This two-part series was written by Pregame Skate contributor Thomas Pepin.

How strong is the correlation between the sequence of wins and winning a series?

The Study: As someone who bets on NHL outrights, after having many of my preseason favorites eliminated from postseason contention (see Exhibit A), I was pretty nervous about what my return (if any) would look like at the end of these Playoffs. Even worse, as the Playoffs have progressed, some of my replacement favorites have also been eliminated (see Exhibit B). So now, with so few of my favorites left and my returns completely in the air, I’ve really been getting into the weeds to try to calm my nerves as the Stanley Cup Final approaches. And here’s where this study comes in.

On many occasions over the years, a team I was hoping would win has been down in a series. Usually, I’ve been less concerned since I know that teams have plenty of games to come back. However, given the situation described above, I thought to myself: Wouldn’t it be nice to be able to anticipate how various series will typically progress?

To look into this question, in this two-part series, I’ll examine the correlation between how a series progresses (e.g. Win Game 1, Lose Game 2, Lose Game 3, Win Game 4, etc. and all other variations) and the series outcome from 2013-14 to 2018-19. In Part No. 1 below, I’ll look at each series length—four, five, six, and seven—separately. In Part No. 2, I’ll cover the data more broadly and offer some discussion and conclusions.

The Results: At each game within each series length, I calculated the correlation between the sequence of wins and the series outcome.

For series that end in four Games (22 of 180 teams), unsurprisingly, the correlation was 1.00 (i.e. perfect) at all points since we know that the team that wins any game will ultimately win the series.

NHL @NHL

No quit. 😤 #Game7 | #StanleyCup

May 31st 2022

161 Retweets1,666 Likes

For series that end in five Games (32 of 180 teams):

• The winner of Game 1 had a 0.250 (weak) correlation with who ultimately won the series

• The sequence of wins through two Games had a 0.735 (strong) correlation with who won the series

• Through three Games had a 0.796 (strong) correlation with who won

• Through four Games had a 0.938 (strong) correlation with who won

• 8 of 16 series winners won Games 1 and 2, and 6 of 16 won Games 2 and 3; the remaining two won Games 1 and 3

• Two won 1,3,4; three won 1,2,3; five won 1,2,4; and six won 2,3,4

For series that end in six Games (68 of 180 teams):

• The winner of Game 1 had a 0.529 (moderate) correlation with who ultimately won the series

• The sequence of wins through two Games had a 0.359 (weak-moderate) correlation with who won the series

• Through three Games had a 0.383 (weak-moderate) correlation with who won

• Through four Games had a 0.431 (moderate) correlation with who won

• Through five Games had a 0.822 (strong) correlation with who won

• 26 of 34 series winners won Game 1 and 18 of 34 won Game 2

• 26 of 34 won at least two of the first three Games

• After four Games, only 13 of 34 were up 3 Games to 1

• Only 2 of 34 came back from losing the first two Games and all sequences other than wins in 2,3,4 (2 of 34); 2,4,5 (1 of 34); and 3,4,5 (2 of 34) had essentially the same frequency (between 3 and 5)

For series that end in seven Games (58 of 180 teams):

• The correlation after Games 1, 3, 4, 5, and 6 were essentially zero

• The correlation after two Games was only 0.179 (weak), and only 4 of 29 came back from losing the first two

Discussion: First, it looks like if a series goes to seven Games, no sequence will give you an indication of who will succeed; according to this study, Game 7 is essentially anyone’s game. For the most part, the same can be said about six-game series, excluding after five Games. However, since no one has a crystal ball to know that a series won’t go to a seventh Game, the strong correlation with sequences after five Games in six-Game series might not tell you much at all.

Second, and again unsurprisingly, securing early wins appears to matter most across the board. As alluded to earlier, I didn’t think this would necessarily be the case, but after running the numbers, it appears to be so. That being said, if you can stay in a series early on—splitting the first two or four—most teams appear to have a good shot.

But again, all of these results come under the assumption that we know how many Games each series will go to since they were categorized in that fashion. While this gives us a clearer picture of which sequences work for a given series length, what will the numbers look like when we don’t know how long a series will go? Stay tuned for Part No. 2 to find out this and more.