Since 1995, home teams have won 53.936 percent of baseball games played in the regular season. Considering that a large portion of home field advantage is, by conventional wisdom, attributed to the effects of home-crowd fans (either on the players, or, more likely, on the umpires), we might reasonably expect the home team to get an even bigger boost during October, when the crowds are bigger, perhaps more partisan, and generally more bananas.
Since 1995, home teams have won 53.986 percent of baseball games played in the postseason. This number is somewhat less reliable because the sample size (552 games) is so much smaller, but if it we conceded it is an accurate representation of the postseason home-field advantage, it would mean that the home teams in the past 18 years of playoffs have won a total of 0.27 more games than they would have won in a regular-season environment. Which is to say, zero games. So, identical.
Surprising and interesting, is all.
(Update: Complicating this is also the fact that "better" teams would probably get to play a few extra home games in October, because of how the playoff rules work. So by a small amount, the home team should be, on average, a bit better than the visiting team in October; in the regular season, the teams are, on average, equal.)
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
If my stubby little fingers pushed the right buttons on the calculator, that 53.986 post season winning percentage comes out to a record of 298-254.
A 59.936 regular season winning percentage would result in a record of 297.73-253.27.
As a team can't actually win .73 games, it rounds up, and comes out perfectly as predicted by the regular season winning percentage.
That is what is commonly known as a rounding error. It is not "still pretty significant" at all.
I think part of what Sam is also saying is that the home field advantage during the playoffs specifically should be amplified when it is not. In general, though, he seems to suggest a home field advantage exists in general. It just doesn't get much of a playoff bump.
1) The visitor can choose pitchers and defenders based on the score. With a one-run lead, they can put in their best pitcher/defense and aggressively substitute based on match-ups. With a tie, they can more strongly consider the impact on future innings.
2) If the visiting team took a lead in the top half, the home team can aggressively use bench players to pinch hit or run, because they know they have to score or lose. If its tied, they know they have to think about additional innings.
Short version: Does 'bats last' necessarily overcome 'pitches/defends last'? What percentage of extra-inning games are won by the home team? Bigger or smaller than the 54% generic home-field advantage?
In the other sports, emotion and ramping up intensity is almost always a goiod thing. In baseball, it is almost always counter-productive. Maybe the heightened home atmosphere degrades performance? Just a thought.
A dedicated analyst could also look at home-field advantage as a function of:
1) Rigor of travel (miles as the crow flies is probably sufficient as a first guess, but some stadiums are hard to get to even after you've landed)
2) Number of days since last travel.
3) Era (Ty Cobb didn't have a charter plane waiting for him.)
4) Number of time zones crossed.
I'd expect that all four are more significant than the +0.27 wins per 552 games of extra postseason "excitement" advantage. #1/#4 are well-documented in the NFL, where east-coast teams do worse than expected when playing night games on the west coast.
Sam links to a review of the book "Scorecasting," which looks into things like this. It attributes HFA to refs being biased to the home team because of the crowd, but that review is skeptical.
Is it possible the amount of scoring has anything to do with it? As far as postseason goes, it seems basketball's the one sport where the better teams usually win series. Baseball and hockey are low-scoring, and their playoffs are a crapshoot (a few years ago, the 5-8 seeds won the first round in the west). So wouldn't luck seem to be a bigger factor when scores are lower? I have no idea if this pertains to soccer playoffs, but they don't play series do they?
From a statistical modeling perspective if the scoring is lower/less frequent than the worse team can win more for reasons similar to the small sample size.
If team A scores a run on average every 0.200 of a game and team B scores a run on average every 0.250 of a game then team A on average scores 5 runs and team B 4 runs and A is a favorite over B. If you imagine that run scoring is a Poisson process with that rate and then simulate games (count how many run events before 1, and if tied the next scoring event wins) you'll see that A will win more of the games, but B will win a decent chunk. If you change that to the same ratio but many more events (so A scores every 0.02 games and B scores every 0.025 games) you have the same ratio of 5:4 but now A scores 50 runs to 40 runs on average and if you run the same sort of simulation A will win many more times.
Scoring in baseball isn't really a Poisson process, since scoring events are linked and correlated, but the model above is a good first order approximation to why teams that are better in sports that are high scoring might well be much more likely to win than "equally better" teams in sports that are low scoring.