The Methodology

The thesis

Consensus enters, divergence leaves. Where the whole market agrees, the price is fair and there's nothing to win. Where it argues — a player one room drafts two rounds before another — that's where an edge hides, for whoever reads the split right. Every board here is some angle on that one idea: where the crowd agrees, where it argues, and where it's moving.

Where the numbers come from

Nothing on these boards is scraped from another ranking site, paywalled feed, or anyone's copyrighted work. If a source ever goes live-by-license, the Data Status page shows it flip — live vs sample is always badged on the board itself, never hidden.

What the colors mean

One language everywhere: every matchup is graded by how many fantasy points a defense has actually allowed to a position, ranked 1–32. GREEN = ranks 23–32, a defense that's been giving it up — soft water. AMBER = the middle — neutral. RED = ranks 1–10, a wall — rough water. Schedule scores (Outlook, Season SoS, Trade) are just the average of those ranks over a window, centered on zero: positive = soft slate, negative = tough. It's a sorting handle, not a rating out of ten.

Facts vs labeled proxies

Most boards are raw facts (a schedule, a price, points allowed). A few use a stand-in where the direct number isn't in the open data — and we label every one:

The Composite score

The Composite is one number built from four real ingredients, each already graded on its own board: Produced (Value Board VOR, 40%), On the field (open snap-count share, 25%), Startable weeks (Consistency math, 20%) and Season ahead (Season SoS slate rank, 15%). Each ingredient is the player's percentile against others at his own position — the score just blends those percentiles at the weights above. It adds nothing of its own.

The substance gate: a score is built only from real, shown inputs. Missing an ingredient shrinks the recipe and tags the row partial. Missing too much — this year's rookies, for example — means no score at all. We never invent a vibes number.

Self-calibrating signals

Draft spreads naturally widen for later picks, so absolute cutoffs would mislabel everyone. The consensus/split signal and the divergence index both grade a player against his own draft-slot peers — a player reads "split" because the market argues about him more than it argues about players like him, not because he's drafted late. Tiers work the same way: a tier breaks where the price gap is big for that position.

What we refuse to do

• No advice. Buy-low/sell-high leans, matchup colors, tier lines — all of it is a read on the water, never a start/sit or trade instruction. You decide.
• No invented movement. The Movers board only ever diffs real stored days. When it doesn't have history, it says "collecting" instead of faking a delta.
• No scraping on these boards. Open data and app-permitted APIs only. When the clean version of a feed costs money, we show a labeled sample until it's licensed — we don't take the scraped shortcut.
• No hidden staleness. Every board carries its refresh date and its live/sample badge.
• Plain words. Every page writes soft/tough, not calm waters/rough seas/easy water — one house style, no cute metaphors to decode.