How this instrument works
Save percentage (SV%) measures the share of shots on goal that a goaltender stops. It is computed as saves divided by shots against, where shots against is simply saves plus goals allowed — every shot that reached the goaltender either got stopped or went in. A goaltender who faces 30 shots and stops 27 has an SV% of 27/30, or 0.900.
The stat exists because raw save counts and goals-against totals depend heavily on how busy a defense keeps a goaltender. A backup who faces 15 shots a night and a starter who faces 35 are not comparable on saves alone, but their save percentages sit on the same 0-to-1 scale regardless of workload. That portability is why SV% became the standard shorthand for goaltender form in scouting reports and broadcast graphics.
In North American hockey, SV% is conventionally written as a three-decimal figure such as .915 rather than a percentage like 91.5%, a quirk carried over from batting-average notation in baseball. International competitions and some other leagues instead report it as a true percentage. Either way, the underlying ratio of saves to shots faced is identical — only the display convention differs.
SV% treats every shot as equally difficult to stop, which is both its strength and its limit. It is easy to compute from box-score numbers and easy to compare across games, but it says nothing about whether a goaltender faced mostly routine shots or a string of breakaways and rebounds. Analysts pair it with shot-quality models when they need a finer read on performance.
- Enter Saves — the total shots the goaltender stopped over the game, stretch of games, or season you're evaluating.
- Enter Goals against — how many of the shots the goaltender faced over that same span went in.
- Read Save percentage — the instrument adds saves and goals against to get shots faced, then divides saves by that total automatically.
- For single-game checks, remember shots against equals saves plus goals against — you don't need to look that number up separately.
- Compare save percentages over spans of at least a few hundred shots faced; a handful of games can swing SV% dramatically on shot luck alone.
Worked example — 27 saves, 3 goals against
A goaltender stops 27 of the shots faced in a game and allows 3 goals. Shots against is 27 + 3 = 30, so save percentage is 27 / 30 = 0.900 — a strong single-game showing, well above the roughly .900-to-.915 range a typical NHL starter posts over a full season.
Change either number and the read shifts fast: allow one more goal on the same 30 shots (27 saves, 4 goals against) and SV% drops to 27/31 ≈ 0.871, a reminder of how much a single soft goal moves the number in a small sample.
Questions
What counts as a shot against for save percentage?
Shots against is every shot on goal the goaltender faced — saves plus goals allowed. Shots that missed the net or were blocked by a skater before reaching the goaltender don't count toward either total, since the goaltender never had a chance to stop them. That's why the formula only needs saves and goals against as inputs; shots against falls out of those two automatically.
What's considered a good save percentage in the NHL?
League-average SV% for full-time NHL starters typically sits around .900 to .910 in modern seasons, with elite goaltenders posting .920 or higher over a full year. Single-game and even single-month figures swing far more widely than that — a .950 night or a .850 night both happen regularly without reflecting a goaltender's true level, which is why analysts lean on full-season or multi-season samples.
Does save percentage account for how hard the shots were?
No. Save percentage treats a routine shot from the blue line exactly the same as a point-blank breakaway — both count as one shot faced, and stopping either counts as one save. A goaltender on a team that allows a lot of high-danger chances can post a lower SV% than a goaltender facing mostly low-danger shots, even with comparable skill. Shot-quality-adjusted models exist precisely to correct for this gap.
Why do single-game save percentages look so extreme?
Because the sample is tiny. A goaltender who faces 20 shots and allows 2 sits at .900, but allow one extra goal on the same shot total and it drops to .850 — a swing that would take dozens of games to matter over a full season. Treat single-game SV% as a snapshot of one performance, not a measure of true talent; that only stabilizes over hundreds of shots faced.
Is save percentage shown as a decimal or a percentage?
Both conventions exist. North American leagues, including the NHL, typically write it as a three-decimal figure such as .915, borrowing the display style baseball uses for batting average. International competitions more often report the same ratio as a true percentage, such as 91.5%. The underlying calculation — saves divided by shots faced — is identical either way; only the notation changes.
How is save percentage different from goals-against average?
Goals-against average (GAA) counts goals allowed per 60 minutes of ice time, so it's sensitive to how many shots a goaltender's team allows them to face. Save percentage instead measures stopping rate per shot faced, which stays comparable whether a goaltender sees 20 shots or 40 in a game. A goaltender behind a weak defense can face a high shot volume and rack up a worse GAA while still holding a respectable SV%.