For 19 Seasons, an NHL Hit Depended on the Rink

Recorded hits climbed for nine seasons, then gave half the gain back

His brother was close. The recount gave Hathaway 43 more hits, and overnight he went from 15th in the league to fourth. The league had recounted one season. The rinks had been counting differently at least since 2005-06, the first season the NHL’s stats site publishes hits. That season both teams together were credited with about 32 hits a game. The count climbed almost every year, to 50 in 2014-15. Then it slid, to about 43 by 2017-18. For seven seasons, 2017-18 to 2023-24, it stayed between 43 and 46, and in 2023-24, the audit year, it was about 45. Then it dropped below that band: just under 43 in 2024-25 and about 41 in 2025-26.

So how much of that rise and fall was players hitting, and how much was the person in the press box writing it down? For 19 seasons the scorer in one rink could log about twice as many hits as the scorer in another, for the same teams. That gap was bigger than the whole boom. In 2024-25 the gaps between rinks shrank almost to what luck alone would leave, and by 2025-26 they were no bigger than luck. Most of the 2024-25 drop came from the rinks that had counted most falling into line. The rise itself came in almost every rink at once, and the counts can’t say whether that was harder hitting or every scorer counting more.

The dark line is the boom. The shaded band behind it is the rinks. In 2005-06 Anaheim’s home games averaged about 17 hits and Dallas’s about 52, over a whole season against every visitor. Every season until 2024-25 the band is wider than the whole climb. In 2025-26 it ran only from 37 to 46. The flat grey line is blocks credited to players. They stayed between about 24 and 31 a game the whole time.

The league saw the problem early. In September 2002, ESPN’s Terry Frei reported that hits, blocked shots, giveaways and takeaways “have been scrubbed on each player’s summary line this season.” The reason, in Frei’s words, was that they “involved subjective judgments that could vary from building to building.” Hits were back in the NHL’s published stats by 2005-06. Every building still had its own scorer.

For 19 seasons, the rink mattered more than the decade

The lean didn’t fade after 2005-06. Take Calgary. From 2007-08 to 2022-23, the Flames’ home games logged fewer hits than their road games every season, and less than 90% of the road count in all sixteen. Calgary’s rink was the league’s stingiest eight times, more than any other. Chicago’s was the most generous six times through 2023-24, also more than any other.

Pittsburgh went the other way. Its rink counted 33% to 52% high every season from 2018-19 to 2022-23. Four of the 15 most-hit games since 2005-06 were played there. The biggest was Dec. 10, 2019: Montreal 4, Pittsburgh 1, and 120 hits, 67 of them by the Canadiens. Ben Chiarot, Jeff Petry and Jack Johnson had eight each.

Calgary’s run was no quirk of one schedule. In the odd-and-even-games check from the table, a rink that counted high in one pile counted high in the other, every season from 2005-06 to 2023-24. A rink’s lean also carried from one year to the next: Calgary’s scorer one season told you a lot about Calgary’s scorer the following season.

Luck plays a part too. A rink that happens to host a few rough nights in a 41-game home season will look generous for a year. With about 41 games each way, luck alone would push a typical rink about 6% off even in a season. The rinks actually sat about 20% off. Take the luck out and the number barely moves: a typical rink still sat about 20% above or below even, every season from 2006-07 to 2022-23. Luck’s 6% doesn’t stack on top of the habit. Lucky nights push some rinks further from even and pull others back toward it, so in a typical season luck added only about a point to the spread.

In 2014-15 the factor ran from 0.68 in Minnesota to 1.44 in Carolina, so Carolina’s scorer logged about twice what Minnesota’s did, each measured against its own team’s road games. That gap is bigger than the whole boom, which from 2005-06 to 2014-15 was a rise of about 56%.

Others found the same thing. Michael Schuckers and Brian Macdonald studied 2007-08 to 2012-13 and found New Jersey’s rink recorded about 41% fewer hits than expected, and Los Angeles’s about 30% more. And hits weren’t the only thing rink scorers shaped: our rebound goals story found that much of the flood of goals from the crease came from how scorers wrote down where goals happened.

The solid line is how far apart the rinks were. The dashed line is how far apart luck alone would push them. For most of these seasons the solid line sat about three times higher. In 2024-25 it dropped almost onto the dashed line. In 2025-26 it dipped just under it: a typical rink sat 3.9% off even, and luck alone that season would usually have put it about 4.3% off.

Luck’s line moves. It is worked out for each season, and in 2025-26 it sat lower than the usual 6%. It is also only a typical size. In some seasons the rough nights happen to spread evenly around the league, and the rinks finish a little closer together than luck would usually leave them. A season like that is what you would expect once the rinks’ habits are gone.

Then, in one summer, Toronto’s rink lost a third of its hits

The recount changed the numbers for games already played. The next season, the counts written down on game nights changed too. Toronto, the busiest rink of 2023-24, shows it most clearly. In 2023-24, Leafs home games averaged about 70 hits, the most of any rink. In 22 of the 41, the two teams combined for 70 or more. The busiest was Flyers at Toronto on Feb. 15, 2024: 103 hits, the most of any game that season. In 2024-25, Toronto’s home games averaged about 47, and one reached 70. Hits in the Leafs’ road games fell only about 10%.

Edmonton’s rink fell further, 37%. In December 2024, Brian Swane of The Hockey Writers saw the Oilers’ hitting collapse. “Mattias Ekholm is down from 4.90 to 1.86,” he wrote, counting hits for every 60 minutes Ekholm spent on the ice. “Personnel change is partially to blame.” And: “The obvious question – why is everyone hitting less? – has no clear answer.” Our numbers point to part of an answer. Ekholm was credited with 90 hits in 40 home games in 2023-24, and 24 in 35 in 2024-25. On the road, 46 in 39 became 18 in 30. Both fell. The Oilers’ road games lost about 17% of their hits too, so some of Edmonton’s drop was real. The home drop was more than twice as steep.

At the other end, San Jose’s rink, the stingiest in 2023-24, rose 15%, and Columbus’s rose 29%. The eight most generous rinks fell 20% on average. The eight stingiest rose 8%.

Each dot is a rink. Dots to the right counted high in 2023-24. Dots near the bottom lost hits in 2024-25. They run downhill together, with Toronto and Edmonton at the bottom right and Columbus at the top left.

Some of that happens every year. Scorers come and go, and a rink that counted very high one season usually counts a little less high the next. So we checked every pair of back-to-back seasons. In a normal year, a rink that counted high kept most of its lean. It came back only about a fifth of the way to even the next season, and stayed generous. The biggest one-year pull before 2024-25 came from 2005-06 to 2006-07, about half the way. From 2023-24 to 2024-25, the rinks came back about nine-tenths of the way, in one summer. Toronto, for one, went from 46% high to about 9%. To come back that far before, the rinks had needed nine whole seasons, from 2005-06 to 2014-15 taken together.

That pull explains most of the first year’s drop. In 2023-24 the league averaged 45.5 hits a game. A few rinks, Toronto at about 70 a home game and Boston at about 62, logged far more than the rest and dragged that average up. The middle rink, with as many rinks above it as below, logged only 43.8. So about 1.7 hits a game of the league average came from the rinks at the top. In 2024-25 league hits fell about 2.6 a game. If the generous rinks simply came down to the middle, that accounts for about 1.7 of the 2.6, roughly two-thirds. The rest of that drop came from somewhere else, and it could be real. League hits fell another 5% in 2025-26, after the rinks had lined up, and that could be real too.

We found no announcement from the NHL explaining why the rinks lined up in 2024-25. The first season with the rinks’ totals lined up also produced a new hit record.

Home scorers may once have given their own team a little extra

The scorer decided more than how many hits went on the sheet. He also decided whose they were. If a home scorer leaned toward his own team, the home side’s share of the hits would show it. In 2005-06, home teams were credited with 53.6% of all recorded hits. That share drifted down to 51.5% by 2014-15 and about 51% after. In 2024-25 and 2025-26, after the rinks lined up, it was 50.9% and 51.1%. If those two seasons show the real home edge in hitting, about two to three points of the early home share came from the home scorer leaning toward his own team. If that was a lean, it was gone by 2015-16. That was nine seasons before the rinks lined up. In those nine seasons the rinks still disagreed widely about how many hits there were. But the home team’s share of them was already down to about 51%, the same as it has been since.

The line slides from about 54% to just above the even-split mark and stays there.

The rinks’ totals have now lined up for two full seasons, so this season is a fresh check on whether a home lean creeps back. There is no sign of one yet. Our play-by-play, the NHL’s game-by-game log of every whistle, shot and hit, covers the first 184 games of 2026-27, and home teams have 49.1% of the hits. Over the same first 184 games, the home share ran from 49.5% to 52.7% every season from 2015-16 to 2025-26, so this year’s start sits a shade below that. It is too early to mean much. A season’s first 184 games have missed its full-season home share by as much as 1.5 points.

The hit count itself is running at 44.1 a game, close to the 43.2 of the first 184 games of 2025-26. That is just as shaky. In the last five seasons, the first 184 games have run as much as 5% below the full season’s rate (2023-24) and as much as 8% above it (2022-23).

The same summer, recorded giveaways doubled

If 2024-25 changed how rinks counted, hits shouldn’t be the only count that jumped. They weren’t. A giveaway is charged when a player hands the puck to the other team with nobody forcing him, say a blind pass up the middle. A takeaway goes to the player who strips the puck off an opponent. Recorded giveaways doubled, from 14.6 a game to 29.4, and takeaways fell from 13.9 to 9.6. Players are very unlikely to have started losing the puck twice as often in one summer. A giveaway written down in 2023-24 and one written down in 2024-25 almost certainly weren’t counted the same way, as our rush story also found.

Both lines wander for years. Then, in the same season the rinks lined up, the giveaway line shoots up and the takeaway line drops.

Blocked shots are a check on the hits story. The same scorers write them down, so if the hits boom had been scorers logging more of everything, blocks would have boomed too. They didn’t. Blocks credited to a player went from about 26 a game in 2005-06 to a high of about 31 in 2023-24 and back to about 28. About one shot attempt in four was blocked in 2009-10 (24.5%), and about one in four in 2025-26 (24.9%). If blocking boomed, it was before the NHL’s published blocks begin in 2005-06.

Even that 2023-24 high may be partly the scoresheet. Credited blocks rose about 2 a game that season, the same season recorded missed shots jumped.

There is a second count of blocked shots: the play-by-play log. Until 2023-24 it matched the credited blocks, about 29 a game in both in 2022-23. Then the log jumped to about 34 a game, while credited blocks reached only about 31. That is the jump our rebound goals story noticed. More than half of it, about 2.6 a game, was a new kind of entry. From 2023-24 the log also marks a shot that hits the shooter’s own teammate, and nobody is credited with blocking those.

What the boom was made of

Back to the question, then: how much of the boom was players hitting, and how much was the person writing it down? The curve has three stretches, the climb and two slides, and each gets a different answer.

Start with the climb. From 2005-06 to 2014-15, hits per home game rose in 26 of 28 rinks, and the middle rink went from about 30 hits a game to 50. When nearly every rink rises at once, it is either harder hitting or every scorer counting more, and comparing rinks with each other can’t tell which. The stingiest rinks did rise fastest. The eight lowest climbed 127%, the eight highest 25%. So part of the climb was scorers coming closer together in how they counted.

Altogether hits fell about 18% from the 2014-15 peak, in two slides with a flat stretch between. The first slide ran from 2014-15 to 2017-18, about 14%. Hits per home game fell in 22 of 30 rinks, and the generous ones fell most: the top eight −23.3%, the bottom eight −3.3%. Players may have hit less, or the busiest scorers may have been reined in. The counts can’t tell those apart.

Then hits held between 43 and 46 a game for seven seasons and stood at 45.5 in 2023-24. The second slide, to 40.9, began in 2024-25, and most of its first year traces back to the scorers. The rinks that had counted most lost most, and they account for about two-thirds of that first year’s drop. The further 5% drop in 2025-26 came after the rinks lined up and could be real.

And all along, the rinks sat further apart than the boom ever moved. In every one of the 19 seasons from 2005-06 to 2023-24, the gap between the busiest and the quietest rink was bigger than the whole climb. In 2024-25, for the first time, it was smaller.

What it means for reading hit totals

For anyone reading a hit total, all of this comes down to one rule: compare hits within a season, not across eras. Inside one season, a hit total is worth a fair amount. A 250-hit season in 2014-15 and one in 2025-26 were counted on different scales. Before 2024-25, a player’s home rink could move his total by tens of percent; our arena-adjusted hits take that out.

Giveaways and takeaways can’t be compared across 2024-25 at all. What got counted changed overnight.

Rink scorers even shape the test of whether hitting wins. Our playoff speed and hitting test, in Prop Lab, got around them by rating each team on its regular-season road games. Those spread a team across every other rink’s scorer, so no single home scorer tilts them. At 5-on-5, with both teams at full strength, playoff teams log about 71% more hits than in their own regular-season road games. Yet the team with more regular-season road hits won 47.1% of 433 playoff games from 2021-22 to 2025-26, a little under half. The more physical team didn’t win more often. Hits partly track time without the puck, since a team chasing the play has more chances to hit someone than a team carrying it.

Hathaway’s hits came home and away alike

Hathaway’s own count stayed level home and away. In 2023-24, the audit year, he was credited with 165 hits in 41 home games and 161 in 41 on the road. In 2024-25 it was 119 in 32 and 138 in 35. That same season, with the rinks’ counts lined up, Sherwood was credited with 462, more of them on the road than at home.

When the league recounted his season, Hathaway took it as a fix. “It seemed some games were not accurate, and I talked to a few guys around the league too, who thought the same,” he said. Going back through every game, he said, “definitely makes for a lot of work, but a more accurate job.”

Show the work

Tests written down

Not in advance (exploratory)

Results computed

October 27, 2026

League totals for every regular season from 2005-06 to 2025-26 from the NHL stats API, plus each game’s home and road rows (25,302 games). Hit factor = both-team hits per home game ÷ per road game, per rink and season. Spread = standard deviation of the log factors across rinks; chance = the spread game-to-game variation alone would give (delta method per rink, averaged); beyond chance = the square root of the difference of their squares. Split-half check: factors from odd against even game ids. Year-to-year: log factors in consecutive seasons. Rink change between seasons: least-squares slope of the log change in a rink’s home-game hits on its old log factor (−1 = every rink’s lean wiped out; near 0 = rinks kept their habits). The audit split cuts each of 2021-22 to 2025-26 at Feb. 8 (2023-24 is the audit season; the others are controls). Game scenes, extremes and rink habits come from the same game rows. Player home and road splits: our boxscore table from 2017-18, and NHL stats skater game rows before that. Play-by-play cross-check from 2017-18, with 2026-27 to date as an out-of-sample look. The 6,826 added hits are ESPN’s count; the games they are spread over are ours. Block share = credited blocks ÷ shot attempts, leaving out shots blocked by a teammate, which the NHL’s attempt totals include from 2023-24 and which are nobody’s credited block.

Continues below

Frequently asked questions

More in The Vanishing Game

• Verdict: Data story

  Hainsworth’s 22 Was Mostly Arithmetic

• Verdict: Data story

  The Slapshot Went From One NHL Goal in Four to One in Twelve

• Verdict: Data story

  Did Garbage Goals Take Over the NHL?

• Verdict: Data story

  The Last Goalie Who Played Every Night

• Verdict: Data story

  Is the Rush Dying in the NHL?

• 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

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.
• Arena scorer bias →How the home rink’s scorer changes counted stats like hits.
• Does hitting win in the playoffs? →Everyone hits more in the playoffs, and the hardest-hitting team does not win more.
• Did Garbage Goals Take Over the NHL? →Rebounds, scrambles and point shots since 2017-18, and what rink scorers changed.
• Does an NHL Game Really Stop Less Than It Used To? →Whistles, faceoffs and the longest stretches without a stoppage.
• Is the Rush Dying in the NHL? →Rush goals and chances by where the possession began, 2017-18 to 2025-26.

Sources and method

1. NHL stats (NHL.com), every season since 1917-18
2. NHL stats API: team realtime totals (hits, blocks, giveaways, takeaways), season and game by game
3. NHL stats API: skater realtime, game by game and season leaders (Boucher, Brown, Clutterbuck, Martin; hit leaders since 2005-06)
4. NHL event summary: Vancouver at Anaheim, Nov. 20, 2005 (zero hits)
5. NHL event summary: Montreal at Pittsburgh, Dec. 10, 2019 (120 hits)
6. The Philadelphia Inquirer, Jackie Spiegel: Garnet Hathaway and the NHL hits audit (Feb. 8, 2024)
7. RMNB, Ian Oland: the NHL’s hits audit statement (Feb. 8, 2024)
8. ESPN, Sean Allen: the NHL hits audit, by player and rink (Feb. 13, 2024)
9. The Hockey Writers, Brian Swane: Oilers hit totals declined dramatically in 2024-25 (Dec. 28, 2024)
10. Daily Faceoff, Tyler Kuehl: Kiefer Sherwood sets the NHL single-season hits record (March 22, 2025)
11. NHL game feed (example 2023-24 game with a teammate-blocked shot)
12. Schuckers and Macdonald: Accounting for Rink Effects in the NHL’s Real Time Scoring System
13. Terry Frei, ESPN: NHL taking the crunch out of numbers (2002)
14. NHL API reference (community documentation of the stats endpoints)

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