LatestArticlesMLBNFLNBANHLFantasyProjections

NHL Coverage ​

Expected goals, possession metrics, and why plus-minus is lying to you.

The Analytics Revolution in Hockey ​

Hockey analytics has come a long way from the early days of Corsi and Fenwick. Modern expected goals models incorporate shot location, shot type, traffic, rebound status, and pre-shot movement to estimate the probability of any shot becoming a goal.

Key Metrics ​

Team-Level ​

  • xGF% — Expected Goals For percentage. Measures quality-adjusted shot generation vs. prevention.
  • CF% — Corsi For percentage. Raw shot attempt differential (a possession proxy).
  • PDO — Shooting % + Save %. Teams far from 100 are likely to regress.

Player-Level ​

  • GAR — Goals Above Replacement. Total value metric.
  • xG Impact — How much a player improves/hurts expected goals when on ice.
  • WAR — Wins Above Replacement, adapted for hockey.

Why Traditional Plus-Minus Is Broken ​

Traditional plus-minus in hockey is one of the worst widely-cited stats in all of sports. It's polluted by:

  1. Teammates — Playing with great players inflates your plus-minus
  2. Zone starts — Players deployed in the offensive zone look better
  3. Goaltending — Your goalie's save percentage directly affects your numbers
  4. Score effects — Trailing teams generate more shots, distorting metrics

Modern adjusted plus-minus models (RAPM) attempt to isolate individual impact, but even these require large samples to stabilize.

The Skeptic's Take: If someone cites raw plus-minus to evaluate a hockey player, politely direct them to this page.