There’s No Place Like Home — The Power of Home-Field Advantage
In soccer (as in many other sports), home-field advantage is widely recognized as a significant factor in team performance. It might be the familiarity of playing in a field that a team has practiced in, the tiresomeness of travel, the contagious energy of the cheering crowds, or just the referee’s involuntary bias. Regardless, playing in the home-field seems to activate some kind of wizardry which pushes a team to win.
It’s no surprise then, that predictive models in sports often take this effect into account in some way, e.g. calculating team strength separately for home and away matches. And the good thing is, to improve the performance of a model, we don’t necessarily need to know why home-field advantage happens, we just need to know how much it matters for whom.
So how much does home advantage actually matter? What does it look like for different leagues, over time, and for different teams?
By League
Of the thirty major leagues analyzed, home-field advantage seems to be the most significant for the MLS, where home teams have won more than twice as often as away teams in matches played from 2008 to 2018. This is in line with analysis done by the MLS using expected goals.
A chi-squared test for each league also show the difference as significant, with p-values under 0.05 for all thirty leagues.
Home-field Advantage Over Time
How has home team advantage changed over time? Narrowing the leagues from above even further to a few European leagues and one North American league, there doesn’t seem to be any discernible pattern. This is with the exception of the Italian Serie A, for which the home team advantage seems to be declining.
By Teams
The English Premier League falls in the middle of the pack of the leagues analyzed, with home teams being 1.6 times more likely to win. By looking at the ratio of home performance over away performance, we can see the home-field effect on wins and draws (a ratio of 1 means more likely to win/draw at home, while a ratio of under 1 means less likely to win/draw at home).
Although home-field advantage definitely has an effect on wins, it also seems to have an effect on increasing the probability of wins AND draws for some teams, which means that they’re less likely to lose altogether.
Effect on Number of Goals Scored
How does home-field advantage affect the number of goals scored? Again zoning in on the English Premier League, teams playing at home seem to score a higher number of goals, and are slightly more likely to score a very high number of goals.
Looking at the goal difference for home and away matches, home matches are likely to have a slightly higher number of goals than away matches — which makes sense considering that home teams win more often.
Team Strength and Home-field Effect
Are teams that are more affected by home team advantage better, or worse teams? To answer this, we can take the Elo Ratings of each team as an estimate of team strength, and plot it with the home team advantage index.
This analysis uses teams from the same 30 leagues as the first analysis above, filtering out teams that have less than 100 matches.
Although there doesn’t seem to be a conclusive pattern, it does seem like that better teams (towards the right side of the graph) are less likely to be highly affected by home-filed advantage, while the deviation is much higher for teams with lower ratings.
There’s no place like home — with a better understanding of home-field advantage in soccer, we can seek ways to account for its effect in creating predictive models for the sport. Considering that the effect changes over time and varies across different teams, it may make sense to dynamically adjust for these two factors (teams and time) in both match outcome prediction and score prediction.
See the repository for this story here
