Examining Home-Field Advantage Phil Birnbaum www.philbirnbaum.com Home Field Advantage (HFA) In baseball, home teams generally win 54% of games Why? Several possible explanations Possible causes Fan enthusiasm Familiarity with park Molding team to park Travel "Home Cooking" Umpires Batting Last Others Can eliminate some Fan enthusiasm Familiarity Nothing significant Batting Last Only a little Travel Nope. Seems to be no correlation between HFA and attendance Not just in close games Pitching last also has its advantages See "The Diamond Appraised" and "Scorecasting" Recently: Umpiring "Scorecasting," by Tobias Moskowitz and Jon Wertheim, released early 2011 Claims refereeing/umpiring is the true cause of HFA Lays out some evidence for several sports "Scorecasting" on umpiring Umpires call more strikes for home team pitchers than for visiting team pitchers The higher leverage (clutchier) the situation, the bigger the effect Umpires actually favor the visiting team in less-important situations "Scorecasting" on umpiring Authors claim umpiring/refereeing accounts for almost all of HFA But … doesn't mesh with other evidence of the incidence of HFA HFA appears in all situations For instance: When one team is ahead by 4+ runs early, that's low leverage. But HFA remains high Home team outscores visitors regardless Inning Overall When one team has 4+ run lead 1 +18% more runs 2 +10 +19 3 +11 +10 4 +8 +13 5 +10 +9 6 +8 +8 7 +7 +8 8 +6 +5 Umpires and leverage Mitchel Lichtman finds some effect of clutchness on called strike HFA, but less than "Scorecasting" found Me too Other estimates are lower John Walsh, "The Hardball Times Annual 2011" J-Doug, "Beyond the Box Score" Home team favored by 0.8 pitches per game Accounts for one-third of HFA Accounts for one-sixth of HFA Dan Turkenkopf, "Beyond the Box Score" Accounts for one-eighth of HFA Not just umpires? There might be other things going on, not just umpires How can we find out? Look at things that don't involve umpires Like what? Maybe … fielding. Once a ball is in play, umpires don't control whether it's a hit or an out If defenses turn more balls into outs at home, would that show that HFA is more than just umpiring? No, not really. Because … Compensating for umpire bias … if umpires are more lenient towards home batters, they get more balls and fewer strikes More favorable counts Better hit balls Harder to field those balls So it could be umpiring after all! Compensating for umpire bias Similarly in other sports In hockey, the referee's main influence is in calling penalties Home teams outscore visiting teams even at full strength But, it could be because visiting teams have to play less aggressively out of fear of referee sensitivity Same for soccer, basketball, etc. How to tell? Need a measure of HFA that is not influenced much by umpires How about wild pitches and passed balls? Objective decision: ball eludes catcher, runners advance WP/PB home field advantage For 2000-2009 (God Bless Retrosheet): Home teams: Road teams: 539 WP+PB per 100,000 pitches 557 WP+PB per 100,000 pitches 3.3 percent difference Statistically significant (barely) WP/PB home field advantage On 0-0 counts only Home: 511 WP+PB per 100,000 pitches Road: 544 WP+PB per 100,000 pitches Even larger effect WP/PB home field advantage Overall difference: about 5 wins over 10 years Works out to .0002 wins per game That's 1/200 of HFA Runs resulting from WP+PB are 1/50 of total runs Not perfect, but reasonable Basketball Free throw shooting percentage is not subject to referee bias HFA in foul shooting is about 0.2 percentage points in favor of the home team Difference of 120 points a year 0.1 points per game Overall HFA is 3 points per game Again, seems reasonable Speed skating There is a home-field advantage in speed skating! Of course, that can't be because of refereeing What could it be? Another theory For some biological reason, humans just perform better at home than on the road Testosterone levels: "players may have tapped into a primal instinct to defend their own territory." But, whatever: some intrinsic evolutionary reason is plausible, because HFA is pervasive and universal And none of the other hypotheses have tested out Competing guesses on HFA "Scorecasting" Craig Wright, "The Diamond Appraised" (1989) Almost 100% umpires. 5% crowd … 5% last AB … 10% familiarity … 10% shaping team to park … 30% home cooking … 40% umpires. Me (among others) 10% umpires … 80% intrinsic/testosterone … 10% other. Where to go from here? Need to come up with ingenious ways to decouple umpiring from other factors If you can think of any, let me know For now, I think there's good evidence that umpiring is only a small part of what's going on Acknowledgements Thanks to readers at my blog for many comments, suggestions, and references on this topic Mike Fast was especially helpful, pointing me to some of the studies mentioned here. Thanks, Mike.