GWTTKB consensus values are derived from real market activity across 100,000+ Sleeper dynasty leagues. Two primary data sources:
We aggregate startup draft results from thousands of Sleeper leagues each season. Startup drafts are the purest market signal — managers are building from scratch with real stakes. A player's startup ADP reflects what the market actually pays, not what people say in polls.
Every completed trade on Sleeper is a data point. We process thousands of trades daily through a market-calibrated engine that adjusts player values based on what they're actually being traded for. If managers consistently trade Player A for packages worth X, that's Player A's real market value.
We built proprietary engines that analyze six years of startup ADP data, completed trade volume, and NFL production statistics to reconstruct daily player values from 2020 to present. No external consensus sources were used — every historical value was computed from scratch using our own formulas applied to real market activity and statistical production over time.
The result: 500+ players with daily value snapshots going back to their rookie seasons. This is the foundation for trajectory analysis, breakout detection, and historical comp matching. When the Coach says "players who posted these stats historically gained 18% within 60 days," that insight comes from our reconstructed value history mapped against real NFL production data.
Raw consensus values treat every point equally. But in dynasty, a 9,000-value player is worth MORE than three 3,000-value players combined. We apply a retention curve that models the real-world consolidation premium:
This means a single elite asset (9,000 × 44% = 3,960 adjusted) is worth more than two mid-tier assets (3,000 × 22% = 660 each = 1,320 adjusted combined). The formula is calibrated against real trade outcomes on KTC and Sleeper.
League power rankings apply the adjusted formula to every asset on each team's roster (players + picks). Teams with concentrated elite talent score higher than teams with spread-out depth — matching how real championship rosters are built.
Power scores are normalized to 0–99 within each league, where the highest-scoring team receives 99.
Each team is classified into one of seven tiers based on optimal starter value, starter-weighted age, bench depth ratio, and pick capital:
The analytics engine maps NFL production data against dynasty value movement to identify breakout and decline signals. We process receiving metrics (target share, air yards share, WOPR, separation, drop rate), rushing metrics (carry share, yards after contact, broken tackles, stuff rate), and passing metrics (CPOE, EPA, pressure rate, time to throw) — cross-referenced against six years of value outcomes.
Player statistics, league data, rosters, transactions, and draft results are sourced from the Sleeper API and publicly available NFL data repositories. All data processing and value calculations are proprietary to GWTTKB.