Data & methods

Receipts,
with boundaries.

This site is generated only from the reviewed public-data layer. Authentication material, account identifiers, private messages, payment data, and raw source pages are outside the build boundary.

Coverage

What the archive can prove

Complete means all nine finished seasons. Partial means exactly that—never implied completeness.

complete

Standings

Regular and final standings for all nine completed seasons.

complete

Drafts

Every pick from 2017 through 2025.

complete

Schedules

Weekly scores and matchup results for completed seasons.

complete

Playoffs

Championship and consolation bracket evidence.

complete

Weekly rosters

1,490 complete team-week snapshots with 26,029 lineup-slot rows. 78 empty slots are counted in coverage but are not converted into player histories.

complete

Transactions

Normalized transaction analytics cover every completed season.

Method

How to read the archive

Identity model

Members are not team names

Member careers, season entries, and historical team names are kept separate. Two teams can share a display name without becoming the same identity.

Ownership model

Complete snapshots, evidence-only players

Complete team-week snapshots; empty lineup slots are counted but are not player histories.

Privacy model

Public-safe IDs only

The generated pages use synthetic public member and player routes. Source account and platform identifiers never enter the site bundle.

Snapshot

2026-07-13T21:25:16-04:00

Dataset 2026-07-13-v1 · Public privacy scan: automated-source-id-scan-passed.

Analytics methods

How the lab reaches a verdict

Every leaderboard keeps its population, denominator, scope, and hindsight limits visible.

Schedule luck

Actual regular-season wins minus expected wins from weekly all-play results against every other team. Positive means the real schedule produced more wins than the scoring profile.

All-play win rate

For each regular-season week, a team receives one win per league score it beats and half a win per tie, divided by all available weekly opponents.

Lineup capture rate

Regular-season starter points divided by the hindsight-optimal legal lineup from that recorded active roster, using the exact 2QB and season-specific flex rules. Reserve-list players are excluded.

Lineup regret

Hindsight-optimal points minus actual starter points. A historical roster snapshot cannot prove every alternative was switchable before kickoff, so this measures captured scoring ceiling—not decision quality or causation.

Scoring volatility

Population standard deviation of regular-season weekly points. Lower values mean steadier scores; higher values mean wider week-to-week swings.

Ownership evidence

Team-weeks, starts, points, and continuous spells come from normalized weekly rosters. Empty lineup slots are coverage evidence, never invented player records.

Transaction classes

Lineup log entries are separated from roster-market action groups. Adds, drops, trades, commissioner actions, and veto activity retain their observed transaction-log meanings without double-counting multi-leg moves.

Ranking thresholds

Career lineup and volatility awards require at least 50 evaluated team-weeks. Shorter careers remain visible but are not forced into misleading qualified ranks.