How Hockey ChangedStandings and results
Home Ice Lost Almost Half Its Edge in 1992. The Power Plays Didn’t Budge.
In one season, 1992-93, NHL home teams lost about 43% of their edge in the standings and more than half their goal margin, and neither came back. All 17 franchises that played through both eras lost ground at home. They kept getting the same extra power plays. Those nearly vanished only once, in the empty arenas of 2020-21, and home teams kept winning anyway.
By Edgehalla ResearchPublished 9 min read
In early February 2021, three weeks into a season played mostly in empty arenas, Brian Swane of The Hockey Writers added up the home teams’ record: 91-41-17. He had planned to write that quiet buildings and lighter travel would take their edge away. “So it seemed like a pretty sound theory,” he wrote. “And in just three weeks, it’s been blown to bits.”
The hot start faded. Home teams played .577 hockey the rest of the way. That same winter, the extra power plays home teams had drawn in every season on record nearly vanished.
That winter put an old suspect on trial. Nearly three decades earlier, in 1992-93, home ice had shrunk in a single season. The referees were one obvious place to look.
On this page
Home teams lost almost half their edge in one season
For most of the NHL’s history, home ice was worth a lot. From the Original Six through the first expansion era, home teams took about 60% of the points on offer, and they still took 59% from 1977-78 to 1991-92. Over the next dozen seasons, they took 55%. An even split would be 50%, so home teams went from about 9 points above even to about 5. The goal margin lost more than half. Home teams had outscored visitors by about 6 goals every ten games. Now it was about 3.
Home teams’ share of points fell in 1992-93; their share of power plays didn’t
Every NHL regular season since 1917-18; power-play chances are recorded from 1977-78
Show data
| Season | Home share of points | Home share of power plays |
|---|---|---|
| 1917-18 | 69% | – |
| 1918-19 | 74% | – |
| 1919-20 | 65% | – |
| 1920-21 | 63% | – |
| 1921-22 | 67% | – |
| 1922-23 | 75% | – |
| 1923-24 | 71% | – |
| 1924-25 | 57% | – |
| 1925-26 | 60% | – |
| 1926-27 | 55% | – |
| 1927-28 | 52% | – |
| 1928-29 | 56% | – |
| 1929-30 | 59% | – |
| 1930-31 | 59% | – |
| 1931-32 | 67% | – |
| 1932-33 | 67% | – |
| 1933-34 | 61% | – |
| 1934-35 | 54% | – |
| 1935-36 | 60% | – |
| 1936-37 | 55% | – |
| 1937-38 | 58% | – |
| 1938-39 | 58% | – |
| 1939-40 | 63% | – |
| 1940-41 | 61% | – |
| 1941-42 | 64% | – |
| 1942-43 | 67% | – |
| 1943-44 | 63% | – |
| 1944-45 | 59% | – |
| 1945-46 | 62% | – |
| 1946-47 | 58% | – |
| 1947-48 | 56% | – |
| 1948-49 | 61% | – |
| 1949-50 | 58% | – |
| 1950-51 | 58% | – |
| 1951-52 | 58% | – |
| 1952-53 | 58% | – |
| 1953-54 | 65% | – |
| 1954-55 | 58% | – |
| 1955-56 | 62% | – |
| 1956-57 | 60% | – |
| 1957-58 | 53% | – |
| 1958-59 | 55% | – |
| 1959-60 | 61% | – |
| 1960-61 | 61% | – |
| 1961-62 | 62% | – |
| 1962-63 | 55% | – |
| 1963-64 | 65% | – |
| 1964-65 | 55% | – |
| 1965-66 | 60% | – |
| 1966-67 | 61% | – |
| 1967-68 | 61% | – |
| 1968-69 | 61% | – |
| 1969-70 | 59% | – |
| 1970-71 | 64% | – |
| 1971-72 | 61% | – |
| 1972-73 | 64% | – |
| 1973-74 | 60% | – |
| 1974-75 | 62% | – |
| 1975-76 | 59% | – |
| 1976-77 | 59% | – |
| 1977-78 | 59% | 52% |
| 1978-79 | 58% | 53% |
| 1979-80 | 59% | 53% |
| 1980-81 | 59% | 52% |
| 1981-82 | 61% | 51% |
| 1982-83 | 61% | 52% |
| 1983-84 | 58% | 52% |
| 1984-85 | 57% | 52% |
| 1985-86 | 58% | 52% |
| 1986-87 | 60% | 52% |
| 1987-88 | 57% | 51% |
| 1988-89 | 58% | 52% |
| 1989-90 | 58% | 53% |
| 1990-91 | 59% | 52% |
| 1991-92 | 61% | 52% |
| 1992-93 | 56% | 52% |
| 1993-94 | 54% | 53% |
| 1994-95 | 58% | 52% |
| 1995-96 | 56% | 51% |
| 1996-97 | 55% | 51% |
| 1997-98 | 54% | 52% |
| 1998-99 | 54% | 51% |
| 1999-00 | 55% | 51% |
| 2000-01 | 55% | 52% |
| 2001-02 | 55% | 52% |
| 2002-03 | 55% | 53% |
| 2003-04 | 55% | 52% |
| 2004-05 | – | – |
| 2005-06 | 57% | 51% |
| 2006-07 | 55% | 52% |
| 2007-08 | 54% | 52% |
| 2008-09 | 56% | 53% |
| 2009-10 | 56% | 52% |
| 2010-11 | 52% | 52% |
| 2011-12 | 56% | 52% |
| 2012-13 | 57% | 52% |
| 2013-14 | 54% | 52% |
| 2014-15 | 54% | 52% |
| 2015-16 | 53% | 53% |
| 2016-17 | 56% | 53% |
| 2017-18 | 56% | 53% |
| 2018-19 | 54% | 52% |
| 2019-20 | 53% | 52% |
| 2020-21 | 53% | 50% |
| 2021-22 | 54% | 53% |
| 2022-23 | 52% | 52% |
| 2023-24 | 54% | 51% |
| 2024-25 | 56% | 52% |
| 2025-26 | 52% | 52% |
- 1991-92: Video replay
- 1992-93: Ejection for starting a fight
- 2020-21: Empty arenas
- Reference: Even (50%)
In 2019 Marshall B. Jones, writing in the journal SAGE Open, put the drop in the 1990s. The published summary of his paper says home advantage “fell by half, from 60% wins (or higher prior to 1991) to 55% wins, and has remained there until the present.” It names two rule changes from the start of that decade as complications: instant video replay in 1991-92 and, the following season, the instigator rule, which threw a player out of the game for starting a fight. We rebuilt every regular-season game since 1917-18, 65,242 of them, and asked a plainer question: in those years, did home teams stop getting their extra power plays?
It happened in 1992-93, the year after replay arrived
Replay arrived in a good year for home teams. In 1991-92 they took 60.5% of the points, their best year since 1982-83. The fall came the next autumn: 55.6% in 1992-93 and 54.0% in 1993-94.
A lot changed that season. Scouting the Refs’ list of rule changes for 1992-93 includes a “Game misconduct penalty for instigating a fight” and a “Minor penalty for attempting to draw a penalty (“diving”)”. When a player from each team went to the box at the same time with both sides at full strength, teams could no longer send on substitutes, “effectively repealing the 1985 rule.” Ottawa and Tampa Bay joined, and the schedule grew to 84 games, with 24 of them played in cities without an NHL team. Any of those, or something no rulebook records, could be the cause. The data can’t say which. It can say when.
We tried every season since 1942-43 as a possible turning point, comparing home results in the six seasons after it with the six before. Eras settle games differently, so we judged each game after 60 minutes. A win in regulation counts as a win. A game still level counts as a half, so an overtime win in the 2010s counts the same as a tie in the 1970s. No season stands out like 1992-93. Home teams’ share of results dropped by about 3 points out of 100 in one step, too big and too sudden to be luck. (The short wartime seasons of the 1940s swing further, but with so few games luck can make swings that size.) The drop shows up in both halves when we split the games in two, odd-numbered and even. It shows up among the teams that existed before the 1990s expansions, and without the two seasons that had games in cities without an NHL team. And all seventeen franchises that played in both eras had a smaller home edge afterwards.
Since 1942-43, no drop in home results stands out from luck like 1992-93’s
Home teams’ share of results in the six seasons from each year, minus the six seasons before it, 1942-43 to 2025-26
points
Show data
| Season | Six seasons after minus six before |
|---|---|
| 1945-46 | −4.3 points |
| 1946-47 | −4.5 points |
| 1947-48 | −3.5 points |
| 1948-49 | −0.9 points |
| 1949-50 | −0.5 points |
| 1950-51 | 1.1 points |
| 1951-52 | 1.6 points |
| 1952-53 | 1.3 points |
| 1953-54 | 0.8 points |
| 1954-55 | −1.5 points |
| 1955-56 | −0.6 points |
| 1956-57 | −1.3 points |
| 1957-58 | −2.4 points |
| 1958-59 | 0.4 points |
| 1959-60 | 0.9 points |
| 1960-61 | 1.6 points |
| 1961-62 | 1.2 points |
| 1962-63 | 1.0 points |
| 1963-64 | 2.8 points |
| 1964-65 | −0.1 points |
| 1965-66 | 1.3 points |
| 1966-67 | 1.4 points |
| 1967-68 | 1.8 points |
| 1968-69 | 1.7 points |
| 1969-70 | 1.0 points |
| 1970-71 | 1.8 points |
| 1971-72 | −0.4 points |
| 1972-73 | −0.8 points |
| 1973-74 | −2.1 points |
| 1974-75 | −2.0 points |
| 1975-76 | −2.6 points |
| 1976-77 | −2.3 points |
| 1977-78 | −1.2 points |
| 1978-79 | −1.0 points |
| 1979-80 | −0.4 points |
| 1980-81 | −0.5 points |
| 1981-82 | 0.0 points |
| 1982-83 | −0.8 points |
| 1983-84 | −1.7 points |
| 1984-85 | −1.5 points |
| 1985-86 | −1.0 points |
| 1986-87 | −0.4 points |
| 1987-88 | −1.2 points |
| 1988-89 | −1.1 points |
| 1989-90 | −0.7 points |
| 1990-91 | −0.9 points |
| 1991-92 | −1.8 points |
| 1992-93 | −3.2 points |
| 1993-94 | −2.8 points |
| 1994-95 | −2.3 points |
| 1995-96 | −2.6 points |
| 1996-97 | −2.6 points |
| 1997-98 | −2.1 points |
| 1998-99 | −1.0 points |
| 1999-00 | −0.4 points |
| 2000-01 | −0.3 points |
| 2001-02 | −0.1 points |
| 2002-03 | 0.3 points |
| 2003-04 | 0.9 points |
| 2004-05 | – |
| 2005-06 | 0.6 points |
| 2006-07 | 0.1 points |
| 2007-08 | 0.1 points |
| 2008-09 | 0.4 points |
| 2009-10 | −0.3 points |
| 2010-11 | −1.2 points |
| 2011-12 | −0.6 points |
| 2012-13 | −0.5 points |
| 2013-14 | −0.7 points |
| 2014-15 | −0.8 points |
| 2015-16 | −0.2 points |
| 2016-17 | 0.3 points |
| 2017-18 | −0.7 points |
| 2018-19 | −0.9 points |
| 2019-20 | −0.5 points |
| 2020-21 | −1.0 points |
| 2021-22 | −1.2 points |
| 2022-23 | −1.1 points |
| 2023-24 | −0.1 points |
- 1992-93: 1992-93
- Reference: No change (0.0 points)
Home teams kept their extra power plays
So was it the referees? If they had changed how they treat home teams, the power-play count is the first place it would show. It shows no change. From 1977-78 to 1991-92 home teams drew 52.1% of the power plays, about one extra power play every 3 games. Over the next dozen seasons they drew 52.0%, still about one extra every 3 games.
In fact, in the two seasons when home results fell, home teams got more extra power plays than before: +0.49 power plays a game in 1992-93 and +0.56 in 1993-94, the biggest home gap in 48 seasons of records. The gap dipped in the late 1990s and recovered by 2000-01, the first season with two referees in every game. The results never recovered.
Home teams kept getting about one extra power play every three games
Home power-play chances minus visitors’, per game, every regular season since 1977-78
PP a game
Show data
| Season | Home minus visitors, power plays a game |
|---|---|
| 1977-78 | 0.22 PP a game |
| 1978-79 | 0.39 PP a game |
| 1979-80 | 0.39 PP a game |
| 1980-81 | 0.37 PP a game |
| 1981-82 | 0.22 PP a game |
| 1982-83 | 0.33 PP a game |
| 1983-84 | 0.38 PP a game |
| 1984-85 | 0.33 PP a game |
| 1985-86 | 0.38 PP a game |
| 1986-87 | 0.38 PP a game |
| 1987-88 | 0.29 PP a game |
| 1988-89 | 0.50 PP a game |
| 1989-90 | 0.49 PP a game |
| 1990-91 | 0.42 PP a game |
| 1991-92 | 0.38 PP a game |
| 1992-93 | 0.49 PP a game |
| 1993-94 | 0.56 PP a game |
| 1994-95 | 0.36 PP a game |
| 1995-96 | 0.21 PP a game |
| 1996-97 | 0.23 PP a game |
| 1997-98 | 0.33 PP a game |
| 1998-99 | 0.24 PP a game |
| 1999-00 | 0.23 PP a game |
| 2000-01 | 0.43 PP a game |
| 2001-02 | 0.36 PP a game |
| 2002-03 | 0.54 PP a game |
| 2003-04 | 0.37 PP a game |
| 2004-05 | – |
| 2005-06 | 0.31 PP a game |
| 2006-07 | 0.40 PP a game |
| 2007-08 | 0.42 PP a game |
| 2008-09 | 0.47 PP a game |
| 2009-10 | 0.30 PP a game |
| 2010-11 | 0.29 PP a game |
| 2011-12 | 0.25 PP a game |
| 2012-13 | 0.22 PP a game |
| 2013-14 | 0.27 PP a game |
| 2014-15 | 0.20 PP a game |
| 2015-16 | 0.31 PP a game |
| 2016-17 | 0.31 PP a game |
| 2017-18 | 0.35 PP a game |
| 2018-19 | 0.21 PP a game |
| 2019-20 | 0.28 PP a game |
| 2020-21 | 0.06 PP a game |
| 2021-22 | 0.31 PP a game |
| 2022-23 | 0.19 PP a game |
| 2023-24 | 0.18 PP a game |
| 2024-25 | 0.26 PP a game |
| 2025-26 | 0.24 PP a game |
- 1992-93: 1992-93
- 2000-01: Two referees every game
- 2020-21: Empty arenas
The individual calls match. We sorted every minor penalty in the game summaries of 1977-78 to 1996-97 by type. The stick and obstruction fouls that rest most on a referee’s judgment, such as hooking, holding, tripping and slashing, went against home teams 48.0% of the time before 1992-93 and 48.0% after.
The whistle did lean in one place. In the last ten minutes of close games, a few more of the calls went against visitors than earlier in those games, about 3 more in every 100 before 1992-93, and about the same after. That could be referees, players, or the home coach’s right to send his players out second, after he sees who the visitors put on the ice. Either way, it didn’t change when home ice did.
The lost goals went missing at even strength
Home teams kept their extra power plays, but they still lost goals. We split the home team’s goal margin in two: special teams (power-play and shorthanded goals) and everything else. Home teams lost about 3 goals every ten games after 1992. About a quarter of that loss came on special teams, and little of it came from the number of chances. Before 1992-93, home power plays scored about 3 more goals in every 100 chances than visitors’ did. Afterwards they scored about 1 more. The other three quarters vanished at even strength, when neither team was on a power play. That count also takes in empty-net goals and penalty shots.
Home teams lost most of their goal edge at even strength
Home goals minus visitors’ goals per game, split into special teams (power-play and shorthanded goals) and the rest
After 1992 the home edge at even strength fell from 0.41 goals a game to 0.17; on special teams it fell from 0.21 to 0.12.
- Special teams
- Even strength and the rest
Shots slipped too, though not all at once. Home teams outshot visitors by about 3 a game before 1992-93 and about 2 after. That gap had been shrinking for more than a decade, so a slow slide fits it about as well as a sudden drop. And the home team’s own off-ice officials count the shots. Our arena scorer study shows how much a home scorer can move a counted stat.
When the crowds left, the extra power plays left too, and home teams kept winning
Then came a season nobody planned for, and with it a real test. In 2020-21, as NBC Sports’ ProHockeyTalk reported, “most of the league’s arenas will not allow fans in.” The seven Canadian teams played only one another, in their own division. Barry Trotz’s Islanders opened their season at Madison Square Garden. “It’s quiet in there. It’s always hopping,” the coach said. “You’re not going to get any help from the atmosphere.”
The power-play part of that help went away. Home teams drew 50.5% of the power plays, the lowest share in 48 seasons and well below the 52.1% of a normal year. Their edge shrank to 0.06 power plays a game. In the all-Canadian division it was 0.01.
Home teams kept winning anyway. Judged after 60 minutes, they took 54.0% of results, against 53.7% in the seasons around it. Swane’s 91-41-17 start was a three-week streak. By season’s end, home teams had taken .593 of the points available to them, close to the .597 of the year before. As ProHockeyTalk’s Sean Leahy put it, “home teams still get last line change.”
Without crowds, power plays evened out; home results didn’t
Home share of power plays, and of results after 60 minutes, 2020-21 against the seasons around it
In 2020-21 home teams drew 50.5% of power plays (52.1% in other recent seasons) and took 54.0% of regulation results (53.7%).
- Home share of power plays
- Home share of results after 60 minutes
One season can’t prove that the whistle never mattered, because the whistle’s part is small. In recent seasons, home teams’ extra power plays were worth about one goal every 22 games. In the standings, that comes to about one game’s result in every 200. Ordinary good and bad luck moves a single season’s home record about three times as much, so one season can’t pick it out. In the 2020 playoff bubbles in Edmonton and Toronto, where “home” was mostly a label, the team listed as home won 45% of 130 games. One tournament is too little to read.
Every team lost home ice in 1992, and no team owns it
If home ice isn’t the whistle, maybe it’s the building. Every era has its legendary home team. The 1975-76 Philadelphia Flyers went 36-2-2 at home and 15-11-14 on the road. The 1972-73 Buffalo Sabres took .808 of their home points and .321 away. The 1983-84 Chicago Blackhawks went 25-13-2 at home and 5-29-6 away. At the other end, the 2022-23 San Jose Sharks won 8 of 41 home games.
None of it repeats. A team’s gap between its home and road records one season tells you almost nothing about its gap the next. We measured the link at 0.05 over 431 pairs of seasons before 1992 and 0.01 over 933 after. If a big home edge carried over from one season to the next, that number would sit near 1. A coin flip would give 0. Give every team the identical home edge, let luck play out a season, and the gaps come out as spread out as the real ones. From 1967-68 to 1991-92, 25 of 459 team-seasons had a home-road gap far out of line with the league’s usual one. Luck alone would put about 23 there. The Flyers sat about 2.4 times as far above their era’s average as luck usually pushes a team in one season. Their great home season is the kind luck hands some team about once every seven years.
Every franchise lost home edge after 1992
Home points share minus road points share, 1967-68 to 1991-92 against 1992-93 to 2025-26, franchises with 200+ home games in both
Show data
| 1967-92 | 1992-2026 | Change | |
|---|---|---|---|
| Pittsburgh Penguins | 0.224 points share | 0.142 points share | −0.082 points share |
| Detroit Red Wings | 0.220 points share | 0.098 points share | −0.122 points share |
| St. Louis Blues | 0.217 points share | 0.091 points share | −0.126 points share |
| Philadelphia Flyers | 0.217 points share | 0.090 points share | −0.127 points share |
| Calgary Flames | 0.219 points share | 0.114 points share | −0.105 points share |
| Chicago Blackhawks | 0.212 points share | 0.083 points share | −0.129 points share |
| Buffalo Sabres | 0.197 points share | 0.099 points share | −0.098 points share |
| Los Angeles Kings | 0.179 points share | 0.104 points share | −0.075 points share |
| Montréal Canadiens | 0.187 points share | 0.087 points share | −0.100 points share |
| Toronto Maple Leafs | 0.186 points share | 0.087 points share | −0.099 points share |
| Boston Bruins | 0.179 points share | 0.077 points share | −0.102 points share |
| New York Islanders | 0.183 points share | 0.081 points share | −0.102 points share |
| Washington Capitals | 0.147 points share | 0.094 points share | −0.053 points share |
| Vancouver Canucks | 0.183 points share | 0.064 points share | −0.119 points share |
| Edmonton Oilers | 0.177 points share | 0.084 points share | −0.093 points share |
| New Jersey Devils | 0.176 points share | 0.073 points share | −0.103 points share |
| New York Rangers | 0.162 points share | 0.046 points share | −0.116 points share |
Whatever happened in 1992 happened to everyone. Over decades a little may stick: Pittsburgh had the biggest home edge of these seventeen franchises in both eras, and the rest line up loosely from one era to the next. With so few teams, that could still be luck. The altitude story doesn’t show up at all. The Colorado Avalanche, at altitude in Denver, have a home edge of .083 over 1,192 home games, a little below the league’s average.
Home ice is mostly something other than extra power plays
Add it up and the extra power plays account for only a sliver of home ice.
- Since 1992, home teams have taken about 4 more results in every 100 than an even split would give them, judging each game after 60 minutes. Before, they took about 9. It is a small edge, and every team gets about the same one.
- The extra power plays are a steady, small part of it: about one game’s result in every 200.
- A team’s great home record one season says almost nothing about its next.
Home ice is at a low, and the power plays still haven’t moved
The 1992 level has slipped again. Judging each game after 60 minutes, the last five seasons, 2021-22 to 2025-26, are the weakest five-season stretch for home teams since at least the 1940s, and 2025-26 gave home teams the second-lowest share of results in 108 seasons, 52.4%. It is too soon to call that a second drop.
In 2025-26, home teams drew 52.1% of the power plays, as they have, give or take a point, since 1977. Swane expected home teams to lose their edge when the buildings went quiet. They lost their extra power plays and kept almost all the rest.
Share this article
Power play tools
Live Edgehalla tools on the same subject.
- Power play tonight PP1 units and power-play matchups for tonight’s games.
- NHL line combinations today Forward lines, defence pairs and PP units for all 32 teams.
- NHL referees Tonight’s officials and how many penalties each crew calls.
- NHL betting history How favourites, puck lines and totals have paid since 2024.
Show the work
- Tests written down
- Not in advance (exploratory)
- Results computed
Every regular-season game from 1917-18 to 2025-26 (65,242 games) and every playoff game (5,110), from the NHL’s game lists, joined to its per-game team reports for penalties (1917-18 on), power-play chances and goals (1977-78 on) and shots. Results are measured two ways: points share (final result, ties a half) and regulation share (games level after 60 minutes count a half), which is the same measure in every era and drives the break tests. Penalty types and the late-game test use the NHL’s game summaries for 1977-78 to 1996-97 (17,216 games). Uncertainties are game-level standard errors. The goal-margin split puts power-play and shorthanded goals in special teams and everything else (even strength, empty nets, penalty shots) in the rest; shootout goals are left out. Not pre-registered: the 1992-93 break year was found by the scan, not fixed in advance, and the late, close-game test was added at the editor’s request after the first results.
| Test | Result | Verdict |
|---|---|---|
| Home results fell in the early 1990s (Jones 2019) | Points share 59.0% (1977-92, 12,360 games) to 55.1% (1992-2004, 13,097); goal margin 0.62 to 0.29 a game | Holds |
More tests below
| Test | Result | Verdict |
|---|---|---|
| The break was 1991-92 (replay) | Six-season step −1.8 points (z -2.04) at 1991-92 vs −3.1 (z -3.7) at 1992-93; best single break 1992-93 | Reframed: 1992-93 |
| Placebo: is 1992-93 special among fake break years? | Most negative step for its luck range (z) of 78 candidate years since 1942-43 (the mid-1940s windows have bigger raw steps, up to about 4.5 points, but z no bigger than 2); only windows overlapping 1992-96 exceed |z| 3 (1992 1993 1995 1996) | Holds |
| Step, not slide | Weighted squared error 0.000278 (step) vs 0.000323 (straight trend); step with trend −3.3 points | Step fits better |
| Same overtime rules on both sides (1983-92 vs 1992-99) | Regulation share 58.3% to 55.3% (z -4.05) | Holds |
Continues below
| Test | Result | Verdict |
|---|---|---|
| Split halves (odd / even game ids) | −3.5 (z -4.4) / −4.4 (z -5.63) points, 1977-92 vs 1992-2004 | Holds in both |
| Not an expansion or neutral-site artefact | Teams that existed in 1990-91 only: −3.4 (z -4.75); without 1992-94: −4.0 (z -6.95) | Holds |
| Every franchise | 17 of 17 franchises with 200+ home games in both eras had a smaller home-minus-road split after 1992 | Holds |
| Playoffs too | Home win share 58.9% (1,089) to 54.2% (1,029), z -2.15 | Same direction |
| Home teams stopped getting extra power plays | Home share 52.1% to 52.0%; gap 0.36 to 0.36 a game (each ±0.04); the gap’s own best break is 2009-10; season-level r with results 0.27 (levels), -0.17 (changes), 48 seasons | Not supported |
Continues below
| Test | Result | Verdict |
|---|---|---|
| Judgment minors shifted toward home teams after 1992 | Home share 48.0% (109,696 calls) to 48.0% (38,592); technical calls (too many men, delay of game) 43.8% before | Not supported |
| Home teams are spared calls late in close games, and that changed in 1992 | Last 10 minutes vs first 50, score mix fixed: −2.9 points (z -3.56, 3,991 late calls) before, −2.6 (z -2.26, 1,983) after; change +0.3 (z 0.19). Raw close-game share late 44.0% vs 46.7% early | Small lean in both eras; no change |
| Where the margin went | Lost 0.34 goals a game: special teams 0.09, the rest 0.25. Special-teams margin 0.21 (1977-78 to 1991-92), 0.12 (1992-93 to 2003-04), 0.10 (2005-06 to 2025-26); the rest 0.41 (1977-78 to 1991-92), 0.17 (1992-93 to 2003-04), 0.17 (2005-06 to 2025-26); games 12,360 / 13,097 / 25,302. Little of the special-teams loss came from the number of chances: home PP% 22.62 vs road 19.99 before, 17.5 vs 16.21 after | Mostly even strength |
Continues below
| Test | Result | Verdict |
|---|---|---|
| Home shot edge fell at the same break | 3.29 to 1.89 shots a game; best break 1991-92, but a straight trend fits as well (0.238995 vs 0.250124) | Slide plus step; counted by the home team’s off-ice officials (see our arena scorer study) |
| Empty arenas removed the power-play edge (2020-21) | Home PP share 50.5% (868 games), 3.2 SD below the other 47 seasons (mean 52.1%, lowest 51.0%) | Holds |
| Empty arenas removed home-ice results (2020-21) | Regulation share 54.0% ±2.9 vs 53.7% (10,184 games); North Division 51.8% ±6.4 vs 53.9% for the same matchups | Not supported (low power) |
| Designated home teams in the 2020 bubbles | Home win share 45.4% of 130 vs 54.3% (1,734 other playoff games since 2005-06) | Noise (z about −2, one tournament) |
Continues below
| Test | Result | Verdict |
|---|---|---|
| Team home edges repeat | Next-season r 0.051 (95% -0.044 to 0.144, 431 pairs) before 1992; 0.005 (-0.059 to 0.069, 933) after; spread SD 0.0962 vs noise 0.0983; the 1975-76 Flyers’ split sits 2.4 standard errors above its era’s average, and 25 of 459 team-seasons in 1967-92 were beyond 2 (about 23 by chance) | No |
| Franchise home edges persist across eras | r 0.57 over 17 franchises (wide range); Pittsburgh top in both | Unproven |
| A second slide since 2021 | 2021-26 regulation share 53.2% vs 54.7% in 2005-20 (about 2.4 standard errors) | Watch, not established |
Frequently asked questions
When did NHL home-ice advantage decline?
In the 1992-93 season. Home teams took 60.5% of points in 1991-92 and 55.6% in 1992-93, and about 55% on average since. Before 1992 they had taken about 59% to 60% in almost every era since the 1940s.
Do NHL home teams get more power plays?
Yes, slightly. Home teams have drawn about 52% of power-play chances since the NHL began recording them in 1977-78, about 0.31 more a game than visitors since 1992. That share did not change when home results fell in 1992-93.
Did NHL home-ice advantage disappear without fans in 2020-21?
No. Home teams took 54.0% of regulation results, close to the 53.7% of nearby seasons. What disappeared was most of their power-play edge: they drew 50.5% of power plays, the lowest share on record.
Which NHL team has the best home-ice advantage?
None that lasts. A team’s home-minus-road split barely predicts the next season’s (correlation about 0.01 since 1992). The best home record since expansion is the 1975-76 Flyers’ 36-2-2, a season the data treats as mostly luck.
Does altitude give the Colorado Avalanche a home-ice advantage?
Not visibly. Colorado’s home-minus-road points split since 1995 is .083 over 1,192 home games, a little below the league average of .095 since 1992-93.
More in Standings and results
- Verdict: Data story
Regular-Season Records Predict NHL Playoff Upsets Almost Exactly
- Verdict: Data story
The Schedule Got Kinder. The Second Night Didn’t.
- Verdict: Data story
Are the NHL’s Stars Pulling Away From Everyone Else?
- Verdict: Data story
Vegas Had the Best First Season of Any NHL Expansion Team
- Verdict: Data story
Does Scoring Depth Win Stanley Cups?
- Verdict: Data story
Same Goals, Half the Blowouts
Sources and method
- NHL stats (NHL.com), every season since 1917-18
- NHL stats API: per-game team penalty, power-play and summary reports (isGame, homeRoad)
- Marshall B. Jones, “The Sustained Reduction-by-Half of Home Advantage in the NHL, 1991-1992 to 2000-2001”, SAGE Open (2019)
- Crossref record and abstract for the Jones study
- Scouting the Refs: NHL historical rule changes
- Wikipedia: 1992-93 NHL season (84 games, neutral-site games)
- The Hockey Writers, Brian Swane: NHL teams dominating on home ice despite absence of fans (Feb. 4, 2021)
- NBC Sports, Sean Leahy: “A continuation of bubble hockey”, the NHL game experience in 2021 (Feb. 3, 2021)
- 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.