The Public Hockey Analytics Map

By Jesse Ambrock

June 6, 2026 · 728 words

The major independent hockey analytics models do not sit inside one interface. Money and proprietary ownership keep the leading work in separate products, which means anyone pulling public data for contract context, prospect translation, or trade evaluation ends up maintaining a short roster of destinations rather than toggling inside a unified platform that has never materialized.

The functional directory therefore consists of the four primary hubs that compile data at different layers plus the four creators whose specific outputs — cards, projections, spatial maps, and manually tracked events — account for most of what circulates on timelines and inside front-office reference materials, along with several other established resources that fill important gaps for prospect evaluation, historical context, and deeper research.

Four Hubs Carry the Compiled Data

HockeyStats.com aggregates microstats, player cards, WAR, and playoff odds inside a clean visual interface and represents the closest current approximation of a one-stop public structure.

MoneyPuck.com functions as the primary destination for game-level models. Team playoff odds, the Deserve To Win O’Meter, expected goals, and the performance of specific line combinations all live there in a compiled form.

Natural Stat Trick supplies the gold standard for raw underlying metrics. It does not emphasize flashy player cards, but it assembles the on-ice data — Corsi, high-danger chances, and the rest — that feeds many of the more visible public models.

Evolving-Hockey.com, built by the Evolving Wild twins, delivers some of the most respected public WAR models along with its own player cards. Deeper features on the site sit behind a paywall.

The Four Names That Power the Visible Work

JFresh produces the brightly colored, percentage-based player cards that appear constantly on X. The cards operate as a visual aggregator: they pull WAR data from Evolving-Hockey and manually tracked microstats from AllThreeZones, then convert the combination into percentile displays.

Dom Luszczyszyn at The Athletic supplies clean tabular team projections and player value cards that frequently attach dollar figures to a player’s worth. His GSVA model drives the output and orients toward game-to-game predictions, standings points, and contract valuation.

Micah Blake McCurdy at HockeyViz creates the red-and-blue heat maps that resemble weather radar over a rink. The models isolate the precise ice locations where a player generates offense or suppresses defense while adjusting for the effects of teammates and competition.

Corey Sznajder at AllThreeZones performs the manual tape tracking that generates the microstats the NHL itself does not publish — successful puck carries into the offensive zone, cross-ice pass completions, and similar event-level detail. Multiple models, including the JFresh cards, depend on his tracked data as a primary input. He remains the unsung foundation for much of the public microstat layer.

Other Reputable Resources

Beyond the core hubs and names above, these established accounts and sites provide high-signal data, visualizations, or analysis that front offices and serious observers routinely consult.

Hockey Reference
The baseline public archive for box scores, historical stats, and career leaderboards. Many advanced models calibrate against its data, and it remains the quickest way to ground any claim in verifiable NHL history.

Elite Prospects
The go-to database for prospect profiles, draft history, league-by-league data, and scouting notes. Essential for any work on draft capital ROI or how production translates across developmental environments.

Patrick Bacon (TopDownHockey)
Develops independent NHL Equivalency and prospect projection models. His public Tableau tools and contributions to compiled sites like HockeyStats supply structured ways to compare prospects across leagues and eras.

CJ Turtoro
Builds clear Tableau visualizations of Corey Sznajder’s AllThreeZones microstat tracking. The dashboards make zone entry, exit, transition, and retrieval data accessible without needing direct access to the raw project.

Hockey-Graphs
A platform for original analytical research and long-form testing of public metrics in specific contexts. It is where many of the deeper studies that inform model-building first appear.

Ownership Keeps the Pieces Separate

Each creator built a distinct mathematical approach and retains control of the formulas. They monetize independently through Patreon, Substacks, or media partnerships such as The Athletic. Assembling every model under a single site would require licensing deals across all of them, an arrangement that has not occurred.

The result is a landscape where the highest-signal public work stays distributed. For anyone who needs the actual sources rather than a secondary summary layer, the durable practice is a personal folder of direct bookmarks to the four hubs and the four primary independent voices.