Friday, March 4, 2011

Home Field Advantage: What You See Depends On Where You Look (And What You Make Of Draws)

A friend recently recommended I read Scorecasting: The Hidden Influences Behind How Sports Are Played and Games Are Won, a new book on numbers and sports (with a focus on decision making) by Tobias Moskowitz and Jon Wertheim. While there isn’t a lot about soccer in the book (it's mostly aimed at the American sports fan of basketball, baseball, football, and to some extent hockey), it’s a fun read. And what there is about soccer is actually really interesting. I recommend it to anyone interested in sports, numbers, and how the people involved in them - players, coaches, and referees - make decisions.

An important part of the book consists of a couple of chapters on the home field advantage in sports, including basketball, hockey, American football, and, you guessed it, soccer. In these chapters, Moskowitz and Wertheim first document that the home field advantage is ubiquitous across different team sports, as well as across leagues and over time. They also argue that most of the advantage home teams seem to have stems from referee decisions, rather than things like travel, schedule, weather, or any other number of factors that have been proposed over the years.

So far so good, but there was one thing about it all that made me want to say "hmm"...

The book has a table that shows the home field advantage to be particularly pronounced in soccer and, according to the authors, across leagues and years. Since we all seem to know about the home team advantage, as I was reading along, I was more interested in the explanation for why it exists than how big it really is. But what stuck with me was the finding that home teams win well over 60% of their matches, a number I initially didn’t pay a whole lot of attention to.

But then I started wondering how this could be true. It seemed to go against everything I remembered about the typical match outcome. So, just to make sure I wasn't hallucinating, I cranked up ye olde dataset from the past five years to crunch a few numbers. For starters, I wanted to know how many matches end in wins and how many end in draws. Turns out, on average, across the leagues for the past five seasons (2005/06 to 2009/10), 26% of all matches end in a draw, which means that 74% of all matches end in a win for one side or the other. So, the odds of having a winner in a match are .74, or basically three quarters of the time. Could it really be that home teams win over 60% of their matches when only 74% end in a win? In other words, do away teams win only slightly more than 10% of the time?

The simple answer is no. Take a look at this distribution of match outcomes.

It's true that more matches end with the home team winning than any other outcome. The numbers vary slightly across leagues, but they're nowhere near 60+%. Instead, around 45-48% of matches end in a home win, and the remainder are roughly evenly split between wins for the away team and draws. So the home team wins about as often as it does not.

So is Scorecasting simply wrong? Well, it depends on how you slice and think about these numbers. Yes, on its face, the finding that home teams win 60% of their matches is incorrect. They win slightly fewer than 50%. But this still begs the question: Do they have an advantage? The answer to that question is also a simple yes. While you wouldn't bet quite as much on them as you might after reading Moskowitz and Wertheim's entertaining book, you should still bet on them if the question is who will win more of the matches that end in a win. Confused? Let me clear it up.

If we forget about draws - that is, throw away about a quarter of all matches - then the magical 60+% re-appears. If you calculate the proportions of matches with a winner that are won by the home side or the away side, you'll see that the home team wins over 60% of those, while the away team wins fewer than 40%.

Does this mean Scorecasting is misleading? I wouldn't go that far, but I would say that it's incomplete. The numbers don't tell the whole story for soccer where, unlike basketball, baseball, or football, we have ties. And ties are important because they generate points. Yes, home teams win more and away teams lose more, but sometimes they come out even. And this means that there is less of a home field advantage in soccer than you might think.