ProdigyChain · Methodology
How ProdigyChain Works
How ProdigyChain ranks 170,000+ hockey prospects: the algorithm, the F17 international factor, the F19 draft factor, and the Selection Bias Index.
The short version
ProdigyChain produces a single weekly rank for every tracked hockey prospect by combining 47 underlying factor scores. The two most-discussed factors are F17, which normalises international scoring across the IIHF age-class system, and F19, which weights draft-eligible production against level, role, and team context. The Selection Bias Index (SBI) corrects for the ice-time and deployment confounders that distort raw point production at the U18/U20 levels.
Every public ProdigyChain surface — player profiles, ranking pages, team rosters, draft recaps, weekly movers — is rendered from the same algorithmic source. Rankings refresh once a week as new game data lands; the resulting rank movement is published on /movers.
Per-factor explainers
Each subpage walks through the underlying formula, the data sources, and the edge cases the factor is designed to handle.
F17 — International scoring
How prospects in IIHF, Russian, Swedish, Finnish, Czech, Swiss, and Slovak leagues get a level-adjusted score that's directly comparable to North-American CHL/USHL output.
Read the F17 explainer →F19 — Draft-eligible production
How draft-year scoring is weighted against age, level, role, and team strength so a USHL Phase I pick and a CHL star can be compared on the same axis.
Read the F19 explainer →SBI — Selection Bias Index
Why pure point production misranks young prospects, and how SBI corrects for ice-time, deployment, and team-context confounders.
Read the SBI explainer →
Data integrity
Source data is fetched from public sources (IIHF, Elite Prospects, league box scores) plus tournament event feeds (USHL Phase I/II, IIHF U18 World Championship). Aggregation runs nightly; rank publication runs every Monday at 00:10 UTC. The cron schedule lives in vercel.json.
We don't hand-edit ranks. If an algorithmic output looks wrong, the fix is in the underlying factor weights or in upstream data normalisation — never the published rank itself.
Want every factor behind a specific prospect?
The 47-factor breakdown, weekly rank history, comp finder, and projected draft fit live on theprodigychain.com.