Same Goals, Half the Blowouts

The blowout halved, and the biggest ones vanished

Start with the final scores, which the NHL has for every game since 1917-18. From 1942-43 to 1966-67, the Original Six years, about one in ten regular-season games ended with a five-goal margin. Since 2015-16 it has been about one in 19. Routs haven’t disappeared. Teams lost by five or more 72 times in 2025-26.

Some routs would happen even in a league of identical teams. Picture one where every club is the same and each goal is a coin toss between the two benches. Most nights the coin splits close. Now and then it lands on one side five times more than the other. That coin-flip league is the dashed line below. It scores at each season’s real rate, and nobody in it is better than anybody else.

The solid line has fallen. Watch the gap between the two lines instead. From the war through the 1970s, real games beat the coin flips for five-goal finals in 36 of 38 seasons. Since 1999-2000 they have fallen short in 24 of 26.

The very top of the scale is gone. No team has won by 11 or more since the Calgary Flames beat the San Jose Sharks 13-1 on Feb. 10, 1993. The last double-digit margin came on Nov. 4, 2016, when the Columbus Blue Jackets beat the Montréal Canadiens 10-0. “While that game was the first since 2003,” the blog Down Goes Brown noted five days later, “the Canadiens actually became the 74th team in league history to lose a game by 10 or more.” Of the 1944 game, it added: “The Rangers managed only nine shots, none of which found the back of the net.”

Scoring held steady, and lopsided rosters explain only part of it

Fewer goals would mean fewer chances to pile up five. Scoring hasn’t dropped, though. Original Six teams scored 2.90 goals a game. Teams since 2015-16 score 2.97. Hand both leagues coin-flip scoring and today’s would produce slightly more five-goal finals than the Original Six’s would.

Lopsided rosters do produce routs. In 1974-75 the expansion Washington Capitals went 8-67-5 and lost 25 games by five or more, 31% of their schedule. Two seasons later the Montréal Canadiens went 60-8-12 and won 22 games by five or more, the most of any team in any season. Put those two clubs in the same league and routs follow, whatever else is going on. Today’s teams sit closer together, and that explains part of the drop.

More equal teams explain a third. How leads behave explains the rest

To find the rest, we sorted each era’s five-goal games into three piles. The first pile is the coin-flip league at that era’s scoring rate. For the second, we let every team play to its real goals for and against that season, so the 1976-77 Canadiens play like the 1976-77 Canadiens and the 1974-75 Capitals like the Capitals. The extra routs that adds go in pile two. Pile three is whatever real games produced beyond that. It comes from how teams play once the score changes: whether a two-goal lead tends to become three, or one.

From the Original Six to now, five-goal games fell from 9.8 in every hundred games to 5.3, a drop of about 4.5. More equal teams took away about 1.5. The way leads behave took away about 3.5. Scoring, a touch higher today, added back about 0.5. That is 1.5 + 3.5 − 0.5 = 4.5 games in every hundred.

Before 1980 the how-leads-behave bar sits above zero. In the Original Six it added about 1.7 routs per hundred games. From 1980 on it sits below zero in every era, and today it takes away about 1.7. The unequal-teams bar is tallest in the expansion years, when those Capitals and Canadiens shared a league. We also split the seasons in two, odd years like 1951-52 in one group and even years like 1952-53 in the other. Both groups show the same flip in every era, so a few freak seasons aren’t behind it.

Leads used to grow. Since the 1980s they shrink

If how leads behave is most of the answer, it should show in the simplest moment in hockey: one team is ahead, and somebody scores next. Say the Canadiens of the late 1950s are up 2-0 on the Bruins. Montréal was the better team, so it should get more than its share of the next goals anyway. We know how many goals each team scored and allowed that season, and that tells us roughly how often Montréal should score next. Then we count how often leaders like it really did. We did that for every regulation goal scored while one team led. Goals into an empty net are left out.

In the Original Six, a team up two scored the next goal 55% of the time. That was about three goals in every hundred more than the two teams’ records called for. A two-goal lead tended to become three. Since 2015-16, a team up two has scored next 46% of the time, about five in every hundred fewer than its record calls for. The team behind gets the next goal more often than its record explains, at every lead size.

A bar above zero means leads tended to grow. Below zero, they tended to shrink. The two earliest eras sit above the line. The expansion years sit close to it. Every era since 1980 sits below it, at every lead size. Most modern bars rest on ten thousand goals or more, far too many for gaps this size to be luck. The smallest, teams up three in the five loser-point seasons, rests on about 3,300 goals, and that one bar could be luck. Odd-year and even-year seasons agree in every era since 1980.

Leads shrink late, and they shrink early too

Some of the swing is the pulled goalie. A team up two with five minutes left sees the other coach wave his goalie to the bench for an extra skater. Set aside the empty-netters. Of the goals left over, today’s leaders score only about one in four (27%). In the Original Six they scored 57%.

A one-goal lead late has gone the same way. In the 1980s and 1990s, a team up one in the last five minutes scored next about nine times in a hundred less often than its record said it should. That shortfall has grown in every era since, to about 22 today, as coaches have pulled the goalie earlier. Our goalie-pull story covers that creep.

Nobody pulls a goalie in the first two periods. Leads shrink there anyway. Since 2015-16, before the third period starts, leaders score next about one to three times in a hundred less often than their records say they should.

Empty nets make close games look like routs

If leads shrink, more games should be close at the end. The box score hides it. Since 2020, a three-goal final (23.5% of games) has been more common than a two-goal final (19.4%). Picture a 3-2 game with two minutes left: the goalie comes out, the leader scores into the empty net, and it goes in the books as 4-2. Add a second empty-netter and it’s 5-2. Take the empty-net goals out and three-goal games fall to 12.0%.

At three goals, the 2020s as recorded tower over the version without empty nets. At one goal the order flips. About one in eight games in 2025-26 was a two-goal win with an empty-netter in it.

Take the empty-netters out, and the final scores show the same shrinking leads as the next goal does. Since 2015-16, 56% of games have finished tied or within a goal. The coin-flip league would give 49%. The 1980s and 1990s were already closer than the coin flips, 48% against 45%. The Original Six had hardly any empty-net goals to take out, and its games finished close less often than the coin flips would have, 43% against 48%. Empty nets barely touch the real blowouts: 5.3% of modern games are five-goal games as recorded, 5.1% without them.

Five goals down, and still not beaten

A team behind that scores next more often should also come back more often. Start with the nights that seemed impossible. On Nov. 29, 2000, the Toronto Maple Leafs led the St. Louis Blues by five goals with 17:06 to play. How often does a team climb out of that? Take how often the Blues and the Leafs scored and gave up goals that season. Play the last 17 minutes forward ten seconds at a time, and count every way St. Louis could draw level before the horn. It comes to under one chance in a thousand. St. Louis scored five times in the third period and won 6-5 in overtime. Every “odds” figure in this story is worked out the same way, from the two teams’ season records.

On Feb. 19, 2008, the Montréal Canadiens fell five goals behind the New York Rangers early in the second period. The odds gave them under one chance in a hundred of pulling even. They won 6-5 in a shootout. On April 10, 1982, Wayne Gretzky’s Oilers led the Kings 5-0 at the Forum after 40 minutes. Los Angeles scored five in the third, the last with five seconds left, and Daryl Evans won it in overtime. “With only five seconds to go, the Kings had completed a historic comeback,” Wikipedia’s account reads; the game “remains the biggest single-game comeback in Stanley Cup playoffs history.” From the moment the Kings first trailed by five, late in the second period, our odds gave them under one chance in a hundred of drawing level.

The regular season has stranger nights. On Feb. 23, 1989, Detroit trailed Pittsburgh 6-0 early in the second and tied it 6-6, Steve Yzerman scoring the last goal. It is the only six-goal deficit erased in the seasons we can follow goal by goal. On Oct. 17 that year, Calgary trailed Quebec 8-3 with 8:33 left. Gary Roberts scored twice in 16 seconds and Jim Peplinski added another 11 seconds later. Then Doug Gilmour and Paul Ranheim scored shorthanded at 19:45 and 19:49. It ended 8-8. The odds put that one at under one chance in a thousand, too, the same as the Blues. In all, 21 teams in our regular-season records have erased a deficit of five goals or more.

Since 1980, teams down two have beaten the odds

Five-goal comebacks are freaks. The everyday test is a team down two. Before 1980 those teams tied it a little less often than the odds said they would. Since 1980 they have tied it more often, in every era.

Before 1980 the “drew level” bar sits just under “the odds” in every era. From 1980 on it sits above it every time. More goals mean more ties. The 1980s and 1990s scored the most, and their teams down two tied it most often.

Count it per hundred. Before 1980, for every 100 ties the odds expected from teams down two, they managed about 96 or 97. Every era since has beaten the odds, at 108 to 115 ties per 100 expected, and odd-year and even-year seasons agree. Down three, modern teams tie it 6.6% of the time, also above the odds. The pulled goalie helps at the very end. Since 2017-18, about one in nine completed two-goal comebacks was tied six-on-five, with the trailing team’s goalie on the bench for an extra skater.

The loser point and the shootout didn’t cause it

Three rules made a late tying goal worth more. Since 1999-2000, a team that loses in overtime still gets a point in the standings, the loser point. Since 2005-06 no regular-season game ends in a tie; a shootout settles it. Since 2015-16 overtime is three-on-three. Our overtime story has the rule texts.

The rise came first. Teams in the 1980s and 1990s were already beating the odds, at 108 ties per 100 expected, years before the loser point. We still tested each rule, comparing the five seasons before it with the five after. Measured against the odds, about one and a half more teams in every hundred down two tied it in the five years after the loser point than in the five before. After the shootout the gain was about half a team. After 3-on-3 it went the other way, about one and a half teams fewer.

Swings that size happen in years when none of these rules changed. We ran the same five-before, five-after comparison around every ordinary season from 1972 on, such as 1985-86, when nothing about overtime changed. For the loser point, about three in ten of those ordinary seasons saw a swing at least as big. For the shootout it was about seven in ten, and for 3-on-3 about three in ten. None of the three rules moved comebacks more than an ordinary year does.

The playoffs, which have none of those rules, tell the same story

Playoff games have no loser point and no shootout, and overtime stays five-on-five until somebody scores. If the rules drove comebacks, playoff comebacks shouldn’t have moved when they arrived. Around the loser point, playoff comebacks moved by almost nothing. Around 3-on-3 they fell by about one tie per hundred. Around the shootout they jumped by about three and a half ties, more than in the regular season. Only 254 playoff teams fell two behind in the four seasons after, though, so that jump could easily be luck too.

The shootout did change who gets the second point. Teams down two won 8.5% of the time in the loser-point years, when games could still end tied, and 12.1% once every game had a winner. Their share of games that earned at least a point barely moved, from 17.1% to 18.4%.

No team has proved it is good at coming back

If comebacks beat the odds league-wide, maybe some teams have the knack. Every era has a team that seems never to be out of a game. In 1979-80 the Philadelphia Flyers went 35 games without a loss, from the middle of October until the first week of 1980. They fell two goals behind 27 times that season, drew level 17 times and won 11. The odds expected about 10 ties. Line up every team’s season we can follow goal by goal, 1,496 of them, and the 1979-80 Flyers rank third for comebacks above the odds. “I’ll never work with a group of athletes like that again,” their coach, Pat Quinn, told ESPN years later.

The next season they drew level 12 times in 31, about what the odds expected. Hot comeback seasons usually go like that. Split any team’s season into its odd-numbered and even-numbered games. A team that comes back often in one set does no better than average in the other.

Each dot is one team’s season. A team with a real knack would be good in its odd games and good in its even games, and the dots would climb from bottom left to top right. They sit in a round blob. Goals for minus goals against behaves differently: a team that outscores the league by a lot one season usually does it again the next. A team’s comeback record carries over next to nothing. Give every one of those 1,496 seasons pure coin-flip luck when it falls behind, and some teams still look like comeback kings while others look like chokers. The real records spread out no further than that.

The Flyers keep turning up, and luck could still explain them

Philadelphia is the one name that keeps turning up. Their 1992-93 team ranks first, with 21 ties in 45 tries. Over 57 seasons the Flyers have drawn level about 43 more times than the odds expected, the most of any franchise. They were above the odds from 1977 to 1997 and again from 1997 to 2017. Since 2017-18 they have been slightly below. Hand all 47 of our franchises pure luck and replay their histories. In about three in ten of those replays, the luckiest franchise ends up this far ahead. So the Flyers may just be lucky. Something about the club may have helped, too. The numbers can’t tell those apart.

A two-goal lead isn’t what it was

• A two-goal lead is less safe than it was in the six-team league. Goals are no rarer.
• A team that keeps coming back is mostly a good team having a lucky year. Next season it will come back about as often as its record says.
• Modern margins overstate how lopsided games were. Check the empty-net goals before reading a 4-1 final as a rout.

The goal that came too late

Detroit’s 16th goal crossed the line a moment after the final horn in 1944, with the Rangers already beaten by 15. Up 15-0, the Red Wings were still coming.

Today a team up two with five minutes left watches the other goalie skate to the bench. Leave out the empty-netters, and the team behind scores 73% of the goals that follow.