How Hockey ChangedThe Vanishing Game
The Skill Gap Between NHL Goalies Shrank 60% From the 1970s to the 2010s
League save % climbed from .873 in 1981-82 to .915 in 2015-16 while the best and worst goalies closed in on each other. It has fallen since, as shots came from closer in and, from 2023-24, possibly because the NHL changed how it records shots.
By Edgehalla ResearchPublished 9 min read
On Feb. 24, 1982, at Memorial Auditorium, Wayne Gretzky scored his 77th goal of the season and broke Phil Esposito’s single-season record. He scored it against Don Edwards and the Buffalo Sabres. The season before, Edwards had stopped .898 of his shots, the best of any NHL goalie who faced 1,000.
Gretzky finished with 92 goals on 370 shots. League goalies stopped .873 of the shots they faced that season, the lowest in 70 seasons of shot counts.
In 2015-16, every goalie who faced 1,000 shots beat Edwards’s number. The worst of them, Montreal’s Mike Condon, stopped .903.
On this page
Save percentage fell, climbed for three decades, then fell again
In the six-team league, from 1955-56 to 1966-67, goalies stopped .904 to .919 of shots, as many as they stopped in the 2000s (.901 to .911). Then the league grew from 6 teams to 21, and save % fell for most of a dozen seasons. It hit bottom in Gretzky’s 92-goal season. From there it climbed for three decades, to .915 in 2015-16. In 2025-26 it was .896, the lowest since 1993-94.
Those decimals hide how many pucks are involved. Going from .915 to .896 means goalies now let in 19 more of every 1,000 shots. This story calls each of those a point, so the slide is 19 points. Goalies faced 72,530 shots in 2025-26. That is about 1,400 more goals.
Save % sank to its low in 1981-82, peaked in 2015-16 and has fallen since
NHL regular season, every season since 1955-56; goalie save % and a second line that also counts empty-net goals
Show data
| Season | Goalie save % | All goals, empty nets in |
|---|---|---|
| 1955-56 | 0.919 | – |
| 1956-57 | 0.911 | – |
| 1957-58 | 0.909 | – |
| 1958-59 | 0.906 | – |
| 1959-60 | 0.908 | 0.907 |
| 1960-61 | 0.908 | 0.907 |
| 1961-62 | 0.906 | 0.905 |
| 1962-63 | 0.908 | 0.907 |
| 1963-64 | 0.916 | 0.915 |
| 1964-65 | 0.909 | 0.908 |
| 1965-66 | 0.904 | 0.904 |
| 1966-67 | 0.907 | 0.906 |
| 1967-68 | 0.910 | 0.908 |
| 1968-69 | 0.908 | 0.907 |
| 1969-70 | 0.912 | 0.911 |
| 1970-71 | 0.903 | 0.902 |
| 1971-72 | 0.903 | 0.901 |
| 1972-73 | 0.896 | 0.895 |
| 1973-74 | 0.896 | 0.895 |
| 1974-75 | 0.890 | 0.889 |
| 1975-76 | 0.890 | 0.889 |
| 1976-77 | 0.891 | 0.889 |
| 1977-78 | 0.889 | 0.887 |
| 1978-79 | 0.883 | 0.881 |
| 1979-80 | 0.883 | 0.881 |
| 1980-81 | 0.876 | 0.874 |
| 1981-82 | 0.873 | 0.871 |
| 1982-83 | 0.875 | 0.874 |
| 1983-84 | 0.873 | 0.871 |
| 1984-85 | 0.875 | 0.873 |
| 1985-86 | 0.874 | 0.872 |
| 1986-87 | 0.880 | 0.878 |
| 1987-88 | 0.880 | 0.878 |
| 1988-89 | 0.879 | 0.877 |
| 1989-90 | 0.881 | 0.878 |
| 1990-91 | 0.886 | 0.884 |
| 1991-92 | 0.888 | 0.886 |
| 1992-93 | 0.885 | 0.883 |
| 1993-94 | 0.895 | 0.893 |
| 1994-95 | 0.901 | 0.898 |
| 1995-96 | 0.898 | 0.896 |
| 1996-97 | 0.905 | 0.902 |
| 1997-98 | 0.906 | 0.904 |
| 1998-99 | 0.908 | 0.906 |
| 1999-00 | 0.904 | 0.902 |
| 2000-01 | 0.903 | 0.901 |
| 2001-02 | 0.908 | 0.905 |
| 2002-03 | 0.909 | 0.906 |
| 2003-04 | 0.911 | 0.908 |
| 2004-05 | – | – |
| 2005-06 | 0.901 | 0.899 |
| 2006-07 | 0.905 | 0.903 |
| 2007-08 | 0.909 | 0.906 |
| 2008-09 | 0.909 | 0.906 |
| 2009-10 | 0.911 | 0.909 |
| 2010-11 | 0.913 | 0.910 |
| 2011-12 | 0.914 | 0.911 |
| 2012-13 | 0.912 | 0.909 |
| 2013-14 | 0.914 | 0.911 |
| 2014-15 | 0.915 | 0.911 |
| 2015-16 | 0.915 | 0.910 |
| 2016-17 | 0.914 | 0.910 |
| 2017-18 | 0.912 | 0.908 |
| 2018-19 | 0.910 | 0.905 |
| 2019-20 | 0.910 | 0.905 |
| 2020-21 | 0.908 | 0.903 |
| 2021-22 | 0.907 | 0.902 |
| 2022-23 | 0.904 | 0.899 |
| 2023-24 | 0.904 | 0.898 |
| 2024-25 | 0.900 | 0.894 |
| 2025-26 | 0.896 | 0.889 |
- 1981-82: Gretzky scores 92
- 1989-90: Pads widened to 12 in
- 2005-06: Pads cut to 11 in
- 2018-19: Fitted chest pads
The second line also counts goals scored into an empty net, so it runs lower: .889 in 2025-26. Its high since 1990 was .911 in 2013-14, and the six-team league reached .915 in 1963-64. Nobody is in net for an empty-netter, so the goalie line leaves them out. Coaches pull goalies far more often now, and the two lines have drifted from 2.2 points apart in 2005-06 to 6.6.
So goalies did get better at stopping pucks. Now put each goalie next to the rest of the league in his own season. Today’s best stand less far above the pack than Ken Dryden and Bernie Parent did 50 years ago, and today’s worst sit less far below it. The goalies bunched together.
The climb was a third fewer goals on the same shots
Start with the scoresheet. Every goal a goalie allows is a goal some shooter scored, so a rising save % could mean better goalies or worse shooters. League totals can’t say which. They can say whether teams shot less or just scored less. The table puts 1983-84, a season as low as 1981-82, next to the 2015-16 peak.
| 1983-84 | 2015-16 | Change | |
|---|---|---|---|
| Shots on goal a game | 61.0 | 59.2 | -3% |
| Goals on goalies a game | 7.74 | 5.04 | -35% |
| Save % | .873 | .915 | +42 points |
In 1983-84 a game had about 61 shots on goal and nearly 8 goals on goalies. In 2015-16 it had about 59 shots and about 5 goals. Same shots, about a third fewer goals.
Some of the later climb came from the power play. Power-play shots go in more often. Since 1998-99, when the NHL’s split by strength becomes complete, goalies have stopped about .85 to .88 of them, against about .90 to .92 at even strength. In 2005-06 about 25% of shots came on the power play. By 2015-16 it was about 15%. Save % rose 14 points over those ten seasons, and fewer power-play shots alone would have produced about 5 of them. The rest came from goalies stopping more shots at each strength, mostly at even strength.
Goalies also grew about five inches from 1980 on, though in each era the taller goalies stopped about as many shots as the shorter ones. The butterfly style, goalie coaches and lighter, bigger gear spread over the same years. League totals can’t say which of those mattered most. Setting goalies against each other, one season at a time, can show who got better. Start in the decade when they were furthest apart.
In the 1970s the best and worst goalies were playing different games
On one count, Bernie Parent’s 1973-74 in Philadelphia is the biggest season in the data. Put an average goalie in front of his 2,006 shots and about 73 more would have gone in. Parent won 47 games. Ken Dryden finished 27 to 37 points above the league in six of his seven full seasons in Montreal (in the other, 1974-75, he was 16 above). Jacques Plante, playing for Toronto in 1970-71, stopped .944, 40 points above the league. No goalie who faced 1,000 shots since 1955-56 has stood further above his own season.
At the other end was Washington. The Capitals of 1974-75 went 8-67-5 and allowed 446 goals in 80 games. Ron Low won 8 of his 48 games and stopped .855, 35 points below the league. Michel Belhumeur played 35 games and won none. That season Rogie Vachon, in Los Angeles, was 37 points above the league. The best and worst regular goalies were 72 points apart.
The chart below sets every regular goalie since 1955-56 against his league the same way. Each panel is a decade. The top line is the best tenth of goalies who faced 800 shots that season. The bottom line is the worst tenth.
The best and worst goalies moved closer to the league after the 1970s
Saves per 1,000 shots above or below the league that season, goalies with 800+ shots
points
1955-56 to 1959-6011 points
1960s12 points
1970s16 points
1980s22 points
1990s12 points
2000s13 points
2010s13 points
2020-21 to 2025-2615 points
Show data
| Season | 1955-56 to 1959-60: Best tenth | 1955-56 to 1959-60: Median | 1955-56 to 1959-60: Worst tenth | 1960s: Best tenth | 1960s: Median | 1960s: Worst tenth | 1970s: Best tenth | 1970s: Median | 1970s: Worst tenth | 1980s: Best tenth | 1980s: Median | 1980s: Worst tenth | 1990s: Best tenth | 1990s: Median | 1990s: Worst tenth | 2000s: Best tenth | 2000s: Median | 2000s: Worst tenth | 2010s: Best tenth | 2010s: Median | 2010s: Worst tenth | 2020-21 to 2025-26: Best tenth | 2020-21 to 2025-26: Median | 2020-21 to 2025-26: Worst tenth |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1955-56 | 8 points | 1 points | −8 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1956-57 | 13 points | 2 points | −8 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1957-58 | 16 points | −1 points | −8 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1958-59 | 13 points | −1 points | −9 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1959-60 | 11 points | 1 points | −11 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1960-61 | – | – | – | 13 points | 1 points | −15 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1961-62 | – | – | – | 14 points | 7 points | −27 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1962-63 | – | – | – | 8 points | 4 points | −6 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1963-64 | – | – | – | 15 points | 2 points | −4 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1964-65 | – | – | – | 15 points | 1 points | −11 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1965-66 | – | – | – | 15 points | −1 points | −12 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1966-67 | – | – | – | 17 points | 10 points | −18 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1967-68 | – | – | – | 23 points | 2 points | −19 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1968-69 | – | – | – | 18 points | 2 points | −8 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1969-70 | – | – | – | 12 points | 3 points | −10 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1970-71 | – | – | – | – | – | – | 16 points | 1 points | −15 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1971-72 | – | – | – | – | – | – | 28 points | −1 points | −16 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1972-73 | – | – | – | – | – | – | 20 points | 7 points | −16 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1973-74 | – | – | – | – | – | – | 13 points | 2 points | −15 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1974-75 | – | – | – | – | – | – | 21 points | −2 points | −19 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1975-76 | – | – | – | – | – | – | 20 points | −1 points | −16 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1976-77 | – | – | – | – | – | – | 21 points | −2 points | −22 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1977-78 | – | – | – | – | – | – | 23 points | −1 points | −20 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1978-79 | – | – | – | – | – | – | 22 points | 1 points | −19 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1979-80 | – | – | – | – | – | – | 16 points | −1 points | −13 points | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 1980-81 | – | – | – | – | – | – | – | – | – | 19 points | 2 points | −16 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1981-82 | – | – | – | – | – | – | – | – | – | 21 points | −2 points | −14 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1982-83 | – | – | – | – | – | – | – | – | – | 23 points | 5 points | −17 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1983-84 | – | – | – | – | – | – | – | – | – | 20 points | 3 points | −17 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1984-85 | – | – | – | – | – | – | – | – | – | 13 points | 4 points | −17 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1985-86 | – | – | – | – | – | – | – | – | – | 20 points | 3 points | −18 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1986-87 | – | – | – | – | – | – | – | – | – | 13 points | 2 points | −7 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1987-88 | – | – | – | – | – | – | – | – | – | 16 points | 5 points | −13 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1988-89 | – | – | – | – | – | – | – | – | – | 18 points | 1 points | −11 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1989-90 | – | – | – | – | – | – | – | – | – | 22 points | 4 points | −14 points | – | – | – | – | – | – | – | – | – | – | – | – |
| 1990-91 | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 6 points | −14 points | – | – | – | – | – | – | – | – | – |
| 1991-92 | – | – | – | – | – | – | – | – | – | – | – | – | 22 points | −3 points | −9 points | – | – | – | – | – | – | – | – | – |
| 1992-93 | – | – | – | – | – | – | – | – | – | – | – | – | 16 points | 2 points | −9 points | – | – | – | – | – | – | – | – | – |
| 1993-94 | – | – | – | – | – | – | – | – | – | – | – | – | 19 points | 2 points | −20 points | – | – | – | – | – | – | – | – | – |
| 1994-95 | – | – | – | – | – | – | – | – | – | – | – | – | 13 points | 4 points | −10 points | – | – | – | – | – | – | – | – | – |
| 1995-96 | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 5 points | −15 points | – | – | – | – | – | – | – | – | – |
| 1996-97 | – | – | – | – | – | – | – | – | – | – | – | – | 17 points | −1 points | −16 points | – | – | – | – | – | – | – | – | – |
| 1997-98 | – | – | – | – | – | – | – | – | – | – | – | – | 13 points | −1 points | −13 points | – | – | – | – | – | – | – | – | – |
| 1998-99 | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 0 points | −15 points | – | – | – | – | – | – | – | – | – |
| 1999-00 | – | – | – | – | – | – | – | – | – | – | – | – | 12 points | 3 points | −12 points | – | – | – | – | – | – | – | – | – |
| 2000-01 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 17 points | 2 points | −11 points | – | – | – | – | – | – |
| 2001-02 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 12 points | −2 points | −9 points | – | – | – | – | – | – |
| 2002-03 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 14 points | 2 points | −15 points | – | – | – | – | – | – |
| 2003-04 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 14 points | 1 points | −12 points | – | – | – | – | – | – |
| 2004-05 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| 2005-06 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 18 points | 1 points | −16 points | – | – | – | – | – | – |
| 2006-07 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 3 points | −12 points | – | – | – | – | – | – |
| 2007-08 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 12 points | 1 points | −13 points | – | – | – | – | – | – |
| 2008-09 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 11 points | 4 points | −10 points | – | – | – | – | – | – |
| 2009-10 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 13 points | 1 points | −11 points | – | – | – | – | – | – |
| 2010-11 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 10 points | 3 points | −16 points | – | – | – |
| 2011-12 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 1 points | −13 points | – | – | – |
| 2012-13 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 4 points | −9 points | – | – | – |
| 2013-14 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 11 points | 1 points | −14 points | – | – | – |
| 2014-15 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 11 points | 2 points | −11 points | – | – | – |
| 2015-16 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 9 points | 3 points | −9 points | – | – | – |
| 2016-17 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 10 points | 1 points | −8 points | – | – | – |
| 2017-18 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 1 points | −11 points | – | – | – |
| 2018-19 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 16 points | 3 points | −12 points | – | – | – |
| 2019-20 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 13 points | 1 points | −12 points | – | – | – |
| 2020-21 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 19 points | 4 points | −6 points |
| 2021-22 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 13 points | 3 points | −11 points |
| 2022-23 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | −1 points | −17 points |
| 2023-24 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 12 points | 1 points | −13 points |
| 2024-25 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 14 points | 2 points | −9 points |
| 2025-26 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | 15 points | 3 points | −13 points |
The bands are widest in the 1970s and pull in after that. The bottom line climbs from about 17 points below the league in the 1970s to about 11 below in the 2010s. The top line comes down too, from about 20 points above to about 12.
The league had tripled from 6 teams in 1966-67 to 18 in 1974-75, and from 1972 a rival league, the WHA, bid for the same players. The WHA’s history describes the two leagues together badly straining the talent pool. Our numbers can’t show why goalies spread so far apart, but our best guess is that there were not enough good ones to go around. Dryden’s 37 in 1975-76, that Vachon season and Parent’s 36 in 1973-74 are still the three best seasons against the league by goalies who faced 1,500 shots.
Then the goalies bunched together
Some of what separates two goalies in a season is bounces. Take a point shot that ticks off a stick and changes direction, or a wrister that rings off the post and out. Give a goalie who is exactly league average 1,500 shots, about a season’s work for a regular starter, and bounces alone can leave him anywhere from about 15 points above the league to 15 below. Our GOAT series used the same math: take the spread that bounces alone would produce out of each season’s spread between goalies, and what is left is skill. With the bounces gone, a typical regular goalie in the 1970s was still about 12.5 points better or worse than average. In the 2010s he was about 5.0. The skill gap shrank about 60%.
After luck, the skill gap between goalies shrank about 60% from the 1970s to the 2010s
How far a typical goalie sat from average on skill alone, in saves per 1,000 shots; goalies with 800+ shots, average of the decade’s seasons
The skill spread between goalies was 12.5 points in the 1970s and 5.0 in the 2010s, with a partial rebound to 6.5 since 2020-21.
The tallest bar is the 1970s and the shortest is the 2010s. The 2020s bar has come back up to about 6.5, though that rests on only six seasons.
Dominik Hasek was the last goalie who stood out like the 1970s stars. He led the NHL in save percentage six straight seasons from 1993-94. Stu Hackel’s NHL.com profile of Hasek quotes a line from Sports Illustrated’s Michael Farber: “He tends goal the way Kramer enters Seinfeld’s apartment.” His 35 points above the league in 1993-94 are the best by a goalie with 1,500 shots since the 1970s. That same season Ottawa’s Craig Billington was 36 points below it.
Nobody has stood out that far since. The best season by a goalie with 1,500 shots since 2000 is Igor Shesterkin’s 28 points above the league in 2021-22. Meanwhile the bottom closed in too. By 2015-16, the year save % peaked, the best and worst of the 42 goalies who faced 800 shots were 29 points apart. The spread between them was no wider than luck alone would produce. Treat that with care. It is one season, and bounces can make a single season look tighter than it was.
All of this counts goalies who faced at least 800 shots. Raise the bar to 1,500 and only the busiest starters are left. More shots each means less luck, and the drop is steeper: 12.4 points in the 1970s, 3.2 in the 2010s, about 74% smaller. The decades in between shift with the bar. At 1,500 shots the 1980s look tighter than the 1990s. At 800 shots they look a little wider.
One good season stopped meaning much
When goalies are this close, one season should say little about the next. Brian Elliott went from near the bottom to the top in a year. In 2010-11, split between Ottawa and Colorado, he stopped .893 and ranked 36th of 38 goalies who faced 800 shots. The next season, in St. Louis, he stopped .940, the best of them all. In 2015-16 he led the league at .930. A year later, in Calgary, he was 26th of 34.
Elliott was not alone. Line up every regular goalie’s save % in back-to-back seasons and the same thing shows across the league. In the 1970s the good ones mostly stayed good the next year (a year-to-year correlation of 0.63). In the 2010s one season told you almost nothing about the next (0.14). Dryden led the league in 1976-77 and again in 1977-78. Since Hasek, only Connor Hellebuyck has topped the goalies with 1,000 shots two seasons running, in 2023-24 and 2024-25.
Even that run has an asterisk. In both seasons Anthony Stolarz stopped a higher share of his shots. He played about half as many games, so he faced too few shots to make the 1,000-shot list, and Hellebuyck led it. With goalies this close, a season of save % tells you more about the team in front of a goalie and his bounces than about him. Our goalie team-effect research takes that up.
No pad rule left a clean mark
If bigger gear helped goalies close the gap, the NHL’s rules shrinking that gear should have knocked save % back down. The league has changed the gear rules many times. A rule that bit hard would show up as a step down right after it. The table sets the three seasons before each rule against the rule’s first season and the two after.
| Rule | First season | 3 seasons before | Rule season + 2 | Change | Even strength change |
|---|---|---|---|---|---|
| Pad width raised from 10 to 12 inches | 1989-90 | .879 | .885 | +5.5 | n/a |
| Pad length capped at 38 inches | 2003-04 | .907 | .906 | −0.9 | +1.6 |
| Pad width cut to 11 inches; equipment reduced; trapezoid | 2005-06 | .909 | .905 | −4.0 | −1.7 |
| Pad thigh rise capped at 55% of knee to pelvis | 2010-11 | .910 | .913 | +3.3 | +1.4 |
| Thigh rise cut to 45%; shallower nets | 2013-14 | .913 | .914 | +1.5 | +1.4 |
| Chest and arm protectors made size-specific | 2018-19 | .913 | .909 | −4.2 | −5.2 |
Neither drop is clean. In 2005-06, smaller pads arrived with a rulebook rewrite after the lockout: two-line passes allowed, bigger offensive zones, and the trapezoid, the area behind the net outside which goalies may not play the puck. Power plays jumped too, from 18.7% of shots in 2003-04 to 24.8% in 2005-06. Save % fell about 4 points, about 2 at even strength, and by 2009-10 it was back above its 2003-04 level. The 2010-11 and 2013-14 rules trimmed the thigh rise, the stretch of pad above the knee. Save % went up after both.
The 2018 chest-pad rule came after the gap had started to widen
The other drop followed the 2018-19 rule that made chest and arm protectors fit each goalie’s size. This time the league made a prediction. Kay Whitmore, the NHL’s vice president of hockey operations, told NHL.com the effect on scoring could not be predicted, but the gap between the best and average goalies should grow. The gap did grow. It had started two seasons before the rule. After luck, the skill gap between goalies was zero in 2015-16, 1.9 points in 2016-17 and 6.8 in 2017-18, the season before the rule. It has run 3.6 to 10.9 since. Save % fell 4.2 points across the rule’s window, in the middle of a slide that began in 2016-17.
Since 2015-16, save % has fallen at even strength and on the penalty kill alike
Gear rules don’t explain the fall since 2015-16, and neither does the power play. About 15% of shots came on the power play in both 2015-16 and 2025-26, and the small change in that mix moves save % by only 0.1 points. Save % fell at even strength and on the penalty kill.
Since 2015-16, save % fell about as much on the penalty kill as at even strength
NHL regular season, from the first season the strength split is complete
Show data
| Season | Even strength | Against the power play |
|---|---|---|
| 1998-99 | 0.916 | 0.872 |
| 1999-00 | 0.912 | 0.866 |
| 2000-01 | 0.914 | 0.861 |
| 2001-02 | 0.916 | 0.872 |
| 2002-03 | 0.918 | 0.869 |
| 2003-04 | 0.922 | 0.867 |
| 2004-05 | – | – |
| 2005-06 | 0.915 | 0.861 |
| 2006-07 | 0.916 | 0.865 |
| 2007-08 | 0.920 | 0.867 |
| 2008-09 | 0.919 | 0.868 |
| 2009-10 | 0.919 | 0.874 |
| 2010-11 | 0.921 | 0.875 |
| 2011-12 | 0.921 | 0.875 |
| 2012-13 | 0.920 | 0.865 |
| 2013-14 | 0.921 | 0.878 |
| 2014-15 | 0.922 | 0.874 |
| 2015-16 | 0.923 | 0.873 |
| 2016-17 | 0.921 | 0.871 |
| 2017-18 | 0.920 | 0.870 |
| 2018-19 | 0.917 | 0.864 |
| 2019-20 | 0.917 | 0.866 |
| 2020-21 | 0.915 | 0.865 |
| 2021-22 | 0.914 | 0.865 |
| 2022-23 | 0.912 | 0.860 |
| 2023-24 | 0.911 | 0.862 |
| 2024-25 | 0.908 | 0.849 |
| 2025-26 | 0.903 | 0.853 |
Both lines slide together after 2015-16. At even strength goalies went from .923 to .903, down 20 points. Killing penalties they went from .873 to .853, down 20. Goalies were slipping whether their team was at full strength or a man short.
The shots got closer, then the NHL changed how it logs them
The next suspect is where the shots came from. A shot from the slot goes in far more often than one from the blue line, so if more of a season’s shots come from close in, save % falls even with the same goalies. Our play-by-play, the NHL’s log of every shot and where on the ice it was taken, covers each season from 2017-18. We took every shot on goal and looked up how often shots from that distance, at that strength, went in from 2017-18 to 2019-20. Add those chances up for a season and you get the save % goalies would have had if only the shot locations had changed. That is the dashed line.
Save % followed the closer shots until 2023-24, then fell below them
Shots on goal with a goalie in net, our play-by-play, NHL regular season; dashed line = save % if only shot locations had changed
Show data
| Season | Actual | Shot mix at fixed rates |
|---|---|---|
| 2017-18 | 0.912 | 0.911 |
| 2018-19 | 0.910 | 0.911 |
| 2019-20 | 0.910 | 0.911 |
| 2020-21 | 0.908 | 0.910 |
| 2021-22 | 0.907 | 0.908 |
| 2022-23 | 0.904 | 0.905 |
| 2023-24 | 0.904 | 0.907 |
| 2024-25 | 0.900 | 0.909 |
| 2025-26 | 0.896 | 0.905 |
- 2023-24: Shot-recording break
Through 2022-23 the two lines move together. The shots did get closer: 29.9% of shots on goal came from inside 20 feet in 2018-19, 32.0% in 2022-23. That accounts for the whole drop. One season doesn’t fit. From 2017-18 to 2018-19 save % fell 2.4 points with almost no change in where shots came from, a dip small enough to be luck.
In 2023-24 the solid line drops away from the dashed one, and by 2025-26 it is 9.0 points below it. That same season the way the NHL records shots changed, and many more attempts, mostly from close range, are now logged as misses (see our shots-on-goal story). David Johnson of Hockey Analysis has suggested that people tracking shots by hand may have counted some as saves that the league’s puck- and player-tracking system now shows were going wide. He offers it as a theory. If it holds, part of the recent drop happened in the record book while the goalies stayed the same.
We are surer about the shrinking gap than about the latest drop
The shrinking gap is the sturdiest finding here. It sets each goalie against his own league, so it doesn’t matter whether save % as a whole went up or down. It holds at 800 shots and at 1,500. It shows up again in how little one season now predicts the next. Two pieces of it are thinner. The 2015-16 result is one season, and one season of bounces can make goalies look closer together than they were. It does not mean every goalie that year was equally good. The rebound since 2020-21 rests on six seasons and began before the 2018 rule.
Some things the numbers cannot pull apart. Bigger goalies, the butterfly, coaching and gear all arrived together. Shot counts before 1983-84 are not checked against an independent source. From 2023-24, closer shots and the change in how shots are recorded are tangled together. And Elliott’s and Hellebuyck’s careers show what the shrinking gap looks like for one goalie. They do not prove it.
The MVP came back to the pack
In June 2025 Hellebuyck won the Hart Trophy, the first goalie to do it since Carey Price in 2015, and his third Vezina, after a .925 season and a league-high 47 wins. In 2025-26 he stopped .895. The league stopped .896. On his 1,547 shots, he finished within a goal of what an average goalie would have allowed.
The leader among goalies with 1,000 shots was Colorado’s Scott Wedgewood, in the busiest season of his career: 45 games, when he had never played more than 37. He stopped .921. Don Edwards led that list 45 years earlier at .898, the season before Gretzky beat him for No. 77.
Share this article
Goalie tools
Live Edgehalla tools on the same subject.
Show the work
- Tests written down
- Not in advance (exploratory)
- Results computed
League save % for every NHL regular season since 1955-56 from the NHL goalie totals (saves ÷ shots against), checked against team totals with empty-net goals. Named goalie-seasons come from the same totals: relative save % is a goalie’s save % minus the league’s that season, in points (× 1,000); goals saved above average (GSAA) is saves minus shots × league save %; ranks are by save % among goalies with 800+ or 1,000+ shots. The skill spread is the spread in save % among goalies with 800+ shots after taking out the spread binomial luck alone would produce; we checked it at 1,500+ shots and against how well one season predicts the next. Rule windows compare three seasons before each equipment rule with three from it. The recent fall is split by strength (NHL saves by strength, 1998-99 on) and by shot mix in our play-by-play (2017-18 on, about 670,000 shots on goal).
| Test | Result | Verdict |
|---|---|---|
| The 1980s low is a shot-counting artefact | 1983-84 (.873) is as low as 1981-82; goals a game fell 35% by 2015-16 | Does not hold |
More tests below
| Test | Result | Verdict |
|---|---|---|
| The gap between goalies shrank | Skill spread −60% (800+ shots), −74% (1,500+); repeatability 0.63 to 0.14 | Holds |
| The top rose, not just the bottom | Best seasons vs the league (1,500+ shots): Ken Dryden 1975-76 +36.8, Rogie Vachon 1974-75 +36.6, Bernie Parent 1973-74 +36.2; best since 2000 Igor Shesterkin 2021-22 +27.6. Best-to-worst gap (800+) 71.8 in 1974-75, 28.8 in 2015-16 | Does not hold: both ends moved in |
| 2015-16 spread wider than luck (800+ shots) | Observed SD 7.0 vs binomial 7.8 points, 42 goalies; skill spread 0 | No wider than luck (one season) |
| Equipment rollbacks reversed the rise | Dips after 2005-06 and 2018-19 both came with other changes; 2010-11 and 2013-14 rose | Does not hold |
Continues below
| Test | Result | Verdict |
|---|---|---|
| The 2018 rule widened the gap | Skill spread 0.0 (2015-16), 1.9 (2016-17), 6.8 (2017-18), then 3.6 to 10.9 | Widening began before the rule |
| The power play explains the fall since 2015-16 | Strength mix +0.1 points of −19.1; even strength .923 to .903, vs facing the power play .873 to .853 | Does not hold |
| Our expected-goals model can judge league save % | Its totals are tuned to match each season’s goals, so it is circular at league level; the shot-mix line uses fixed 2017-18 to 2019-20 goal rates by distance band and strength instead | Not used |
| Shot quality drove the recent fall | Fixed rates −5.6 vs actual −5.6 (2018-19 to 2022-23); 2017-18 to 2018-19 −2.4 with almost no mix change, under two standard errors; 2025-26 gap 9.0 points after the recording break | Partly holds |
Frequently asked questions
What is the average save percentage in the NHL?
League goalies stopped .896 of shots in 2025-26 (saves ÷ shots against, empty-net goals excluded), the lowest since 1993-94. Counting empty-net goals it was .889, the lowest since 1992-93.
What was the highest league save percentage in NHL history?
In the modern era, .915 in 2015-16. The six-team league recorded similar or higher numbers (.919 in 1955-56, the first season with shots against), but shot counts from that era are less reliable.
Who had the best save percentage season in NHL history?
Among goalies who faced 1,500 shots, Tim Thomas stopped .938 for Boston in 2010-11, and Dominik Hasek .937 for Buffalo in 1998-99. Measured against the league that season, the best were Ken Dryden in 1975-76 (37 points above it), Rogie Vachon in 1974-75 and Bernie Parent in 1973-74; on 1,000 shots, Jacques Plante’s .944 in 1970-71 (40 above).
Why did NHL save percentage go up from the 1980s to the 2010s?
Goals fell by about a third while shots on goal barely changed, and over the same decades the best and worst goalies moved closer together. Fewer power-play shots explain part of the climb after 2005-06. The rest coincides with bigger goalies, the butterfly style, coaching and gear, which league totals cannot separate.
Why is NHL save percentage down?
From 2018-19 to 2022-23, shots on goal came from closer in, and that accounts for the whole drop. From 2023-24 save % fell further than the shot mix predicts, in the same season the NHL’s shot records changed. One analyst, David Johnson of Hockey Analysis, has suggested some stops once counted as saves are now logged as missed shots.
Did smaller goalie pads lower save percentage?
Not visibly. Save % fell after the 2005-06 equipment cuts, but those came with a wider rulebook rewrite, and save % was back above its pre-lockout level by 2009-10. The 2010-11 and 2013-14 pad cuts were followed by higher save %, and the 2018-19 chest and arm rules came during a slide that began in 2016-17.
More in The Vanishing Game
- Verdict: Data story
NHL Fights Fell to 0.18 a Game in 2018-19, Then Climbed Past the 0.20 of Gordie Howe’s Era
- Verdict: Data story
Is the Rush Dying in the NHL?
- Verdict: Data story
The Last Goalie Who Played Every Night
- Verdict: Data story
Did Garbage Goals Take Over the NHL?
- Verdict: Data story
For 19 Seasons, an NHL Hit Depended on the Rink
- Verdict: Data story
Hainsworth’s 22 Was Mostly Arithmetic
Sources and method
- NHL stats (NHL.com), every season since 1917-18
- NHL stats API: goalie summary and saves by strength; team and skater summaries
- NHL.com: This date in NHL history, Feb. 24 (Gretzky’s 77th goal, against Don Edwards and the Sabres)
- NHL.com, Stu Hackel: Dominik Hasek, 100 Greatest NHL Players (six straight save % titles; Farber quote)
- Sports Illustrated: Connor Hellebuyck becomes first goalie to win Hart Trophy since 2015 (June 12, 2025)
- Hockey Analysis, David Johnson: Are changes in NHL shot data tracking hiding that teams are getting better at defending? (March 21, 2025)
- Boston Globe: the evolution of goalie pad rules (2013)
- NHL.com, Nicholas J. Cotsonika: NHL, NHLPA add chest and arm pad rules for goalies (2018)
- Scouting the Refs: NHL historical rule changes
- Wikipedia: World Hockey Association (1972-1979)
Research and analysis, not betting advice. Past results do not guarantee future returns; bet only what you can afford to lose.