How Hockey ChangedRules
Nearly One Game in Four Now Goes Past 60 Minutes
Overtime has been added, scrapped, brought back, shrunk to four skaters, backed by a shootout and cut to three-on-three. Since overtime returned in 1983-84, the share of NHL games still tied after regulation has crept up. We charted every season since 1917-18 and tested whether tied teams ease off to bank the point.
By Edgehalla ResearchPublished 8 min read
More takeaways below
- Three-on-three cut shootouts by more than a third. Shootouts decided 13.8% of games in 2014-15 and 8.7% a year later, the first season of 3-on-3.
On this page
How many NHL games go to overtime?
In 2025-26, 326 of 1312 games (24.8%) were tied after 60 minutes, the most since 2014-15. That is high but not unusual for the loser-point era: every season since 1999-2000 has been between 20.7% and 25.6%.
Share of NHL games still tied after regulation
Regular season. From 1942-43 to 1982-83 there was no overtime, so every tied game ended as a tie.
Show data
| Season | Went past 60 minutes | Ended in a tie |
|---|---|---|
| 1917-18 | 5.6% | 0.0% |
| 1918-19 | 11.1% | 0.0% |
| 1919-20 | 8.3% | 0.0% |
| 1920-21 | 10.4% | 0.0% |
| 1921-22 | 10.4% | 4.2% |
| 1922-23 | 16.7% | 4.2% |
| 1923-24 | 12.5% | 0.0% |
| 1924-25 | 11.1% | 3.3% |
| 1925-26 | 18.3% | 8.7% |
| 1926-27 | 22.7% | 8.2% |
| 1927-28 | 21.8% | 16.8% |
| 1928-29 | 27.7% | 21.4% |
| 1929-30 | 23.2% | 13.2% |
| 1930-31 | 21.4% | 15.0% |
| 1931-32 | 28.6% | 18.2% |
| 1932-33 | 23.6% | 17.1% |
| 1933-34 | 27.8% | 17.6% |
| 1934-35 | 19.4% | 12.5% |
| 1935-36 | 25.5% | 17.7% |
| 1936-37 | 22.4% | 14.6% |
| 1937-38 | 22.9% | 18.8% |
| 1938-39 | 19.6% | 14.9% |
| 1939-40 | 17.9% | 12.5% |
| 1940-41 | 27.4% | 18.5% |
| 1941-42 | 18.5% | 7.1% |
| 1942-43 | 2.0% | 21.3% |
| 1943-44 | – | 10.7% |
| 1944-45 | – | 11.3% |
| 1945-46 | – | 15.3% |
| 1946-47 | – | 14.4% |
| 1947-48 | – | 18.9% |
| 1948-49 | – | 15.6% |
| 1949-50 | – | 19.5% |
| 1950-51 | – | 21.4% |
| 1951-52 | – | 18.1% |
| 1952-53 | – | 22.4% |
| 1953-54 | – | 15.7% |
| 1954-55 | – | 23.8% |
| 1955-56 | – | 17.6% |
| 1956-57 | – | 19.0% |
| 1957-58 | – | 16.2% |
| 1958-59 | – | 15.7% |
| 1959-60 | – | 17.1% |
| 1960-61 | – | 18.6% |
| 1961-62 | – | 17.1% |
| 1962-63 | – | 22.4% |
| 1963-64 | – | 16.7% |
| 1964-65 | – | 13.8% |
| 1965-66 | – | 13.3% |
| 1966-67 | – | 14.8% |
| 1967-68 | – | 17.1% |
| 1968-69 | – | 16.9% |
| 1969-70 | – | 20.0% |
| 1970-71 | – | 15.8% |
| 1971-72 | – | 16.8% |
| 1972-73 | – | 13.9% |
| 1973-74 | – | 16.7% |
| 1974-75 | – | 16.8% |
| 1975-76 | – | 15.8% |
| 1976-77 | – | 15.7% |
| 1977-78 | – | 18.3% |
| 1978-79 | – | 16.3% |
| 1979-80 | – | 16.9% |
| 1980-81 | – | 19.6% |
| 1981-82 | – | 17.6% |
| 1982-83 | – | 15.1% |
| 1983-84 | 16.7% | 10.2% |
| 1984-85 | 17.7% | 12.4% |
| 1985-86 | 16.1% | 9.4% |
| 1986-87 | 17.6% | 11.1% |
| 1987-88 | 17.4% | 11.5% |
| 1988-89 | 17.7% | 11.5% |
| 1989-90 | 18.5% | 11.9% |
| 1990-91 | 19.8% | 13.3% |
| 1991-92 | 19.2% | 13.3% |
| 1992-93 | 16.4% | 9.9% |
| 1993-94 | 19.6% | 12.8% |
| 1994-95 | 16.2% | 12.0% |
| 1995-96 | 18.9% | 12.9% |
| 1996-97 | 20.1% | 13.5% |
| 1997-98 | 20.5% | 15.5% |
| 1998-99 | 20.1% | 14.6% |
| 1999-00 | 22.7% | 12.7% |
| 2000-01 | 22.3% | 12.4% |
| 2001-02 | 22.0% | 12.1% |
| 2002-03 | 25.4% | 12.8% |
| 2003-04 | 25.6% | 13.8% |
| 2004-05 | – | – |
| 2005-06 | 22.8% | – |
| 2006-07 | 22.8% | – |
| 2007-08 | 22.1% | – |
| 2008-09 | 22.9% | – |
| 2009-10 | 24.5% | – |
| 2010-11 | 24.1% | – |
| 2011-12 | 24.4% | – |
| 2012-13 | 22.5% | – |
| 2013-14 | 25.0% | – |
| 2014-15 | 24.9% | – |
| 2015-16 | 22.4% | – |
| 2016-17 | 23.5% | – |
| 2017-18 | 23.3% | – |
| 2018-19 | 21.3% | – |
| 2019-20 | 23.1% | – |
| 2020-21 | 22.5% | – |
| 2021-22 | 22.0% | – |
| 2022-23 | 23.0% | – |
| 2023-24 | 20.7% | – |
| 2024-25 | 20.7% | – |
| 2025-26 | 24.8% | – |
- 1921-22: OT capped at 20 min
- 1928-29: 10-min OT
- 1942-43: OT dropped
- 1983-84: 5-min OT
- 1999-00: Loser point, 4-on-4
- 2005-06: Shootout
- 2015-16: 3-on-3
The rules explain the shape. Early overtime ran until someone scored, then was capped at 20 minutes for 1921-22, which is when ties first appear. In 1927-28 it became 10 minutes of sudden death, and from 1928-29 teams played a full 10 minutes that could end level. On November 21, 1942, the league dropped regular-season overtime because of wartime train schedules, and for 41 seasons a tie after 60 minutes was final. Five-minute sudden-death overtime returned in 1983-84.
What the loser point did
From 1999-2000, a team that lost in overtime kept a point, and overtime became four-on-four. A team tied late in the third was now guaranteed a point and could still win two. If teams responded, more games should reach overtime tied.
Scoring changes the odds of a tie by itself: fewer goals means more ties. So we set a baseline for every season: the chance that two teams, each scoring at half the league’s regulation rate, finish 60 minutes level by luck alone (a Poisson model). It is not a model of hockey, just a yardstick that moves with scoring.
Games tied after 60 minutes, against a scoring baseline
Actual share vs the share expected from each season’s regulation scoring
Show data
| Season | Tied after 60 minutes | Scoring baseline |
|---|---|---|
| 1942-43 | 21.3% | 15.1% |
| 1943-44 | 10.7% | 14.2% |
| 1944-45 | 11.3% | 15.0% |
| 1945-46 | 15.3% | 15.7% |
| 1946-47 | 14.4% | 16.2% |
| 1947-48 | 18.9% | 16.9% |
| 1948-49 | 15.6% | 17.6% |
| 1949-50 | 19.5% | 17.5% |
| 1950-51 | 21.4% | 17.6% |
| 1951-52 | 18.1% | 18.0% |
| 1952-53 | 22.4% | 18.8% |
| 1953-54 | 15.7% | 18.7% |
| 1954-55 | 23.8% | 18.3% |
| 1955-56 | 17.6% | 18.2% |
| 1956-57 | 19.0% | 17.7% |
| 1957-58 | 16.2% | 17.3% |
| 1958-59 | 15.7% | 17.0% |
| 1959-60 | 17.1% | 16.8% |
| 1960-61 | 18.6% | 16.7% |
| 1961-62 | 17.1% | 16.6% |
| 1962-63 | 22.4% | 16.7% |
| 1963-64 | 16.7% | 17.4% |
| 1964-65 | 13.8% | 17.0% |
| 1965-66 | 13.3% | 16.5% |
| 1966-67 | 14.8% | 16.7% |
| 1967-68 | 17.1% | 17.3% |
| 1968-69 | 16.9% | 16.7% |
| 1969-70 | 20.0% | 17.0% |
| 1970-71 | 15.8% | 16.3% |
| 1971-72 | 16.8% | 16.5% |
| 1972-73 | 13.9% | 15.9% |
| 1973-74 | 16.7% | 16.1% |
| 1974-75 | 16.8% | 15.5% |
| 1975-76 | 15.8% | 15.6% |
| 1976-77 | 15.7% | 15.8% |
| 1977-78 | 18.3% | 15.9% |
| 1978-79 | 16.3% | 15.4% |
| 1979-80 | 16.9% | 15.3% |
| 1980-81 | 19.6% | 14.6% |
| 1981-82 | 17.6% | 14.3% |
| 1982-83 | 15.1% | 14.6% |
| 1983-84 | 16.7% | 14.5% |
| 1984-85 | 17.7% | 14.6% |
| 1985-86 | 16.1% | 14.5% |
| 1986-87 | 17.6% | 15.1% |
| 1987-88 | 17.4% | 15.0% |
| 1988-89 | 17.7% | 14.9% |
| 1989-90 | 18.5% | 15.0% |
| 1990-91 | 19.8% | 15.6% |
| 1991-92 | 19.2% | 15.5% |
| 1992-93 | 16.4% | 15.2% |
| 1993-94 | 19.6% | 16.1% |
| 1994-95 | 16.2% | 16.8% |
| 1995-96 | 18.9% | 16.3% |
| 1996-97 | 20.1% | 17.0% |
| 1997-98 | 20.5% | 17.9% |
| 1998-99 | 20.1% | 18.0% |
| 1999-00 | 22.7% | 17.6% |
| 2000-01 | 22.3% | 17.6% |
| 2001-02 | 22.0% | 18.1% |
| 2002-03 | 25.4% | 18.0% |
| 2003-04 | 25.6% | 18.3% |
| 2004-05 | – | – |
| 2005-06 | 22.8% | 16.8% |
| 2006-07 | 22.8% | 17.2% |
| 2007-08 | 22.1% | 17.7% |
| 2008-09 | 22.9% | 17.3% |
| 2009-10 | 24.5% | 17.6% |
| 2010-11 | 24.1% | 17.7% |
| 2011-12 | 24.4% | 17.9% |
| 2012-13 | 22.5% | 17.9% |
| 2013-14 | 25.0% | 17.9% |
| 2014-15 | 24.9% | 17.9% |
| 2015-16 | 22.4% | 18.0% |
| 2016-17 | 23.5% | 17.8% |
| 2017-18 | 23.3% | 17.1% |
| 2018-19 | 21.3% | 16.9% |
| 2019-20 | 23.1% | 17.0% |
| 2020-21 | 22.5% | 17.2% |
| 2021-22 | 22.0% | 16.6% |
| 2022-23 | 23.0% | 16.5% |
| 2023-24 | 20.7% | 16.6% |
| 2024-25 | 20.7% | 16.9% |
| 2025-26 | 24.8% | 16.7% |
- 1999-00: Loser point
Extra ties above the baseline, by rule era
Percentage points of games tied after 60 minutes beyond the scoring baseline
Excess ties were 2.6 points in 1983-99 and 5.7 points in 1999-2004, after the loser point arrived.
The jump at 1999 is clean. Every season since 1999-2000 is at least 3.8 points above the baseline. Only 1 of the 16 seasons from 1983-84 to 1998-99 reached that. It fits Jason Abrevaya’s 2004 study of the same rule change. Four-on-four overtime arrived the same season, so the data cannot separate the two. One puzzle: the tie era before 1983, when a tie also gave both teams a point, shows little excess. A tie then meant one point each with no chance at two, which may be less tempting than a free point plus overtime.
Do tied teams ease off late in the third?
If teams play for the point, it should show in the last ten minutes of tied games. We took every regular-season and playoff game from 2017-18 to 2025-26 that was tied entering the 41st and 51st minutes, and counted shot attempts per minute while it stayed tied. Playoff games are the control: there is no loser point, so nobody has a reason to coast.
Shot attempts per minute while tied
Both teams, all situations, 2017-18 to 2025-26
In tied regular-season games, attempts fell from 1.893 to 1.833 a minute from minutes 41-50 to 51-60; in tied playoff games they fell from 1.999 to 1.859.
- Minutes 41-50
- Minutes 51-60
Attempts dip late in both. The regular-season dip (−3.2%) is, if anything, smaller than the playoff dip (−7.0%). The ratio of the two is 1.041, with a 95% range of 0.980 to 1.105: no sign of regular-season coasting, and the playoff sample is small.
| Seasons | Regular season | Playoffs | Regular ÷ playoffs (95% range) |
|---|---|---|---|
| 2017-18, 2019-20, 2021-22, 2023-24, 2025-26 | 0.961 | 0.960 | 1.001 (0.926–1.082) |
Continues below
| Seasons | Regular season | Playoffs | Regular ÷ playoffs (95% range) |
|---|---|---|---|
| 2018-19, 2020-21, 2022-23, 2024-25 | 0.977 | 0.890 | 1.098 (0.998–1.207) |
| All nine | 0.968 | 0.930 | 1.041 (0.980–1.105) |
How 3-on-3 shrank the shootout
The shootout arrived in 2005-06 and peaked at 15.0% of all games in 2009-10. Three-on-three overtime, from 2015-16, decides most overtime games before a shootout: shootouts fell to 8.7% of games that season and 5.9% in 2024-25, before 9.1% in 2025-26.
How games past regulation end
Share of all regular-season games, since the shootout began
Show data
| Season | Decided in overtime | Shootout |
|---|---|---|
| 2005-06 | 11.1% | 11.8% |
| 2006-07 | 9.5% | 13.3% |
| 2007-08 | 9.4% | 12.7% |
| 2008-09 | 10.0% | 12.9% |
| 2009-10 | 9.5% | 15.0% |
| 2010-11 | 12.0% | 12.1% |
| 2011-12 | 9.7% | 14.7% |
| 2012-13 | 9.0% | 13.5% |
| 2013-14 | 10.5% | 14.5% |
| 2014-15 | 11.1% | 13.8% |
| 2015-16 | 13.7% | 8.7% |
| 2016-17 | 15.4% | 8.0% |
| 2017-18 | 15.2% | 8.1% |
| 2018-19 | 14.5% | 6.8% |
| 2019-20 | 15.2% | 7.9% |
| 2020-21 | 15.0% | 7.5% |
| 2021-22 | 14.2% | 7.8% |
| 2022-23 | 15.8% | 7.2% |
| 2023-24 | 14.5% | 6.3% |
| 2024-25 | 14.8% | 5.9% |
| 2025-26 | 15.8% | 9.1% |
- 2015-16: 3-on-3
What to keep in mind
- Early seasons are short. The NHL played 18 to 44 games a team before 1931, so one season can swing several points.
- The baseline is deliberately crude. Real scores are not independent Poisson counts: trailing teams push, leading teams sit back, and empty-net goals stretch margins. That is why the baseline sits below the actual line in most eras. What matters is the change at 1999, not the level.
- The coasting test covers 2017-18 on, when our play-by-play starts. It cannot speak to the late 1990s, when the incentive was new.
Show the work
- Tests written down
- Not in advance (exploratory)
- Results computed
Every NHL regular-season game since 1917-18 (65,242) by final period from the NHL stats API. The coasting test uses every shot attempt in our play-by-play from 2017-18 to 2025-26 in games tied entering the 41st or 51st minute, with playoffs as the no-loser-point control and the seasons split in two halves as a holdout.
| Test | Result | Verdict |
|---|---|---|
| More games go past regulation since the loser point, beyond scoring | +2.6 pts (1983-99) vs +5.5 to +6.1 (since 1999) | Holds |
| Tied regular-season teams ease off late (vs playoffs) | ratio 1.041 (0.980–1.105) | Not seen |
| Same test, odd vs even seasons | 1.001 vs 1.098 | Not seen in either half |
Frequently asked questions
What percentage of NHL games go to overtime?
About 23% since the loser point arrived in 1999-2000, and 24.8% in 2025-26. From 1983-84 to 1998-99, with five-minute overtime and no point for an overtime loss, it was about 18%.
When did the NHL get rid of ties?
In 2005-06, when the shootout was added to decide games still tied after overtime. From 1942-43 to 1982-83 there was no regular-season overtime at all, so about 15-20% of games ended in ties.
Does the loser point make teams play for overtime?
Overtime became more common after it: games tied after 60 minutes rose about three points more than scoring explains. But shot attempts late in tied games do not drop more in the regular season than in the playoffs, so there is no visible sign of teams sitting back.
How often do NHL games go to a shootout?
About 6% to 9% of games since 3-on-3 overtime began in 2015-16 (9.1% in 2025-26), down from 12% to 15% between 2005-06 and 2014-15.
Why did the NHL stop overtime in 1942?
Regular-season overtime was discontinued on November 21, 1942, because of wartime train restrictions. It did not return until 1983-84.
Sources and method
- NHL stats (NHL.com), every season since 1917-18
- NHL stats API: every regular-season game with its final period, 1917-18 to 2025-26
- Scouting the Refs: NHL historical rule changes
- Wikipedia: 1999–2000 NHL season (point for an overtime loss, four-on-four overtime)
- Purdue News: Abrevaya on the incentive effects of the NHL overtime rule (2002); paper in Journal of Sports Economics 5(3), 2004
Research and analysis, not betting advice. Past results do not guarantee future returns; bet only what you can afford to lose.