Corsi, Fenwick, xGF%, PDO — the acronyms scare people off, but reading NHL possession stats comes down to one comparison: is a team's share of shot attempts above or below 50%, and does its shooting luck (PDO) look sustainable or due for a crash. A team sitting above 52% in CF% and xGF% is generating the puck more than its opponents most nights; a team under 48% is getting outshot regardless of what the scoreboard says. The catch is that raw numbers get skewed by score effects and zone starts, so the read only holds once you adjust for game state.
- Reading NHL possession stats starts with CF% and xGF% against the 50% break-even line, not the standings.
- PDO near 100 means shooting and save percentage are balanced; above 102 or below 98 signals regression is coming.
- Score-adjusted Corsi beats raw Corsi because trailing teams inflate shot totals chasing the game.
- Zone starts and quality of competition change what a CF% number actually means for a given roster.
- In 2026, xGF% is the analytics community's preferred possession metric over plain Corsi because it weighs shot quality.
Why This Matters
Standings lie in small samples. A team can sit third in its division in November on the back of a hot goalie and a friendly shooting percentage, then collapse by March once the bounces even out. Possession numbers exist to answer a different question than the standings do: is this team actually controlling play, or is it riding variance?
That distinction matters for anyone breaking down a matchup, building a parlay, or trying to figure out which team is built to hold up in a seven-game series. SportsOrca's NHL coverage leans on this framework constantly — a club with a losing record but a 54% xGF% is a different bet than a club with the same record and a 46% xGF%, even though the standings show them as twins.
How to Read NHL Possession Stats: The Core Numbers
Four numbers do almost all the work. Learn these and you can read any possession page on Natural Stat Trick or MoneyPuck without translation.
| Metric | What It Measures | Break-Even | Strong | Weak |
|---|---|---|---|---|
| CF% (Corsi For %) | Share of all shot attempts — on net, missed, blocked | 50% | 52%+ | Below 48% |
| FF% (Fenwick For %) | Share of unblocked shot attempts | 50% | 52%+ | Below 48% |
| xGF% (Expected Goals For %) | Share of expected goals, weighted by shot location and quality | 50% | 53%+ | Below 47% |
| PDO | On-ice shooting % + save % combined | 100 | Over 102 is unsustainable | Under 98 signals positive regression coming |
CF% and FF% tell you volume — who's throwing more attempts at the net. xGF% tells you quality — who's generating attempts from the slot instead of the point. PDO tells you luck — whether the record matches the process or is running hot/cold.
A team can post a strong CF% and still lose games all year if its PDO sits at 97; that's a team getting the process right and the bounces wrong. Flip it, and a team with a 46% CF% riding a 103 PDO is a team that's going to fall off a cliff once shooting luck normalizes. Reading NHL possession stats means checking both columns together, never one alone.
CF% (Corsi): 50% Is the Break-Even Line
Corsi For % is the original possession stat, built from every shot attempt at 5-on-5 — goals, saves, misses, and blocks. Anything at exactly 50% means a team is trading shot attempts evenly with whoever it plays. Push past 52% and a team is out-attempting its opponents on a nightly basis, which correlates with more goals scored over a full 82-game season.
The weakness: Corsi treats a point shot the same as a slot one-timer. That's why the analytics community shifted attention toward xGF% over the last decade — it's still Corsi's cousin, but it weighs where the shot came from.
xGF% (Expected Goals): The Quality-Adjusted Version
Expected Goals For % assigns a scoring probability to every shot attempt based on location, shot type, and situation, then totals up each team's share. A team sitting at 53%+ isn't just taking more shots than its opponents — it's taking better ones. This is the number SportsOrca's NHL analysis leans on most heavily when separating a team that's actually dominant from one that's just shot-happy from the perimeter.
xGF% and CF% usually move together, but when they diverge, the gap is the story. A team with a 51% CF% and a 55% xGF% is manufacturing high-danger chances efficiently rather than firing volume from bad angles.
PDO: 100 Is the Regression Anchor
PDO adds a team's on-ice shooting percentage to its on-ice save percentage. League average nets out to roughly 100 across a full season, so any team parked well above or below that number in 2026 is running on shooting or goaltending luck that tends to correct.
A club posting a 102.5 PDO with a losing possession share is winning games it probably shouldn't be winning — a classic regression candidate. A club at 97.5 PDO with a strong xGF% is getting unlucky and is a buy-low situation the standings don't reflect yet. Goaltending swings PDO hard, which is why checking the NHL's best goalies for a playoff push alongside team PDO explains a lot of the gap between a team's record and its underlying process.
Why Possession Numbers Vary Team to Team
Raw CF%, FF%, and xGF% never tell the full story on their own — context changes what the number means.
- Score effects: teams trailing late pile up shot attempts chasing the game, inflating their raw Corsi; teams protecting a lead sit back and let attempts against climb. Score-adjusted or close-score splits strip this out.
- Zone starts: a fourth line that starts 70% of its shifts in the defensive zone will post a worse CF% than its actual play deserves.
- Quality of competition: a shutdown pair facing top opposing lines every night carries a tougher possession workload than a matchup-sheltered scoring line.
- Goaltending performance: a hot goalie can mask a losing possession battle for months before PDO regresses the record back toward the process.
- System and pace: dump-and-chase teams generate different shot-attempt volume than puck-possession, cycle-heavy systems, even at equal talent levels.
- Sample size: 15-20 games isn't enough to trust a possession number; the analytics community generally wants 25+ games before treating CF% or xGF% as stable.
Is a High Corsi Number Always Good?
A high Corsi number is good, but not automatically decisive — a team can post a 55% CF% while generating mostly low-danger perimeter shots and still lose the expected-goals battle. Cross-check CF% against xGF% before calling a team dominant; if the two disagree, the quality-adjusted number wins the argument.
What's a Good PDO in the NHL?
A good PDO sits close to 100, since that's the league-average anchor shooting percentage and save percentage regress toward over a full season. Anything sustained above 102 or below 98 for 20+ games is a signal, not a trend — expect the number to move back toward the mean.
Do Possession Stats Predict Playoff Success?
Possession stats predict sustained regular-season performance more reliably than they predict single playoff series, since a hot goalie or a shooting-percentage heater can carry a lower-possession team through a short sample. Teams with strong xGF% profiles tend to be built for longer runs because their process doesn't depend on puck luck holding up.
“A team with a losing record and a winning xGF% is a bet on the process catching up to the standings, not the other way around.”
See who's leading the Hart Trophy race
Track which players are backing up strong possession numbers with results in 2026.
FAQ
How do you read NHL possession stats quickly?
Check CF% and xGF% against the 50% break-even line first, then check PDO to see if the team's record matches its process. Above 52% CF%/xGF% with a PDO near 100 is a genuinely strong team; anything else needs a closer look.
What is Corsi in hockey?
Corsi is the count of all shot attempts at 5-on-5 — goals, saves, misses, and blocks — expressed as a percentage of the total between two teams. It's the oldest and most widely cited possession stat in NHL analytics.
What is the difference between Corsi and Fenwick?
Corsi counts every shot attempt including blocked shots, while Fenwick excludes blocked shots and only counts attempts that reached or missed the net. Fenwick is considered slightly cleaner since blocking ability varies by team.
What is xGF% and why does it matter more than Corsi now?
xGF% is Expected Goals For %, which weighs each shot attempt by location and quality instead of counting every attempt equally. It matters more than raw Corsi because it separates teams generating high-danger chances from teams padding stats with low-percentage shots.
Is a 50% possession stat good or bad?
A 50% CF% or xGF% is break-even — the team is trading shot attempts evenly with opponents. Good starts around 52%, and elite teams typically run 53%+ across a full 2026 season.
What does PDO tell you about an NHL team?
PDO combines on-ice shooting percentage and save percentage, and it regresses toward 100 over a full season. A team well above or below 100 is running on luck that historically corrects.
How many games before possession stats are reliable?
Most analysts want at least 25 games before trusting a team's CF% or xGF% as a stable read, since smaller samples get distorted by hot streaks, injuries, and schedule quirks.
Do zone starts affect possession stats?
Zone starts affect possession stats significantly — a line that starts most shifts in its own end will post a worse CF% than its actual ability, since it's pinned back before the shift even begins.
One Last Thing
The possession number worth checking first isn't raw CF% — it's score-adjusted or close-score CF%, because trailing teams chase the game and inflate their own shot totals in a way that has nothing to do with real strength. Skip that adjustment and a team gets credit for possession numbers it only built after falling behind by two goals. SportsOrca's NHL breakdowns run possession alongside goaltending and rookie performance for exactly this reason — the raw number and the adjusted number can tell two different stories about the same team in 2026.
