, weighted equally
Each source fails differently: the trade market reflects real payments but lags news, the crowd market moves fast but overreacts, and expert ranks are careful but ordinal, saying who ranks above whom, never by how much. Blending all three keeps any single failure mode from taking over.
The trade market
The trade market’s values come from completed trades in real leagues: actual market behavior, not opinion. The pipeline pulls separate 1QB and Superflex boards at half PPR, matching the crowd market’s format, keyed on each player’s native id.
The crowd market
The crowd market’s values come from head-to-head comparison voting, covering both formats. Unlike the trade market, it also prices draft picks, and every future pick value starts here. It’s fetched at most once a day and cached to disk.
Expert consensus ranks
Expert consensus is a rank, not a value, so it can’t join the average until stage 02 lifts it into value space. The feed covers 500+ ranked players per format from about two dozen experts, refreshed daily. Every pull checks depth first: a restricted key returns only the top 10, which would skew the blend toward elite talent, so a truncated response falls back to a public dataset instead, keeping the blend at three sources.
Why the weights are equal
All three sources carry weight 1.0, a deliberate starting point, not a claim they’re equally accurate: weights get tuned only once enough trade history earns the change.
WEIGHTS = { trade market: 1.0, crowd market: 1.0, expert consensus: 1.0 }