Subscribe
Carey Price of the Montreal Canadiens crouched in his crease, seen from directly above the net

A first-shot goal is not automatically a bad goal

A goal on the first shot gets treated as the one a goalie can't allow. We put all 67,422 first shots NHL starters have faced since 1997-98 in context against every other shot.

David Hutchison | | 8 min read
Key Takeaways
  • NHL starters have faced 67,422 first shots since 1997-98, and they stop the first one more often than they stop the rest. The size of the gap, and the reason for it, are both measured.
  • A first-shot goal predicts very little about the rest of the night. The gap that does exist is small, and part of it has nothing to do with the goalie.
  • The first shot of the game is not the only first shot. Goalies coming on in relief, most of them mid-period and usually right after a goal against, get their own numbers.
  • Save percentage has fallen across the league for a decade, and the fair first-shot bar for a goalie playing today has moved with it.
  • Even Carey Price, tied for the top first-shot save percentage since 1997-98, gave up 30. See where Rinne, Lundqvist, Brodeur, and Hellebuyck rank.

We’ve been frustrated for a while by how a goal on the first shot gets talked about. In the stands, on the bench, and in the car ride home, it gets treated as the one goal a goalie simply cannot give up, and as proof he wasn’t ready.

But one shot can’t tell you that. A first-shot goal says no more about a goalie than any other single goal does. To know whether a goalie really struggles with the first shot, you have to look at all of his first shots, and compare them with what every other goalie does with theirs.

So that’s what we did. The NHL’s game data has logged every shot on goal, and the goalie who faced it, since the 1997-98 season. For every regular-season game from then through Oct. 27, 2026, we found the first shot each starting goalie faced: 67,422 of them.

We started with save percentage, which is a blunt tool. It counts a dump-in from the red line the same as a one-timer from the slot. So we also looked at where those first shots come from, and that turned out to matter as much as who was in net.

.931League save % on a starter’s first shot
.908League save % on every shot
1 in 15Starts where the first shot goes in

It happens about once every 15 starts

Across the league, 4,647 of those 67,422 first shots went in. That is a .931 save percentage, well above the .908 goalies posted on all shots over the same seasons.

With two starters a night, a first-shot goal turns up in about one NHL game in 8. A goalie who starts 60 games at the league rate will give one up four or five times a season.

And it predicts very little about the rest of the night. Starters who gave up a goal on the first shot stopped .904 of the shots they faced the rest of the game; starters who stopped it, .908. Across more than 4,600 of those games, that gap is too consistent to be luck, but it is small: about one extra goal for every 10 games or so, and some of that is the other team simply being better that night. A first-shot goal is not a sign that a bad night is coming.

Line chart: league save percentage by season since 1997-98, starters' first shot above all shots every season
The starter’s first-shot save percentage has run above the overall rate in every full season since 1997-98. The NHL did not play in 2004-05.

First shots are easier shots

Since 2018-19, the NHL’s data has also recorded where on the ice each shot was taken. The first shot a starter faces usually comes from farther out: a median of 36 feet, against 32 feet for every other shot. A quarter of first shots come from inside 20 feet, compared with 31% of the rest.

A first shot is also rarely a second chance. Only 3.3% of first shots came within three seconds of another attempt by the same team, against 10% of other shots. Only 8% came on a power play or other special-teams situation, against 20% of the rest. The first shot arrives a median of 1:42 into the game, and about a third come in the first minute.

Bar chart: share of shots on goal by distance, first shots against all other shots, 2018-19 to 2026
Share of shots on goal by distance from the net, regular season 2018-19 to Oct. 27, 2026. Empty-net goals excluded.

Put simply, goalies stop the first shot more often because it is usually an easier shot. Account for distance, second chances, and special teams, and most of the gap between first shots and every other shot goes away. The sidebar at the end has the detail.

Today’s bar is lower

So the fair yardstick for a goalie’s first shot is the league’s first-shot save percentage, not his overall one. And that number has moved.

Save percentage has fallen across the league for a decade, from .915 on all shots in 2015-16 to .896 last season, and first shots have fallen with it, from .942 in 2016-17 to .917 last season. The .931 average since 1997-98 leans on the higher-saving years, so it is too high a bar for a goalie playing now.

Last three full seasons (2023-24 to 2025-26)

.922League save % on a starter’s first shot
.900League save % on every shot
1 in 13Starts where the first shot goes in

Over the last three full seasons, starters stopped .922 of their first shots and .900 of all shots, and a first shot went in about once every 13 starts. For a goalie today, about .922 is the fair bar. Keep that in mind on the lists that follow: a goalie who played mostly in the 2010s was measured in higher-saving years than one who broke in recently.

Coming in cold, and coming out for the second

The higher save percentage belongs to the first shot of the game, not to every first shot. Goalies coming on in relief stopped their first shot at .909 (4,928 of them), right in line with the .908 league save percentage on every shot.

Most of those goalies didn’t get an intermission to get ready. Of the relief appearances we could place, 2,797 started in the middle of a period, more than four in five of them right after a goal against. Those goalies had been sitting since the pregame warm-up and walked into play against shooters who had been on the ice for minutes, and they still stopped that first shot at .909. That deserves more credit than it sounds.

The other 1,681 came in at the start of a period, coming out of the same intermission as everyone else. They stopped the first shot at .916, much like a starter’s first shot of the second period (.914) or the third (.918). Once the game is under way, the first shot of a period looks much like the rest, whoever is in net.

Since 1997-98: Price at the top

Among the 91 goalies who have played at least 300 games since 1997-98, Carey Price and Ondrej Pavelec share the top first-shot save percentage at .957, and Price did it over nearly twice as many starts: 30 first-shot goals in 699 starts, about 18 fewer than a league-average goalie would have allowed. Of those 91 goalies with 300 or more games, Frederik Andersen (.952) and Igor Shesterkin (.948) rank highest among the ones who played last season or this one.

At the other end, Brent Johnson has the lowest rate, .889 in 280 starts, and Jacob Markstrom gave up the most extra goals: 60 in 566 starts, about 21 more than a league-average goalie would have since 2010-11. That is roughly one extra goal every 27 starts. His .907 on every other shot is right around league average, and so is Philipp Grubauer’s .909. This is about first shots, not about either goalie.

Dot chart: first-shot save percentage with 95% ranges for the top 5, familiar names and lowest 5 goalies since 1997-98
The dot is each goalie’s first-shot save percentage. The shorter the line around it, the more sure we can be of it; the sidebar at the end explains the lines. † NHL career began before 1997-98, so only part of it is counted.
GoalieSeasonsFirst-shot goals / startsFirst-shot save %All shots save %Goals vs. average*
Top 5
Carey Price2007-202230 / 699.957.917+18.2
Ondrej Pavelec2007-201816 / 373.957.907+9.7
Jaroslav Halak2006-202325 / 556.955.915+13.3
Cory Schneider2008-202218 / 389.954.918+8.8
Frederik Andersen2013-202626 / 537.952.913+11.0
Lowest 5
Brent Johnson1998-201231 / 280.889.904-11.7
Philipp Grubauer2012-202740 / 371.892.908-14.4
Jacob Markstrom2010-202760 / 566.894.906-21.0
Chris Mason1998-201329 / 275.895.910-10.0
Arturs Irbe†1997-200433 / 333.901.906-10.0

* Goals a goalie stopping first shots at the league’s .931 would have allowed on the same shots, minus the goals he actually allowed. Positive is better. † NHL career began before 1997-98, so only part of it is counted.

Goalies playing now

Frederik Andersen #31 Carolina Hurricanes goalie sprays water bottle during stoppage in play at Lenovo Center
Frederik Andersen’s .952 on the first shot is the highest among goalies who played last season or this one and have 300 or more games. Photo by Katherine Gawlik/Icon Sportswire

For goalies who played in 2025-26 or so far in 2026-27, we lowered the bar to 150 games played, which brings in 50 goalies. Kevin Lankinen, Thatcher Demko, and Frederik Andersen lead for now. Andersen is the leader among goalies playing now who have also reached 300 games.

With this few starts, one goal moves a save percentage by more than six points, so this list can change quickly. Earlier in a career it swings much more. Devin Cooley had stopped the first shot he faced in 40 of his first 43 NHL games (.930), counting five he came into in relief. When the first shot he faced against Edmonton on Oct. 28, 2026, went in, it fell to .909. One shot in one game moved it 21 points. Of the top five, only Andersen is clearly better than today’s .922 bar.

Do the leaders simply get easier first shots? Only slightly. From 2018-19, the first season the NHL recorded where shots were taken, we worked out what each goalie’s first shots should have been saved at, given how far out they came from, whether they came right after another attempt, and whether a team was on the power play. Among goalies who have faced at least 150 first shots since then, the 10 best stopped .948 of them and the 10 worst .899. Going by the shots alone, they should have stopped .926 and .923. Lankinen and Demko led while facing slightly harder first shots than average. The shots explain very little of the gap between the best and the worst.

Dot chart: first-shot save percentage with 95% ranges for goalies who played in 2025-26 or 2026-27
Played in the 2025-26 or 2026-27 regular season, at least 150 games played.
GoalieSeasonsFirst-shot goals / startsFirst-shot save %All shots save %Goals vs. average*
Top 5
Kevin Lankinen2020-20279 / 200.955.895+4.8
Thatcher Demko2017-202612 / 256.953.909+5.6
Frederik Andersen2013-202626 / 537.952.913+11.0
Joey Daccord2018-20278 / 168.952.903+3.6
Igor Shesterkin2019-202717 / 325.948.917+5.4
Lowest 5
Philipp Grubauer2012-202740 / 371.892.908-14.4
Jacob Markstrom2010-202760 / 566.894.906-21.0
Adin Hill2017-202721 / 206.898.904-6.8
Lukas Dostal2021-202716 / 171.906.897-4.2
Spencer Knight2020-202714 / 150.907.903-3.7

* Goals a goalie stopping first shots at the league’s .931 would have allowed on the same shots, minus the goals he actually allowed. Positive is better. † NHL career began before 1997-98, so only part of it is counted.

Where the big names sit

Readers will ask, so here are some familiar goalies among the 91 with at least 300 games played. Most sit within a few goals of average across a whole career.

And even the gaps between them may not mean much. For each goalie, we worked out the range his true first-shot save percentage most likely falls in, given how many first shots he has faced (statisticians call it a 95% confidence interval). For every goalie in this table, that range includes the league average of .931. Even Rinne’s runs from .930 to .963. So on first shots, the numbers can’t tell any of them apart from an average goalie, and the order could come down to a handful of bounces. The sidebar at the end, “The numbers behind it,” explains how the ranges work.

Pekka Rinne Nashville Predators goalie in yellow-blue mask holding stick in crease during NHL game
Pekka Rinne stopped the first shot at .949, the highest of the familiar names below. He goes into the Hockey Hall of Fame with Carey Price on Nov. 9.
Rank of 91GoalieFirst-shot goals / startsFirst-shot save %All shots save %Goals vs. average*
8Pekka Rinne34 / 667.949.917+12.0
16Henrik Lundqvist49 / 871.944.918+11.0
19Jonathan Quick46 / 810.943.910+9.8
20Ed Belfour†30 / 522.943.910+6.0
30Dominik Hasek†27 / 453.940.922+4.2
42Martin Brodeur†67 / 1024.935.912+3.6
43Marc-Andre Fleury67 / 1017.934.912+3.1
44Roberto Luongo67 / 1013.934.919+2.8
54Connor Hellebuyck43 / 617.930.916-0.5
60Andrei Vasilevskiy45 / 595.924.916-4.0
61Juuse Saros35 / 459.924.912-3.4
66Jake Oettinger24 / 305.921.910-3.0
67Sergei Bobrovsky64 / 799.920.912-8.9
75Patrick Roy†31 / 371.916.918-5.4
80Tristan Jarry28 / 315.911.906-6.3

* Goals vs. average, as above. † Only the part of the career from 1997-98 on is counted.

In the playoffs

We didn’t dig as deeply into the playoffs, mostly because the samples for each goalie are so small: Marc-Andre Fleury has faced the most first shots in the postseason since 1997-98, and that is fewer than 170. The overall pattern still holds, though. Starters stopped .934 of their first shots (319 goals on 4,850), against .917 on all shots, or one first-shot goal every 15 starts.

What it means for the goalie in the crease

We started with a frustration: a goal on the first shot gets treated as the one goal a goalie cannot give up, and as proof he wasn’t ready. We checked that against 67,422 first shots since 1997-98, and the numbers don’t support it.

Every goalie gives up first shots. Across the league one has gone in about once every 15 starts, and about once every 13 starts over the last three full seasons. Carey Price, tied for the top first-shot save percentage since 1997-98, still allowed 30 of them. And a first-shot goal says very little about the rest of the night: starters who gave one up stopped .904 of the shots they faced the rest of the game, and starters who stopped the first shot, .908. A goal on the first shot is not a sign that a bad night is coming.

Goalies do stop the first shot more often than the rest, but mostly because it is usually an easier shot: taken from farther out, rarely a second chance, and rarely on the power play. The fair bar for a goalie playing now is about .922, and over a few hundred first shots the gaps between goalies are mostly luck. One goal moved Devin Cooley’s first-shot save percentage 21 points.

These are NHL numbers, not minor hockey numbers, but they point the same way. The next time the first shot goes in, the coach on the bench, the parent in the stands, and the goalie in the net can treat it as what it is: one goal. Judge the goalie by the whole night and the whole season, not by one shot.

Have thoughts on first-shot goals? Members can add them in the comments below, or anyone can email us at info@ingoalmag.com.

SidebarThe numbers behind it

For readers who want the statistics: what the lines on the charts mean, how much of this is luck, and how we did it.

Reading the lines on the goalie charts

The dot is the goalie’s first-shot save percentage. The line around it is a 95% range: where his true rate most likely sits, given how many first shots he has faced. It is wide for a few hundred shots and tighter for a thousand.

Where a line crosses the dashed league mark, that goalie can’t be told apart from an average one; his numbers could simply be luck. Only a line that sits entirely clear of the dashed mark points to a real difference. Over a full career, that is true of a handful of goalies at most.

Among goalies playing now, every range in the top five, and three of the five at the bottom, includes the league’s .931. Measured against today’s .922 bar instead, Frederik Andersen sits clearly above it.

How much of it is luck?

A lot of it. If every goalie were exactly league average, chance alone would put about 5 of the 91 outside the expected range. Nine land outside it: Carey Price, Ondrej Pavelec, and Jaroslav Halak above, and Cam Talbot, Arturs Irbe, Chris Mason, Jacob Markstrom, Philipp Grubauer, and Brent Johnson below.

Split each goalie’s first shots into odd and even games, and the two halves barely agree (a correlation of 0.23). A goalie’s first-shot save percentage also only loosely tracks how he plays the rest of the night (a correlation of 0.22). The order inside the top five, or the bottom five, could easily shuffle with another season of shots.

How much luck matters depends on how many shots there are. One goalie’s few hundred first shots leave a wide range; the league’s tens of thousands leave almost none. That is why a league-wide gap in this story can be real even when it is small, while a gap between two goalies can be luck even when it looks big.

Shot quality: how much of the gap is the shots themselves

From 2018-19 on, starters saved .925 of their first shots and .904 of every other shot. If they had stopped first shots at the same rate as other shots from the same distance, with the same mix of second chances and special teams, their first-shot save percentage would have been .919. So most of that gap is the shots themselves.

The rest of the night barely changes either way. After allowing a goal on the first shot, starters saved .904 of the shots they faced for the rest of the game; after stopping it, .908. Some of that small gap is the other team being better that night, and playing with a lead.

Shot mix barely separates goalies. For the 54 goalies who faced at least 150 first shots from 2018-19 on, the save percentage their first shots should have produced, going by distance, second chances, and special teams, runs only from .916 to .933, and it tracks what they actually stopped only weakly (a correlation of 0.28). Split each goalie’s first shots over this shorter stretch into two random halves, and the halves don’t agree at all (0.02). Over a few hundred first shots, the gap between the leaders and the rest is mostly luck, not easier shots.

Expected goals: a preliminary look, for interest only

We also ran the first shots through InGoal’s expected-goals model, which weighs each shot by how often shots like it go in. The model is preliminary, these numbers are for interest only, and they may well change depending on the model used.

From 2020-21 through Oct. 27, 2026, more than six seasons, the model rates the average first shot at 0.076 expected goals, against 0.098 for every other shot. On that basis it expects first shots to be saved at .924. Starters saved .924: 1,163 goals on 15,252 first shots, against 1,166 expected. By this measure, shot quality accounts for the whole gap.

What it can’t do yet is rank goalies. Even over six seasons, no starter has faced more than 356 first shots. That would normally mean about 27 goals, and luck alone can push that up or down by about 10, about as much as the whole spread between the best and worst goalies in the data.

How we did it

  • Source: the NHL’s shot-by-shot game data, regular season 1997-98 to Oct. 27, 2026, plus playoffs. The NHL did not play in 2004-05.
  • First shot: the first shot on goal or goal a starting goalie faced in a game. The first two goalies to face a shot in a game are counted as the starters; relievers are reported separately. Empty-net goals and shootouts are left out.
  • Goalies qualify by games played (300 since 1997-98, 150 for goalies who played in 2025-26 or 2026-27); their first-shot numbers come from the games they started.
  • 33,711 regular-season games. 30 games with incomplete data (fewer than 20 shots and goals logged, many listing only the goals) were left out.
  • Checked against official box scores in a 120-game sample: 260 of 265 goalie-games matched shot for shot, and the rest were off by one.
  • Relievers: a goalie counts as coming in mid-period if the goalie he replaced had faced a shot earlier in that period, and at the start of a period if not. 451 relief appearances couldn’t be placed (the starter came back, or both starters left). The data doesn’t say why a goalie came in, so an injury replacement can’t be told apart from a goalie who was pulled.
  • The 95% ranges are Wilson score intervals.
  • Shot distance, second chances (a shot within three seconds of another attempt by the same team, including one that missed the net or was blocked), and special teams are recorded from 2018-19 on.
Save