TL;DR: xG measures shot quality (0–1 probability of scoring), Corsi measures possession via shot attempt share, GSAX measures goaltender value above average, and PDO measures luck. Together they give a more accurate picture of team and player performance than goals and plus/minus alone.
Hockey analytics has moved from the fringes to the front office. Every NHL team now employs analysts, every broadcast mentions Corsi, and terms like “expected goals” have entered mainstream hockey conversation. But what do these metrics actually measure — and more importantly, what do they tell us that traditional stats miss?
This guide covers the essential NHL analytics metrics, explains what each one measures, and shows how teams use them to make decisions.
Goals, assists, and plus/minus have dominated hockey discourse for decades. The problem: they’re heavily influenced by luck, teammate quality, and small sample sizes.
A player can score 30 goals in a season while playing on a dominant line, receiving power play time, and benefiting from a hot goalie. Strip those factors away and the underlying performance might be mediocre. Analytics metrics attempt to isolate what a player actually contributes — independent of luck and context.
What it measures: The probability that a shot attempt results in a goal, based on historical data.
Expected goals assign a value to every shot based on factors like:
A shot from the slot on a rebound after a cross-ice pass might have an xG of 0.25 — meaning historically, 25% of shots in that situation result in goals. A shot from the blue line with no traffic might be 0.03.
Why it matters: Over a full season, players and teams whose actual goals exceed their expected goals have typically been “lucky” — they’ve outscored their process. Players who underperform their xG have been unlucky or are playing in front of a poor goalie. xG regresses toward reality over time, making it a better predictor of future performance than raw goals.
Key metrics:
What they measure: Shot attempt volume.
Both are expressed as percentages:
Why they matter: Puck possession drives shot attempts. Teams that control the puck more generate more Corsi events. Over large samples, Corsi is a strong predictor of future winning percentage — more so than actual goals in the short term, because it removes goaltending luck.
Limitations: Not all shots are equal. A Corsi event from centre ice carries the same weight as a tap-in from the crease. This is why expected goals, which weight shot quality, have largely superseded raw Corsi as the primary possession metric.
What it measures: Goaltender performance relative to shot quality faced.
GSAX compares how many goals a goalie actually allowed versus how many they were expected to allow based on the shots they faced (using xG values). A GSAX of +10 means the goalie saved 10 more goals than an average goalie facing the same shots would have.
Why it matters: Save percentage treats a blocker-side shot from the blue line the same as a backdoor tap-in. GSAX accounts for shot quality, giving a much more accurate picture of goaltender contribution.
Top GSAX seasons tell us who is genuinely elite versus who has benefited from playing behind a shot-suppressing defence.
What they measure: How teams transition through the neutral zone.
Tracked manually (or increasingly by tracking technology), zone entries record:
Why they matter: Controlled zone entries (carry-ins) lead to significantly higher shot quality and expected goals than dump-ins. Teams and players who consistently carry the puck into the zone — rather than dumping and chasing — generate better offensive opportunities. This is why analytically-focused coaches emphasise neutral zone play.
What they measure: Shot attempts from the most dangerous areas of the ice.
The “high-danger” zone is typically defined as the area directly in front of the net (the royal road area) and includes slot shots and cross-ice passes converted to shots. High-danger chances are a subset of expected goals — they represent the highest-probability scoring opportunities.
What it measures: Shooting percentage + save percentage, combined.
PDO is the sum of a team’s on-ice shooting percentage and save percentage. The league average is always 100 (or 1.000). Teams above 100 are “running hot” — they’re scoring on a high percentage of their shots and their goalie is stopping more than expected. Teams below 100 are “cold.”
Why it matters: PDO regresses toward 100 over time. A team with a PDO of 105 in the first two months is probably due for regression — their actual results will likely worsen even if their underlying play stays the same. PDO is a red flag or green flag for analysts when evaluating whether a team’s record reflects their true quality.
Modern NHL front offices don’t use any single metric in isolation. The typical analytical workflow:
Analytics don’t capture everything. Leadership, compete level, defensive zone awareness, and the ability to perform under playoff pressure remain difficult to quantify. A team of analytically perfect players can still lose in the playoffs.
The most sophisticated teams use analytics as one input among many — combining data with traditional scouting, coaching intuition, and player interviews. The goal isn’t to replace human judgment but to reduce the impact of cognitive biases and small-sample conclusions.
Understanding these metrics gives you a fundamentally different lens on the game. When a team’s xGF% is consistently above 55% but they’re losing games, you’re watching a team that’s playing well but running cold — and likely to bounce back. When a player’s Corsi is elite but their goals are down, the goals will probably return.
The numbers don’t tell the whole story. But ignoring them means missing half of it.
xG (expected goals) assigns a probability to each shot based on historical data. A shot from directly in front of the net on a rebound might carry an xG of 0.25 — meaning 25% of similar shots result in goals. A shot from the blue line with no traffic might be 0.03.
Corsi (CF%) is the percentage of all shot attempts directed at the opposing net while a player or team is on ice. It includes shots on goal, missed shots, and blocked shots. Above 52% is positive; above 55% is elite.
Corsi counts all shot attempts equally — a blue-line shot counts the same as a slot shot. Expected goals weight each shot by its probability of scoring. Most modern analysts prefer xG because it captures both volume and quality.
GSAX (Goals Saved Above Expected) compares a goalie’s actual goals allowed versus expected goals allowed based on shot quality faced. It’s the most accurate measure of goaltender value because it accounts for the difficulty of shots faced.
PDO is shooting percentage plus save percentage, always averaging 100 across the league. It measures luck — teams above 100 are running hot and tend to regress, teams below 100 are cold and tend to improve. It’s the most useful tool for identifying whether a team’s record is sustainable.