clock menu more-arrow no yes

Filed under:

One-Goal Games

New, comments

Originally appeared in Hockey Prospectus, June 18, 2009:

In baseball, you often hear that a team's performance in one-run games is all luck. After all, while overall winning percentage is highly-correlated from year-to-year, it varies wildly in one-run games. Winning teams - Atlanta and the Yankees made the playoffs more than ten years in a row - tend to blow out their opponents and not have particularly good records in one-run games. In hockey, it looks no different: from 1979-80 to 2003-04, the year-to-year r^2 of overall winning percentage was 0.37, while it was 0.05 in one-goal games.

However, you would not expect a team's record in one-goal games to be uncorrelated with its overall record in the NHL. Offensive levels have been so low since the mid-1990s that the value of a single goal is much higher than a single run in baseball, which has double the scoring of hockey. Indeed, at 2.5 goals per game (2003-04 NHL season), almost half of NHL games were decided by one goal, while at 4 goals per game (1982-83), just one-third were one-goal games.

Let's look at the annual correlation between overall and one-goal-game winning percentage plotted against the NHL's scoring levels:



While the trend is certainly downwards as scoring increases, several points confirm that it's possible to have both high-scoring and a high correlation between overall performance and performance in one-goal games. As scoring levels rise, the number of wins each team gets in one-goal games decreases, but there was no disincentive for them to play to win these games.

Unlike baseball, it's possible to suppress the other team's offense in hockey by focusing on defense at the expense of offense - the same is true in football and soccer. However, in baseball, when the opposing team is batting, you're already completely focused on defense. In a one-run game, you can gain a slight advantage by using defensive replacements in the field and a lights-out relief pitcher, but the overall impact is minimal if relievers can only go one inning and your defensive replacements can't score you a run if you need one.