How Huffman Sports Wire builds its boards
A practical guide to our sources, ranking signals, draft models and checks.
Version 1.0 · Updated
Download full methodology PDF · Download metrics glossary PDF
Our boards organize evidence; they do not reveal what a team will do. Reported statistics, independent evaluations and draft scenarios answer different questions. This guide explains how those inputs become rankings, where judgment enters, and what unavailable data means. The metric panels show the calculations behind the labels.
1. Data sources and coverage
SportsDataIO is our primary configured live-data provider. Access differs by league and endpoint, so a working score feed does not guarantee complete roster, transaction or player-stat coverage. CollegeBasketballData supports college basketball datasets where available. Public NCAA statistics also support designated college-production features; their scope is identified separately from scouting rankings.
For NFL prospects, SportsDataIO college season statistics lead. CollegeFootballData may fill missing fields or demonstrably older cumulative production, with reviewed snapshots supplying remaining gaps. Players are matched by name and school, not name alone. Stable team catalogs can support navigation during outages, but they cannot supply invented live records or rosters.
Authored ranking snapshots and published reporting are separate inputs. We use ranking signals rather than reproducing scouting prose. Public availability alone does not establish redistribution rights, and a stored list is not proof of a complete provenance audit. Manual News Break reporting remains original, attributed editorial work, distinct from automated performance summaries.
2. Ranking methodology
The NFL starting order is an authored consensus snapshot combining four public ranking views. Different list lengths create coverage differences, so a player omitted from a short list is not automatically a poor prospect. That precomputed order supplies source rank; the published model converts rank into baseline points before considering production.
Production is compared within position groups using weighted statistical factors and percentile placement. Required factors must be complete. An unrated player receives no production adjustment, not a zero-performance penalty. Offensive-line production uses team context rather than claiming individual blocking measurement. Bounded editorial adjustments are another signal, not replacement statistics or an official scouting grade.
Final points determine the NFL order, with source rank resolving equal totals. This makes the arithmetic inspectable without suggesting that tiny differences predict professional success. Rankings describe the available evidence and model choices; they are not injury diagnoses, betting probabilities or guarantees.
3. Constructing boards and draft scenarios
A prospect board orders players independently of a club's immediate needs. A mock draft adds pick ownership and roster context. NFL needs use active depth-chart coverage and available player ratings to compare position groups and prioritize five needs. Our depth order and need labels are independent evaluations, not official coaching plans.
Pick grades combine board value with an ordinal need bonus on a 70-point scale. Class grading starts with the average pick score, then applies the displayed class-scale conversion. A public weekly mock is a dated snapshot, separate from a fan's private simulator. Changing a scenario does not rewrite a published edition.
Other sports use different constructions. The NBA mock is an authored prospect order, not an invented statistical formula. CBB's player board starts with NBA mock order, then alternates committed 2026 recruits returned by CollegeBasketballData and qualified 2025–26 statistical leaders; enrichment statistics do not independently reorder that board. MLB combines three public rank lists, with limited MLB-only performance context for verified matches. Neither board claims complete minor-league or recruiting coverage.
Team polls are separate from player boards. The CBB editorial Top 25 uses fixed owner ballots and verified final-result points. It is not the separate preseason projection model. Always read the panel's model name, season and explanation before comparing its points with another feature.
4. Validation, freshness and corrections
Validation begins with identity and scope: confirm the player and school or club, distinguish provider IDs, check season labels, and apply sample thresholds. NFL production requires complete factors; MLB excludes ambiguous name-only matches. Missing values remain unavailable. They must not become fabricated salaries, zero statistics or unsupported claims of full coverage.
Refreshes check source availability and coverage. NFL production can retain a prior good snapshot when a newer response has worse coverage. That protects continuity but does not make old data live. Source notes, dates and partial or provisional labels matter, especially while final-game statistics lag scores or schedule feeds fail.
Automated tests exercise ranking, need, grade and provider-boundary behavior. They do not constitute an independent audit, establish publication rights or eliminate editorial mistakes. Corrections should identify the affected page, player or team, season and supporting source. Contact huffmansportswire@gmail.com; distinguish a factual error from disagreement with an independent evaluation.
Metrics explained
NFL board points and production adjustment
Baseline = max(5, 1000 − 5 × (source rank − 1)); production adjustment = rating − 50.
An unrated prospect has a neutral adjustment. Final points add baseline, available production adjustment and bounded editorial adjustment, with a floor of zero. Points are an ordering device, not a percentage chance of being drafted.
Position-aware production percentiles
Compare complete weighted production factors within a position group.
A percentile describes relative placement, not certainty. Different position groups use different factors; team offensive-line context is not an individual blocking grade. Missing required factors mean unavailable, even when some counting statistics exist.
NFL pick value, need and class grades
Delta = pick number − board rank; value = clamp(45 + 0.75 × delta, 0, 60).
Need ranks 1–5 add 10, 7.5, 5, 2.5 and 0 points; unranked positions add 0. Clamp the pick score to 0–70. A positive delta means a player remained available beyond board rank. For classes, round 50 + 3 × (average pick score − 47.5) to one decimal place, then clamp to 0–70. A class score is not the raw average.
CBB statistical-leader selection
Selection score = 2 × PPG + 1.2 × RPG + 1.5 × APG + 3 × SPG + 3 × BPG.
The leader pool requires at least ten games. This score selects statistical candidates for the alternating portion of the player board; it does not replace the leading NBA mock order or describe the separate team Top 25.
MLB consensus and eligible production
Consensus = average source rank across three lists; an omission receives rank 125.
Ties consider source count and best rank. Qualified MLB hitters need 100 plate appearances, or 75 at-bats when plate appearances are unavailable; pitchers need 20 innings. The separate production index emphasizes OPS for hitters and ERA/WHIP for pitchers, reweighting available components. It is not comprehensive MiLB measurement.
Data source references
- SportsDataIO: Primary configured live provider; coverage varies by league and endpoint.
- CollegeFootballData: Narrow NFL prospect-production supplement for missing or older cumulative fields.
- CollegeBasketballData: College basketball data; board season and sample eligibility remain explicit.
- NCAA statistics: Public college-production context, not an independent scouting rank.
Explore the prospect board · Read the weekly mock draft · Sourcing and corrections policy