InfyCalculator
🏃

Save Percentage Calculator: Goalie SV%, GAA and Goals Saved Above Average

Calculate a goaltender's save percentage (.938 style and as a percent), goals-against average for any game length, goals saved above average (GSAA) against the league rate, and how many goals a target save percentage allows.

Shots against (shots on goal)
Goals against
Minutes played
Regulation game length
League-average save percentage
Target save percentage
Loading calculator…

How it works

Save % = (shots against − goals against) ÷ shots against · GAA = goals against ÷ minutes played × regulation game length (60) · GSAA = shots against × (1 − league save %) − goals against

Save percentage is the share of shots on goal a goaltender stops, written in North America as a three-digit decimal like .938 and in international hockey as a percent (93.8%). Only shots on goal count: a shot that misses the net or is blocked is not a shot against, and every goal was a shot on goal, so shots against = saves + goals. Goals-against average is the number of goals allowed per regulation game: goals ÷ minutes played × 60 in the NHL and most leagues (enter a different game length under More options for leagues with shorter periods, or 90 for soccer). GAA depends heavily on how many shots the team allows, which is why save percentage is the cleaner measure of the goalie. Goals saved above average compares the goalie with a league-average one facing the same shots: shots × (1 − league save %) − goals against. Positive means goals prevented. It does not adjust for shot quality the way expected-goals models (GSAx) do. Empty-net goals are not charged to a goalie, and shootout attempts are kept separate.

Worked example

A goalie faces 32 shots and allows 2 goals in 60 minutes. Saves = 30, so save % = 30 ÷ 32 = 0.9375, shown as .938 (93.8%). GAA = 2 ÷ 60 × 60 = 2.00. A .900 league-average goalie would allow 32 × 0.100 = 3.2 goals, so GSAA = 3.2 − 2 = +1.2. To post at least .920 on 32 shots, the goalie can allow at most ⌊32 × 0.08⌋ = 2 goals.

Sources: Hockey-Reference: glossary (SV%, GAA) · Hockey-Reference: NHL league averages by season · Hockey-Reference: 2024-25 NHL goalie stats (GSAA) · Omni Calculator: save percentage and GAA with regulation time · Wikipedia: Save percentage

Frequently asked questions

What is a good save percentage?

In the NHL the league average has been about .896 to .908 in recent seasons (.900 in 2024-25, .896 in 2025-26, per Hockey-Reference). Around .915 is very good and a full season above .925 is elite. Below .900 over a large sample usually means a struggling goalie. Lower levels and youth hockey run lower.

How do you calculate save percentage?

Divide saves by shots on goal. If a goalie faces 30 shots and allows 2 goals, saves are 28 and save percentage is 28 ÷ 30 = .933, or 93.3%.

How is goals-against average calculated?

GAA = goals against × 60 ÷ minutes played. Connor Hellebuyck allowed 125 goals in 3,741:22 minutes in 2024-25, so his GAA was 125 × 60 ÷ 3,741.37 = 2.00. For a league with 51-minute games, multiply by 51 instead of 60.

What is GSAA (goals saved above average)?

It is the number of goals a goalie saved compared with a league-average goalie on the same number of shots: shots × (1 − league save %) − goals against. Hellebuyck faced 1,664 shots and allowed 125 in 2024-25; with a .900 league average, GSAA = 166.4 − 125 = +41.4 (Hockey-Reference lists 41.6 using the unrounded league rate).

Why is it written as .938 instead of 93.8%?

North American hockey reports save percentage as a proportion to three decimal places, the way it appears on scoreboards and stat sheets. The IIHF and many European leagues show the same number as a percent, 93.8%.

Do empty-net goals count against a goalie's save percentage?

No. A goal scored while the goalie is pulled is charged to the team, not the goalie, so it is left out of the goalie's goals against and shots against. Shootout goals are also kept out of save percentage and GAA.

Related calculators

This calculator is for educational purposes. Double-check important results before acting on them.