If you watched Game 3 of this year’s Stanley Cup Final, you saw something that looked impossible. Vegas led Carolina 4-0, deep into the third period. Then Carolina scored. Then again. Then again. Three goals in 39 seconds, a record. The building came apart, and every person watching had the same thought at the same moment: the Hurricanes have the momentum now, and Vegas is about to fold.
Vegas won the game in double overtime.
That is the problem with momentum. We feel it so strongly that we forget how often it lies to us. The comeback that stalls. The team that cannot be stopped and then gets stopped. We remember the runs that finished and quietly forget the ones that fizzled, and we end up believing in a force that may not be there.
So we tested it. We pulled every goal from the 2026 playoffs, 477 of them across 80 games, from the NHL’s public play-by-play, each tagged with its exact time and strength state. Then we asked one question. When a team scores, does it become more likely to score again in the minutes that follow, the way the word momentum promises? Or does the next goal behave like an independent event that memory later stitches into a story?
After a Goal, Who Scores Next?
Start with the simplest version. After a goal, who scores the next one?
If scoring builds momentum, the team that just scored should take the next goal well over half the time. It does not. Across the 397 cases in these playoffs where another goal followed in the same game, the team that scored last also scored next 51.4% of the time. The 95% confidence interval runs from 46.5% to 56.3%, and a binomial test against a 50/50 split returns p = 0.62. In plain terms, that is a coin flip. We cannot distinguish it from chance.
This is worth sitting with, because it is not what the sport tells you. The team riding the wave, feeding off the crowd, tilting the ice, by the time the next goal is scored, it is the other team about as often as not.
The Two-Minute Window
The obvious objection is that momentum does not mean the rest of the game. It means right now, the next two minutes, while the building is loud and the scoring team is supposedly surging. So we narrowed the window to goals that were followed by another goal within 120 seconds.
In that two-minute window, the team that just scored gets the next goal 44.8% of the time (n = 67, 95% CI 33.5% to 56.6%). That is below half. If anything, the moments right after you score look slightly more dangerous for you, not less, which would fit a team relaxing for a shift while the other side answers hard. We will not lean on that, the confidence interval is wide and the binomial test is not significant (p = 0.46). But the claim momentum actually makes, that scoring makes your next goal more likely in the immediate aftermath, is simply not in the data. The point estimate sits on the wrong side of 50.
Stretch the window to three minutes and it is 50.0% on the nose (n = 102). Walk it out bin by bin, zero to one minute, one to two, two to three, and so on, and nothing climbs meaningfully above 50 or stays there. There is no decay curve to find because there was no peak to decay from.
But it Was a Power Play
There is a fair objection here, and the honest version of this question has to take it seriously.
Many goals come on the power play, and a power-play goal usually ends the advantage that produced it. The penalty expires, the other team returns to full strength, and sometimes they get a power play of their own. So part of the other team scoring next is not about momentum at all. It is the structure of the game resetting. That mechanism would drag the scoring team’s next-goal rate down for reasons that have nothing to do with psychology.
So we stripped the power-play goals out and looked only at even-strength goals, five skaters a side, nobody handed an advantage. This is the clean test.
At even strength, the team that just scored gets the next goal 52.7% of the time (n = 279, 95% CI 46.8% to 58.5%, p = 0.40). Still a coin flip. For comparison, after a power-play goal it is 48.0%, consistent with the structural-reset explanation rather than a momentum one. Taking out the special teams does not rescue the idea. When the teams are even and no one has an edge handed to them, scoring a goal tells you almost nothing about who scores the next.
That is the quiet finding. The momentum we narrate so confidently, the swing, the surge, the wave, mostly is not there once you line up the goals and count.
Why it Feels So Real
If the effect is not in the data, why is the feeling so strong? Because some comebacks do finish, and those are the ones that stay with us. Carolina’s three goals in 39 seconds is unforgettable. The far more common sequence, a team scores, presses for the next one, and gets nothing, is forgotten by morning. Of the 151 goals in these playoffs that were followed by another within five minutes, the other team answered 49% of the time. Nearly half. We keep the highlight and discard the rest, then mistake the highlight for the pattern.
It is worth being precise about what we are not claiming. We are not saying nothing matters after a goal. Power plays change scoring odds a great deal. Pulling the goalie changes them. A kill at the right moment changes them. Those are structural and visible in the rules. What we cannot find is the loose, mystical version, the idea that the act of scoring lights a fire that makes the next goal more likely. That version looks like a coin flip wearing a narrative.
The One Thing About This Final That is Real
Here we have to be careful, because it would be easy to overreach. This specific Final has felt wilder than usual. Vegas blew a 4-0 lead and a 3-1 lead. Every game has been tied at some point in the third period. The series is full of the swings that make people shout about momentum.
We will be straight: four games and 33 goals is far too small a sample to say anything reliable about momentum inside this series alone. Anyone telling you the data proves Carolina or Vegas has the momentum heading into Game 5 is selling the same illusion in a playoff jersey.
But one thing about this Final is real and measurable. It is a higher-scoring series than the rest of the playoffs, by a clear margin.
The Final has averaged 8.2 goals per game. The rest of the 2026 playoffs averaged 5.8. A two-sample t-test puts that gap at p = 0.04, unlikely to be chance. So when people say this Final has been wild, they are right, but for a reason simpler and more honest than momentum. There are just more goals. More goals mean more lead changes, more comebacks, more moments that feel like a tide turning. The chaos is real. The cause is volume, not magic.
Watch it Happen in Game 5
Carolina took Game 5 by a score of 4-2 to lead the series 3-2. The lead held this time. In Games 3 and 4 it did not, Vegas coughed up a 4-0 lead and a 3-1 lead. Same series, opposite outcomes, and the broadcast called all of it momentum. The data says none of it was. A goal changes the score. It does not change the odds of the next one the way the word momentum promises.
If you want something real to hold onto, forget who has the momentum. Watch the power plays, watch the goalies, watch whether this stays an 8-goal track meet or finally tightens into the defensive series everyone predicted. Those move the odds. The surge after a goal, the wave, the sense that the ice has tilted and will not tilt back, that is the momentum illusion. It is a strange truth about hockey that the moment which feels most certain is the one the numbers refuse to confirm. The comebacks are real. Momentum is the story we add afterward.
Data: NHL public play-by-play, 2026 Stanley Cup Playoffs, 477 goals across 80 completed games, each with period, time-in-period, and strength state (even, power play, short-handed, empty net) decoded from the situation code. For every goal followed by another goal in the same game (n = 397), we recorded whether the next goal belonged to the same team and the elapsed time. Same-team-scores-next rates are reported as proportions with 95% Wilson confidence intervals; departures from 0.50 were tested with two-sided binomial tests, and none reached significance, including the two-minute window (44.8%, p = 0.46) and the even-strength subset (52.7%, p = 0.40). Power-play goals were separated from even-strength goals to isolate structural reset effects from momentum. Time structure was examined in 30- to 120-second bins from 0 to beyond 10 minutes; no bin departed significantly from 0.50. The Final’s scoring volume was compared against the rest of the playoffs with a two-sample t-test on goals per game (8.2 vs 5.8, t = 2.12, p = 0.037). The Final is reported for scoring volume only; with four games it is too small a sample for any momentum claim. Interactive figures built with Plotly.js.