Home Field Advantage and Extra Innings: Some Continuing Research

Last week at Baseball Prospectus, Rob Mains did some digging into home field advantage and found a very curious effect: home teams did worse in extra inning games than in regular-season games. More specifically, he found that home teams won roughly 54% of games overall but only roughly 52% of extra inning games. There are no two ways about it: that’s strange.

Mains looked into many potential explanations for this discrepancy: team quality, pitcher quality, games that were tied going into the ninth, and various ways of looking at how teams have adapted to the zombie runner era. Today, I thought I’d throw my hat into the ring with a slightly different way of thinking about why home teams are less successful in extras than they are overall.

My immediate thought when I heard this problem was something Ben Lindbergh mentioned on Effectively Wild: home field advantage accrues slowly, and extra innings have fewer innings than regulation. The minimum scoring increment in baseball is one run, naturally. Home field advantage is clearly less than a run per inning; it’s less than a run per game. I like to think of home field advantage as fractionally more plays going the home team’s way. A called strike here, a ball that lands in the gap instead of being caught there, and eventually one of those plays might put an extra run on the board.

Put another way, home teams are getting on base at a .324 clip this year, and away teams are reaching at a .315 clip. Over four batters (an average half-inning), that’s a negligible difference. It’s less than half a baserunner over a full game, even. The difference in slugging percentage is similar. What if extra innings are closer merely because there are simply fewer observations to overcome baseball’s inherent randomness?

To answer this question, I flipped things around. Extra innings are sudden death; what if the overall game was as well? I took inning-by-inning scores for every game starting in 2015 and rewrote the rules of baseball. More specifically, I awarded the first team to end an inning with a lead the victory. 1–0 visitors after one? They win! 2–1 home team after three? You guessed it, the fans are going home happy.

Over roughly 19,000 games, my alternate rules would have slightly lowered home teams’ winning percentage. More specifically, using sudden death rules, the home team “won” 53.2% of games. In reality, they won around 54%. That feels like a point in my theory’s favor: sudden death rules, where ending an inning with the lead ends the game, lead to a lower home field advantage.

You could poke holes in this theory. For one thing, 53% isn’t 52%. For another, teams don’t play to the score in early innings, but they surely would if the game could end after one inning. But I actually think this style of counting overstates a home team’s advantage in sudden death scoring rules. Why? Because we’re starting with the first inning, that’s why. The first inning is the best inning for the home team. Jeff Zimmerman wrote a comprehensive review of potential reasons in 2013, and I’m also partial to Matt Swartz’s look at the situation in 2010.

One thing is inarguable: home teams have more of an advantage in the first inning than they do in subsequent ones. As such, our sudden death sample set isn’t as representative as I’d hope. It has a disproportionate number of first innings, and teams are on far more equal footing in all the other innings. If we’re looking for the true effect of sudden death rules on home field advantage, we should probably remove the first inning from the equation.

To do that, you need to make some rules. I settled on one version, though we’re just slicing up already completed baseball games in different ways and you could always choose to slice them differently. I searched up games that were tied after the first inning, then applied the same sudden death rules thereafter.

That still left me with a sample of nearly 10,000 games. And there’s good news for my theory: in games where the first inning was tied, applying sudden death rules thereafter would have produced a 51.6% winning percentage for home teams. Unfortunately, there’s a countervailing effect: removing the first inning removes some of the home team’s overall advantage. Not only is the game getting shorter, but the home team’s biggest advantage has also already passed, and the game is still tied. In reality, games that were tied after the first produced a 52.5% winning percentage for the home team.

This method of estimating the effect of sudden death scoring rules is clearly imperfect. By using games played under regular rules and treating them as though they were happening in extra innings, I’m introducing some bias. Teams would have played to the score more frequently if a 2–1 deficit after three innings meant the game was over. Presumably, though, both teams would have played differently, and I still think the effect is noteworthy. Reduce the amount of innings in a game, and you’re reducing the home team’s advantage.

Consider this: if we played 50 inning games of baseball, home field advantage would presumably be meaningfully higher. Home teams score 4.76 runs per nine innings, and visiting teams score 4.50 runs per nine. That difference is meaningful — it works out to a .540 winning percentage, as we already discussed — but one home run, or one runner scoring instead of being tagged out, overwhelms that difference. Over 100 innings, that scoring advantage would be one and a half runs, and a break would give the visiting team a victory less often. If games were 500 innings, no one would watch, but the home team would be favored by 15 runs in every game, assuming teams were otherwise equal.

Is this the only explanation for what’s going on with home field advantage in extra innings? Certainly not. It could come down to tactics, or pressure, or some kind of selection bias; the teams who reach extra innings might be built in a certain way. I didn’t take the time to rule all of those out, though Mains did a great job in his investigation of the phenomenon. I do think that my explanation is a big part of the overall answer, though. There are plenty of potential reasons why home field advantage is less pronounced in extra innings games. As best as I can tell, a good chunk of that effect comes down to the very rules of extra innings themselves, rather than anything the teams are doing about them.