Sources and weights
| Source | What it measures | Weight | Terms |
|---|---|---|---|
| DynastyProcess values | Expert consensus (built from FantasyPros expert rankings), 1QB and superflex | 40% | Open data repository, GPL-3.0 |
| Sleeper dynasty ADP | Where players go in real dynasty drafts on Sleeper | 30% | Free public API, no key |
| Fantasy Football Calculator dynasty ADP | Where players go in dynasty mock drafts | 15% | Free ADP API for personal and commercial use; attribution requested |
| nflverse production | PPR points per game this season (3+ games), adjusted for age | 15% | CC BY 4.0. Contains information from nflverse; we changed it into per-game production ranks. |
Why these weights: experts get the largest share because their rankings already weigh age, situation and talent. Draft markets get 45% combined, because what managers actually pay is what you can trade for. Sleeper gets more weight than mock drafts because those are real leagues. Production is 15% so that current output counts without letting a hot month swamp long-term value.
Putting different sources on one scale
- Each source ranks players within their position (for example WR12).
- That positional rank is turned into a value with the DynastyProcess value curve for the position, on a 0–10,000 scale. Superflex uses the superflex curve, which lifts quarterbacks.
- The composite is the weighted average of the sources that have the player. A player needs at least one expert or market source; production alone is not a dynasty value.
Consensus and buy/sell signals
Each player shows a consensus badge: Sources agree (spread under 35%), Some disagreement (35–65%) or Sources disagree (65% or more). When the two draft markets value a player at least 35% below the experts, we flag a possible buy; 35% above, a possible sell. These are prompts to look closer, not advice.
Start/sit scores
Only one free weekly projection source is available to us (Sleeper, which publishes RotoWire projections), so start/sit blends three independent signals instead:
If a signal is missing, the others are re-weighted. When the signals differ by 6+ points, the tool says so.
When a source fails
The data refresh runs every morning. Each source is fetched separately; if one fails or has too little data (for example, Fantasy Football Calculator’s dynasty ADP can be empty early in a season), it is skipped, the remaining weights are re-normalised, and the page notes it. If every value source failed, yesterday’s values are kept.
Sources we checked and did not use
- KeepTradeCut: its terms forbid scraping and using its full values in other tools.
- FantasyCalc: its API terms do not allow using its values in another trade calculator.
- ESPN and Yahoo projections: no documented public API for this use (Yahoo needs a registered app and account; ESPN’s endpoints are undocumented and its terms limit reuse).