How These Values Are Made
Values: DynastyProcess (from FantasyPros ECR) · updated 2026-09-04
Source
Every number comes from values.csv in the
DynastyProcess repository (GPL-3.0).
DynastyProcess maps FantasyPros dynasty expert consensus rankings onto an exponential
decay curve. We present that output: no curve of our own, no edits to the
value_1qb or value_2qb columns. Redistribution keeps the
same license and attribution.
Update cadence
Upstream publishes on Fridays: weekly, not daily. We poll daily and archive every version (340 snapshots since 2019-02-13). Upstream also goes quiet. Since 2023: 6 gaps over 14 days, the longest 62 days (2026-02-27 to 2026-04-30). During a gap the site shows the real date of the last version, never today's.
| Gap start | Gap end | Days |
|---|---|---|
| 2026-02-27 | 2026-04-30 | 62 |
| 2023-04-21 | 2023-05-26 | 35 |
| 2024-02-23 | 2024-03-27 | 33 |
| 2025-06-06 | 2025-06-27 | 21 |
| 2023-01-27 | 2023-02-15 | 19 |
| 2023-02-24 | 2023-03-12 | 16 |
Eight data gates
A failed blocking gate means no publish; the previous version stays live. Gates 1, 7 and 8 apply only across a normal 6–9 day gap.
| # | Gate | Rule | Fail |
|---|---|---|---|
| 1 | Row count | Within ±60% of previous version | block |
| 2 | Freshness | scrape_date ≥ previous version, ≤ 10 days old | block |
| 3 | Completeness | player, pos, value_1qb, value_2qb non-empty | block |
| 4 | Range | Values 0–12,000; QB, RB, WR, TE each ≥ 40 rows | block |
| 5 | Continuity | Top 10 overlap with previous ≥ 6 | warn |
| 6 | Control | Six known-bad fixtures each trip their expected gate | block |
| 7 | Liveness | ≥ 100 players changed since previous version | block |
| 8 | Picks | Median pick change ≤ 15% outside Jan–Mar re-anchor window | warn |
Agreement with FantasyCalc
Measured 2026-09-05 against FantasyCalc's 1QB dynasty list, read once for this check, never in the build. Spearman ρ = 0.962 (n = 356 matched players).
| Cohort | Shared players |
|---|---|
| Top 10 | 9/10 |
| Top 25 | 19/25 |
| Top 50 | 44/50 |
| Top 100 | 92/100 |
| Top 300 | 289/300 |
Disagreement sits at ranks #96–#353, where nobody has consensus. If your league treats KeepTradeCut or FantasyCalc as the currency, use them. This site is a second opinion with a public method.
Limits
No league sync: the Sleeper API is non-commercial only. Only 1QB and Superflex exist as separate datasets; TE premium and league size have no independent data, so they are not toggles. Ranks past #96 have no consensus anywhere.
Not used
KeepTradeCut is not scraped (its terms forbid it). The FantasyCalc API is not called by the site (non-commercial, no substantive substitute). The Sleeper API is not used.