Analytics: Home Ice Advantage

As sports fans, we’ve all heard the cliches: “That’s a tough place to play.” “That’s a long road trip.” “We’ve got home ice advantage!” But how important is home ice really? And are certain rinks that much more difficult to play in, or are the teams that call them home just that good? The good news is we don’t have to guess: with decades of game results at our disposal, let’s take a look at the numbers. The source for all the game data is College Hockey News, and the statistics are my original work.

First, just to get it out of the way: since the 1901-1902 season, the home team in college hockey wins 58.9% of the time. There are some problems with this number, however, that I want to explore more fully.

1. Bigger (and usually better) schools tend to schedule more home non-conference games with weaker teams. This is probably not as pronounced as in basketball or football, but it could have still have an effect if the home schedule is favorable to most teams.
2. Every game, even neutral site games, technically has a “home” team and “away” team. These games are included in the number above and really shouldn’t be.
3. Does home ice advantage change over time? Is it worth more or less now than it used to be?
4. Does home ice advantage vary by school? Do some fanbases create a more intimidating environment for their foes?
5. How much does attendance affect home ice advantage? Does the home team win more often when they have more fans in attendance?
6. How much does distance affect home ice advantage? Does traveling across several time zones make it harder to compete?

First, to account for the potential skewing of differing non-conference schedules, let’s look only at conference games. Because conference matchups are almost always at true home sites, this should also filter out nearly all neutral site games. In conference matchups, home teams have won historically at a 56.3% clip. Because of these nice attributes of conference games, we’ll be using this dataset for most of the analysis.

Over Time

Now, let’s look at the home winning percentage over time. The chart below shows the home ice win percentage since the 1960s. Data before this time is too sparse to be of any use to us.

Although there is a lot of noise, this graph appears to have three distinct periods to it. The first period spans from 1963-64 to 1993-94 and very gradually descends. The second period spans from 1993-94 to 2002-03 and ascends fairly rapidly. The third period spans from 2002-03 to the present and descends gradually again. Based on the happenings through college history, I propose the following theory: Home winning percentage tends to decrease over time as schools get familiar with each other and new environments become less intimidating. It’s only during periods of upheaval such as conference realignment or adding new schools to the league that home winning percentage increases. The rising period in the chart above is during the realignment period when many Division II teams were jumping to Division I before Division II was permanently abolished in the 1999-2000 season. During this upheaval, many new schools were getting acclimated to each other, and so home ice advantage increased due to the unfamiliarity with the new rinks the schools were playing in. If you have a better theory, please let me know in the comments.

By School

We can also break down the data by school. Accurately comparing schools proved a bit more tricky. What I ultimately settled on is what I’m (creatively) calling Home Ice Advantage, or HIA. It is calculated by dividing the school’s overall losing percentage by their home losing percentage, and then subtracting one. This can be expressed roughly as the answer to the question, “How much less likely is a team to lose on their home ice than overall?” Other measures I considered diminished the home ice advantage of either good teams or bad teams; this considers all equally.

The graph below shows the HIA for all time (still only considering conference games) for all current D1 schools, with WCHA teams highlighted in color. Keep in mind that home ice advantage only indicates the difference in home performance versus overall performance. If a team has a very well traveling fanbase, for example, that allows them to perform at a high level even when away from home, it would mask some of the home ice advantage effect. Unfortunately, there is no way to control for this effect as a team with a well-traveling fanbase is indistinguishable from a slightly better team without the same fan benefit.

The chart above shows all-time home ice advantage. What if we limited to the period of more interest to most Husky fans: since Mel’s return in 2011-12?

This may at first seem like a bad thing: our Huskies have gotten worse at home since they became respectable again! However, this may be because Huskies fans have traveled very well recently or because MTU has merely increased their away performance, rather than any decrease in home performance. This also confirms what WCHA fans have known instinctively for a while: Mankato, Marquette, and Big Rapids are tough places to play.

By Attendance

If you’ve ever wondered how much a difference you, as a fan, make to the success of your team, the following charts are for you. The first one charts a histogram of HIA versus absolute attendance. The positive effect is there, but there’s a lot of noise.

The following chart shows HIA as a function of the percentage of the arena filled. Here the correlation is much stronger: A home team in a full arena is about 25% less likely to lose than at a neutral site, while a home team in an empty arena is only about 7% less likely to lose. This is the same as the difference between the #3 school (Boston College) and the #54 school (Bentley) in the all-time home ice advantage chart above!

Does it matter how big of arena a team calls home? The answer is a resounding no! The chart below shows absolutely no correlation between arena size and the HIA it provides. It just goes to show that it’s not the size of the arena that matters, it’s how you use it.

By Time Zone

The last factor we’ll look at is time zones. As anyone who has let Dirk’s dulcet tones lull them to sleep during an 11:07 Alaska game knows, time zones are a bear. The chart below shows simple home winning percentage (not HIA) for conference games by time zones. The x-axis is the number of time zones crossed by the visiting team, with positive numbers indicating a western team playing in the east, and negative numbers an eastern team playing in the west. The y-axis is the winning percentage of the home team in those games.

Immediately the 3- and 4-hour data points jump out: this is because they mostly represent the Alaska teams, which have a long history of mediocrity. So that +3 and +4 data points mostly represent Alaska teams traveling to the lower 48 (resulting in very high home winning percentage), while the -3 and -4 data points mostly represent lower 48 teams traveling to Alaska (resulting in very low home winning percentage). For this reason, I decided to exclude those outliers from the trend line. The remaining five data points show a very clearly descending trend; I didn’t quite expect that. I would have expected that the bigger the time change, the more the home team would win. In fact, a team is actually more likely to win if they move to the east rather than stay in their own time zone. This is likely because when a team plays farther east, the game feels early in the day to them, so they have more energy, while to the west the game feels later at night.

To conclude, where a game is played does matter, and we as fans have a much larger role in our team’s success than we might have suspected. While the Huskies may not have the best home ice advantage in the league, we can improve that by packing the Mac for every home series. And someone please tell Suzanne Sanregret to avoid games out west!

Cover image credit Britton Anderson