Does an NHL Game Really Stop Less Than It Used To?

Thirteen minutes without a whistle

That stretch of play, from the faceoff after Chychrun’s goal to Wedgewood’s cover on Brodie, is the longest in our nine seasons of play-by-play, the NHL’s log of every faceoff, shot and whistle with its time on the clock. That log covers 11,052 regular-season games since 2017-18. Thirteen minutes without a whistle almost never happens. To find out whether an NHL game really stops less than it used to, start with the faceoffs.

After Wedgewood covered Brodie’s shot, play restarted with a faceoff, as it does after every whistle. So the faceoff count is a whistle count. In 2025-26 a game had 56.0 faceoffs, overtime included. Take away the draw that opens each period and about 53 were restarts after a whistle. Half the stretches of play between them were shorter than 47 seconds.

So a 2025-26 game had about 53 whistles in 60 minutes. Most stretches were short: half ended inside 47 seconds. A few ran long, about 15 in 100 lasting two minutes or more, and runs like Toronto’s pull the average up to a little over a minute (65.0 seconds).

The clock stops at every whistle, so the time from one faceoff to the next is exactly the time the puck was live. In Toronto that night it was 13:22. The longest run of 2025-26, 12:10, came in Columbus in April. The game clock can’t see the dead time in between: TV timeouts, video reviews, the wait before each faceoff. That dead time is why Toronto’s first period still took 29 minutes on the wall clock. Every stretch figure here covers the first three periods of every regular-season game since 2017-18, when our play-by-play starts. Overtime is left out.

The short answer is yes against the late 1990s, and about even against 2009-10. In between, the count fell after the 2004-05 lockout, climbed back by the mid-2010s and fell again from 2017-18. The whistles that went away most recently were offsides and goalies smothering the puck. And a handful of goalies stop the game far more often than the rest.

A 1997-98 game had about ten more whistles

The NHL’s team faceoff tables go back to 1997-98, so that is as far back as this story can reach. That first season tops them. A game had 65.9 faceoffs. Every season through 2003-04 had between 64.7 and 65.9.

Even the quietest team that season stopped more than nearly every team now. The Boston Bruins’ games had 59.7 faceoffs, the fewest in the league. In 2025-26, 31 of the 32 teams had fewer, and only Carolina’s games, at 60.7, had more. The busiest games of 1997-98 were Chicago’s, at 70.2.

Follow the top line. It runs flat near 65 until the lockout, drops to about 56, climbs past 61 and comes back down to 56.

The drop after the lockout came all at even strength: 8.8 fewer even-strength faceoffs a game from 2003-04 to 2005-06. Faceoffs on the power play went up, from 15.1 to 19.8, because referees called far more penalties that season. The rule book changed, too. The red line no longer stopped two-line passes. Tag-up offside came back: an attacker caught inside the blue line could skate back out to clear himself, and play went on instead of stopping. And a team that iced the puck could no longer change lines. The faceoff tables can’t say which whistles vanished. The NHL’s log of why each whistle blew starts, in our sample, in 2009-10.

Over the whole span, restarts after a whistle fell by about 9.9 a game. Power plays explain part of it. Each one starts with a whistle, and a game had 3.5 fewer power-play chances, which leaves 6.4. Goals went the other way. A game had 0.9 more of them, and each goal brings its own faceoff. Those extra goal faceoffs hid part of the drop. Add the 0.9 back to the 6.4, and the ordinary whistles (icings, offsides, frozen pucks and pucks out of play) fell by about 7.3 a game.

Three things blur that split. A fight or a pair of offsetting minors stops play without a power play, so if those got rarer, that drop lands in the ordinary whistles too. A goal in overtime ends the game, so no faceoff follows it. And some goals come on a delayed call: the referee’s arm is up and the other team scores. If the call was a minor, the goal wipes it out. No power play starts, and nothing is counted twice. A major still goes on the clock, though. Then a single faceoff follows both the goal and the start of the power play, and the split subtracts that one faceoff twice, once as a goal and once as a power play.

Today’s game stops about as often as a 2009-10 game

Pick a different year and the answer changes, because the count didn’t slide in a straight line. By 2007-08 it had dropped to 56.5. Then it climbed back, to 61.3 in 2014-15, before falling again from 2017-18. Since 2020-21 it has sat between 55.8 and 56.9.

Put 2009-10 next to today and almost nothing is left to explain: 56.8 faceoffs a game then, 56.0 in 2025-26.

So does an NHL game really stop less than it used to? Yes, compared with the late 1990s, by about ten whistles a game. Yes, compared with the mid-2010s, when a game had more than 60 faceoffs. And about the same, compared with 2009-10.

The two lines lie almost on top of each other, so a game today has the same mix of short and long runs as one from those years. These lines come from a sample of games we pulled from the NHL’s own log (more on it below), and it agrees on the averages too: stretches ran 63.6 seconds in 2009-12 and 64.8 in 2023-26.

The mix inside those games changed. Today’s games have more icings and fewer offsides and goalie freezes than the 2009-13 games did. The totals still came out about even: 43.6 stoppages a game in the 2009-13 sample games, 43.2 in 2021-26. Those totals count only the whistles the log tags with a reason, like an icing or an offside. Goals and penalty calls are filed separately, and so is the draw that opens each period, though a faceoff follows every one of them. That is why both totals sit well below the 56 faceoffs a game.

The Devils played the quietest hockey in the league

In 2009-10 the league as a whole stopped about as often as it does now. Team by team, though, some games stopped far less than others. The quietest teams are worth a look, because they hint at what makes a game stop.

The quietest belonged to the New Jersey Devils. In 2009-10 their games had 51.8 faceoffs, against a league average of 56.8. No team’s games have stopped less in any season since 1997-98. The Devils had the fewest in the league in eight of the twelve seasons from 2002-03 to 2014-15, and the second-fewest in two more.

Where the Devils’ line touches the dashed one, they were the quietest team that season. It happens again and again from 2002-03 to 2014-15.

Their games also had fewer shots. In 2009-10 the Devils and their opponents put 57.0 shots a game on goal, against 60.6 across the league. The most common whistle in hockey is a goalie covering a shot, like Wedgewood on Brodie, so fewer shots leave fewer pucks to cover. Teams whose games had fewer shots usually had fewer whistles, in 27 of the 28 seasons. But shots explain only a small part of why one team’s games stop more than another’s, and we didn’t test whether they explain the Devils.

After the Devils, the quietest games moved around the league, and the next quiet team brings the goalies into the story. The Islanders had the fewest three seasons running, from 2018-19 to 2020-21, all three with Barry Trotz as coach. Two of their goalies from those years, Semyon Varlamov and Ilya Sorokin, are among the four goalies least likely to stop play, as we show below. We haven’t tested whether they are why the Islanders were quiet. In 2025-26 the Dallas Stars had the fewest, at 53.2. Carolina had the most, at 60.7, nearly what the whole league averaged in 2014-15. The Hurricanes also led in 2019-20, 2020-21 and 2023-24.

Offsides and goalie freezes took the most whistles out

So which whistles went? Our play-by-play says how each stretch ended: with a goal, a penalty call, the period horn or some other whistle. From 2017-18 to 2025-26 a game went from 60.2 stretches to 55.4. Goals added 0.30 a game. Penalty calls fell by 0.31. The other whistles fell by 4.7.

Only the first pair of bars moves much. Penalties, goals and period ends stayed where they were.

“Other whistle” covers a lot. Our database doesn’t keep the reason for each one, but the NHL’s own log does. It tags each whistle “icing”, “offside”, “goalie stopped after a shot on goal” and so on. So we pulled 408 games, 24 spread evenly through each season from 2009-10 to 2025-26, and grouped the seasons into four blocks of four or five years each: 2009-13, 2013-17, 2017-21 and 2021-26. In today’s game, goalies covering the puck are the biggest single reason play stops, about 16 a game. Pucks out of play add about 9, icings 8 and offsides 4.

Watch the offside bars shrink after 2013-17, and the icing bars step up after 2013-14.

Offside became a fixation in the 2010s. In February 2013 Matt Duchene scored for Colorado against Nashville off a long stretch pass. He had crossed the blue line ahead of the puck. “I was 100% offside,” he said after the game. Nashville’s coach, Barry Trotz, said, “The league already verified that it should have been an offside.” Colorado won 6-5. The coach’s challenge for offside arrived in 2015-16. In 2019-20 coaches challenged for offside 75 times, and 56 of those goals came off the board, the officiating website Scouting the Refs reported. For 2020-21 the league eased the rule. Before, an attacker’s trailing skate had to be touching the blue line when the puck crossed it. Now the skate could be lifted in the air, as long as it was still above the line and not all the way past it into the zone.

In our sample, offsides fell from 6.3 a game in 2013-17 to 4.2 in 2021-26. The drop is too big to be luck, and it showed up in both halves of the sample. It also started before the 2020-21 rule: offsides were already down to 4.9 in 2017-21. That block includes 2020-21 itself, but its three seasons before the new rule were about as low. Whatever started the fall came before the new rule, and the sample can’t say what it was.

Goalie freezes fell from 17.7 a game to 15.6. Fewer shots don’t explain it. Shots on goal fell late in that window, but a game in either block averaged about 60 shots. Per 100 shots on goal, freezes went from about 30 to 26.

Icings went the other way, from 7.0 a game in 2009-13 to 8.3 in 2013-17. In 2013-14 the NHL brought in hybrid icing. The linesman now blows the play dead as soon as he judges the defender will win the race to the puck, instead of waiting for him to touch it. The league also told linesmen to stop waving off icing on a pass a teammate could have reached. That second change should mean more icings called. The rise lines up with the new rules. The sample can’t show that the rules caused it.

The coach’s challenge barely shows up as a whistle of its own. Video reviews and challenges coded as their own stoppage went from 0.11 a game in 2013-17 to 0.26 in 2021-26. Even those come at a stop that already happened. The NHL’s report logs Keefe’s challenge as its own stoppage, but play had already stopped for Chychrun’s goal, so the challenge made that break longer without adding a whistle. TV timeouts work the same way. The log usually lists one as a second reason on a whistle that was already coming, like the timeout that followed Wedgewood’s cover on Brodie, so they rarely add a whistle of their own. You can watch the log keep pace with the rule book. Teams lost the right to call a timeout after icing the puck in 2017-18, and icings followed by a team timeout fell from 0.29 a game in 2009-13 to 0.03 in 2021-26.

Some goalies stop the game far more than others

Goalie covers were one of the two whistles that went away, so the men in net are part of why a game stops less. They also cause the most whistles, because a goalie freezing the puck is the biggest single reason play stops. Do all goalies freeze alike? The sample is too small to tell, but our play-by-play has every save. It logs every shot on goal with the goalie who faced it, and every whistle with its time. So for every save since 2017-18 we checked whether a whistle came within three seconds. Wedgewood’s stop on Brodie counts. A save that a defenceman clears up the ice does not.

We count each of those quick whistles as a goalie cover. The log can’t show the glove closing on the puck, only a save and then a whistle, so part of what this count catches may belong to the team in front of him or to the rink’s scorer, who logs the times. Both come up below. First we checked the count against the NHL’s own tags in the 216 sampled games since 2017-18. The tags list 3,407 goalie freezes. Our three-second count found 3,511, about 3 too many in every 100.

The full count agrees that goalie covers fell. Across the league, a whistle followed about 32 to 33 saves in 100 every season from 2017-18 to 2022-23. From 2023-24 it was about 30.

The line steps down once, in 2023-24, and stays down: about 2.4 fewer whistles per 100 saves. Part of that step is the log, not the goalies. Through 2022-23 a save was most often followed by its whistle one second later. From 2023-24 it was two seconds, so some covers now land just past our three-second cut. Count every whistle within ten seconds of a save instead, and the step shrinks to about one fewer per 100 saves. Our sample of the NHL’s tags sees freezes falling from 2013-17 to 2021-26, which agrees that some of the fall is real. Between its last two blocks of seasons it looks nearly flat. That last block, 2021-26, starts two seasons before the 2023-24 step, and 24 games a season is too few to catch a change this small.

Goalie by goalie, the spread is wide. Among the 46 goalies with 5,000 or more saves, Igor Shesterkin has the lowest share: a whistle follows 27.1% of his saves. James Reimer has the highest, at 35.5%. Shesterkin makes about 27 saves a game. If Reimer made those same saves, about two more of them would end in a whistle each night.

The two goalies from that night in Toronto sit near the two ends. Each figure on the chart covers all of a goalie’s saves from 2017-18 to 2025-26, for every team he played for. Petr Mrazek is second from the bottom, at 27.3%. Wedgewood is at 35.4%, just short of 5,000 saves. With enough saves to qualify he would sit second from the top, a hair behind Reimer’s 35.5%.

Those are straight shares of every save, and the 2023-24 step flatters a goalie who made more of his saves since then. Measure each goalie against the league in his own seasons and the ends shuffle a little. Semyon Varlamov and Mrazek edge just below Shesterkin. Jeremy Swayman edges just above Reimer, and Wedgewood tops them all. Shesterkin and Reimer still sit about eight points apart.

Each goalie’s number stays put from year to year. Split his seasons into two piles, odd years and even years, and the two piles give nearly the same number for each goalie. His home games and his road games agree, too.

The whistle habit mostly moves with the goalie

So who is in net changes how often a game stops, if the habit really is the goalie’s. It could still belong to the team in front of him, or to the scorer in his home rink who logs the times. Goalies who changed teams give a way to check.

Take Thomas Greiss. In his Islanders seasons, a whistle followed 30.9% of his saves and 33.1% of his partners’. He moved to Detroit, where the other goalies were at 34.6%. In his Detroit seasons Greiss was at 31.0%. His new partners stopped play more. He didn’t. Both stints came before 2023-24, so the league-wide drop doesn’t touch his numbers.

We found 28 goalies who made 1,500 or more saves for each of two teams. One thing could fool us. In 2023-24 the whole league’s rate dropped, so a goalie whose new team came after the drop could look as if he had changed when the league had. Spencer Knight shows how. A whistle followed 30.9% of his saves in Florida and 28.7% in Chicago, close to his new partners’ 27.8%. But his Chicago seasons all came after the drop. So we compared each goalie with the league average in the same seasons. Against that average, Knight went from 0.9 points below in Florida to 1.4 below in Chicago. His Chicago partners were 2.3 below, so he stayed closer to his old self.

We settled on that adjustment only after seeing the count without it, so here are both. With it, a goalie’s number at his new team stayed closer to his own old number than to his new teammates’ in 18 of the 28 moves. Without it the count was 14 of 28, an even split, because the league-wide drop pulled down the later seasons of goalie and teammates alike, as it did for Knight.

With or without the adjustment, a goalie who stopped play often at his old team mostly kept doing it at his new one. His new team’s habit barely predicted his once the league drop was taken out, and only loosely before. Across all 28 moves the pull was small, too. When a goalie’s new partners were whistled 10 more times per 100 saves than his old ones, his own rate rose by about half a whistle per 100 on average, or about one before the adjustment.

Mrazek is the plain case before any adjustment. In his Carolina seasons, a whistle followed 33.8% of his partners’ saves and 27.4% of his own. He moved to Chicago, where the other goalies were at 28.6%. In his Chicago seasons Mrazek was at 27.2%. His partners’ rate fell about five points, and his own hardly budged. (The 27.3% above covers all his saves, for every team.) Against the league average, though, his Chicago seasons sat a little closer to his partners than to his old self. He is one of the ten goalies who moved that way.

The dots climb from lower left to upper right. Goalies who stopped play often before a move kept doing it after. Not every case fits. Jonathan Quick went from 31.9% in Los Angeles to 26.9% with the Rangers, right where Shesterkin sits. That is only 28 moves, so a smaller team effect can’t be ruled out.

Each arena has its own scorer logging the times, and that shows up a little. Visiting goalies in Seattle and Boston were credited with these whistles about two points more often than in their other road games. In Colorado and Pittsburgh it was about one and a half points less. Some of that is luck. With the number of saves each rink sees, chance alone would push a typical rink about half a point high or low. The 32 rinks spread about twice that wide, and Seattle and Boston sit further out than chance alone usually reaches. So the scorer, or the rink, plays a part. But that is a point or two, against the eight points between Shesterkin and Reimer. On the road, where other arenas’ scorers keep the log, Reimer still had the highest rate of the regulars and Shesterkin the fourth lowest.

Maybe one goalie gloves a puck that another would steer to the corner. Maybe his defencemen leave him fewer loose pucks to play. The log can’t tell us which, or how much of the habit is really his defence’s.

The 2023-24 drop is a different kind of change. It hit the whole league at once, goalies and their teammates alike. By our three-second count it took about 2.4 whistles out of every 100 saves, and about one once the later whistle times are allowed for. The log doesn’t say why that last whistle went. Those are goalie covers that a game today no longer has, and with the lost offsides they are the whistles that went away most recently.

What fewer whistles mean on a given night

Here is what all that means for the game on your TV tonight.

• A 2025-26 game had about ten fewer whistles than one in 1997-98, and about as many as one in 2009-10. It is the second-lowest count of the 28 seasons the NHL has tables for; 2020-21 was lower.
• Half the stretches of play now last under 47 seconds, up 4 seconds from 43 in 2017-18, and about 15 stretches in 100 run two minutes or more.
• How often a night stops depends partly on who is in net. Facing the same saves, Shesterkin and Reimer would differ by about two whistles a game.

And now and then, a game still runs on and on without a whistle.

The longest run of 2025-26 ended at the horn

On April 4, 2026, Winnipeg played in Columbus. At 7:50 of the first period a linesman whistled an offside. After the faceoff, nobody blew a whistle for 12 minutes and 10 seconds, the longest run of 2025-26. In real time the first period took 26 minutes. The game had 33 stretches in its first three periods, tied with Ottawa at Vancouver on April 19, 2022 for the fewest of any game since 2017-18. Kyle Connor scored both Winnipeg goals in a 2-1 win.

Toronto’s run, the longest of the era, had ended with the most common whistle in hockey, a goalie covering a shot. Wedgewood made that cover, and a whistle follows 35.4% of his saves. In the other net that night was Mrazek, at 27.3%.

In Columbus, in those last twelve minutes of the first period, Connor Hellebuyck and Jet Greaves faced five shots between them. Neither goalie covered the puck once. The horn ended it.

Show the work

Tests written down

Not in advance (exploratory)

Results computed

October 26, 2026

Faceoffs a game are the NHL’s team faceoff totals divided by team games, every regular season 1997-98 to 2025-26; team-seasons use the same tables, with shots for and against from the NHL’s team summary. Stretches of play are every regulation stretch in 11,052 regular-season games from 2017-18 to 2025-26 (627,218 stretches), from faceoff to the next faceoff or period end, ordered by clock time: in 2017-18 and 2018-19 the log stores some events out of time order, so every game is sorted by the clock first. That leaves 121 of the 627,218 stretches with no whistle logged at the end; they still count, running to the next faceoff. Stoppage reasons come from the NHL’s play-by-play for 408 games (24 a season, evenly spaced, 2009-10 to 2025-26), compared by era with standard errors (SE: the typical gap luck alone opens between two samples this size) and an even/odd split of the sample. The goalie measure (added 2026-10-27) counts, for every save in regulation since 2017-18, whether a stoppage came 0 to 3 seconds later on the same period clock. The mover check takes each goalie’s first and last team with 1,500+ saves, measures every rate in points above the league rate in the same seasons, and compares his later rate with his earlier one and with his later teammates’ rate (other goalies with 1,000+ saves on that team in the same seasons).

Continues below

Frequently asked questions

More in Rules

• Verdict: Data story

  The Overtime Faceoff’s 30-Second Head Start

• Verdict: Data story

  Spearing Hit Zero. Hooking Never Went Away.

• Verdict: Data story

  A Shootout Shooter Needs More Than 200 Tries Before Skill Outweighs Luck. The Most Anyone Has Taken Is 135.

• Verdict: Data story

  How Much Do NHL Referees Change the Power-Play Count? The Busiest Crews Call About 14% More Than the Quietest

• Verdict: Data story

  NHL Scoring Doubled in 1929-30 Under the New Forward-Pass Rules

• Verdict: Data story

  NHL Overtime Has Risen From 18% of Games to 23% Since 1999

All How Hockey Changed research →

Keep reading

• How Hockey Changed: NHL data stories →One chart-led story a day about how the NHL game, its players and its rules have changed.
• The NHL shift got shorter, then grew back →Average shift length since 1997-98, and how much fewer faceoffs explain.
• The power play is disappearing →Power-play chances halved since 2005-06.
• Shots on goal fell. Shooting didn’t. →Fewer shots on goal since 2017-18, while scoring rose.
• NHL goalie hub →Starters, save percentage and workload for every goalie.
• NHL referees →Tonight’s officials and how many penalties each crew calls.

Sources and method

1. NHL stats (NHL.com), every season since 1917-18
2. NHL stats API: team faceoffs by season, 1997-98 on
3. NHL stats API: team summary (shots for and against), by season
4. NHL API: game play-by-play with stoppage reasons (sampled games)
5. NHL official play-by-play report: Arizona at Toronto, March 10, 2022
6. NHL official play-by-play report: Winnipeg at Columbus, April 4, 2026
7. Pension Plan Puppets, Omar White: Maple Leafs spot the Coyotes a big lead, and then overtime happened (March 10, 2022)
8. Scouting the Refs, Josh Smith: Avs’ Duchene scores epic offside goal (May 7, 2013)
9. Scouting the Refs, Josh Smith: NHL modifies offside rule to extend blue line to vertical plane (Dec. 22, 2020)
10. Hockey-Reference: Barry Trotz coaching record
11. NHL Official Rules 2021-22, Rule 15.2: a goal on a delayed minor wipes out the minor; majors still go on
12. NHL API reference (community documentation of the stats endpoints)
13. Scouting the Refs: NHL historical rule changes

Research and analysis, not betting advice. Past results do not guarantee future returns; bet only what you can afford to lose.