We know that baseball has a home field advantage. The advantage has changed over time, with home teams in the .530-.540 win-percentage range in recent years.
What’s responsible for home-field advantage? Ad hoc explanations abound – road-weary players under-performing due to fatigue or lousy bedding, fielders and pitchers accustomed to the particular of their particular ballpark, umpires playing to the crowd, the emotional impact of the fans. But if we’re interested in why home-field advantage exists, we need to know what it’s composed of.
Obviously, home-field advantage is in part a park-specific phenomenon; the home field advantage for the Rockies will differ considerably from the home field advantage of the Marlins. In addition, it’s easy to measure the value of home field advantage, but it’s much more difficult to decipher how much of an observed home field advantage came from leveraging the particularities of a park’s features by acquiring players who fit the park best. The distinction between observed home field advantage in the past and likely/potential home field advantage in the future is crucial.
In order to determine the home field advantage created by each individual park, we should first determine the typical home field advantage and then express each park’s advantage relative to the average advantage. I’ll be looking at the effects of each individual park in future articles; the purpose of this article will be to establish the average home field advantage.
Using STATS batted ball data, interleague play excluded, I’ve compiled the observed home field advantage factors for 1999-2004. I used the same methodology as in my Dodger Stadium article, so if you’re not sure what any of this means, check that out. The only difference was that I put $GB ahead of $LD in the events chain this time.
These are the numbers as expressed from the point of view of the home team’s batters – in other words, how many times the home team’s batters are hit by pitches divided by how many times they were hit by pitches on the road.
$HBP $foul $BB $K $gb $ld NL 1.016 0.984 1.070 0.945 0.992 1.014 AL 1.068 0.987 1.079 0.960 0.992 0.996 Groundballs out $h $xb $3b $roe NL 0.985 1.047 1.078 1.063 1.016 AL 0.989 1.033 1.037 1.019 1.028 Flyballs out $hr $h $xb $3b $roe NL 1.000 0.987 1.007 0.975 1.099 1.097 AL 0.999 1.013 0.988 0.995 1.158 1.182 Linedrives out $hr $h $xb $3b $roe NL 1.015 1.054 0.994 1.051 1.161 0.972 AL 1.020 1.114 0.993 1.007 1.172 0.778
And because the distinction between line drives and fly balls gets hairy, check out all airballs:
Airball out $hr $h $xb $3b $roe NL 0.998 0.991 1.005 1.019 1.114 1.064 AL 1.004 1.026 0.989 1.003 1.164 1.075
Home teams are more likely to reach base on groundballs, and they’re also more likely to turn any hit into a double or any extra base hit into a triple. In the NL, airballs actually saw a tiny homefield disadvantage in terms of home runs, but that was offset by slightly fewer outs. In the AL, it was reversed, as some extra home runs were offset by extra outs. There also was an advantage on foul outs, but it’s not really significant.
The most overwhelming factor, however, is walks and strikeouts, as has been observed previously. Home teams are much more likely to draw walks and much less likely to strike out.
The effect of walks and strikeouts clearly trumps the other effects of home field advantage in significance. I used Base Runs to estimate run scoring for the NL and AL home and road offensive numbers. Combining Base Runs and outs, I then used Pythagopat to determine the resulting win percentages for the home teams. To demonstrate the relative impacts of the various components, I grouped the effects into four categories – groundballs, airballs, balls and strikes, and foul outs – and saw what the home team’s win percentage would look like if every other effect was neutral. In other words, if home field advantage affected only groundballs and left airballs, balls and strikes, and foul outs unaffected, how often would the home team win a game between two completely equal teams?
'99-'04 NL W% AL W% No Change 0.5364 0.5279 Groundballs 0.5115 0.5085 Airballs 0.5025 0.5001 Balls and Strikes 0.5219 0.5192 Foul Outs 0.5004 0.5004
Since we’re just using Base Runs, baserunning isn’t included (although James Click looked at that angle for Baseball Prospectus). The impact of foul outs is pretty minimal, though consistent. In the NL, airballs had some impact while in the AL, they had none. The impact of groundballs was a big deal in the NL, but fairly small in the AL. The impact of balls and strikes reigned supreme, constituting about two thirds of the total home field advantage. Here too the NL saw a slightly bigger home field advantage.
These results certainly had me wondering why the NL saw a bigger home field advantage. To see if the variability in the sample was obviously responsible, I checked the numbers for 1999-2001 versus 2002-2004:
'99-'01 NL W% AL W% No Change 0.5304 0.5243 Groundballs 0.5095 0.5051 Airballs 0.5028 0.4992 Balls and Strikes 0.5187 0.5187 Foul Outs 0.5004 0.5004 '02-'04 NL W% AL W% No Change 0.5426 0.5316 Groundballs 0.5147 0.5109 Airballs 0.5023 0.5010 Balls and Strikes 0.5252 0.5195 Foul Outs 0.5003 0.5003
Looks like it’s probably partially a sample issue. However, it may also have something to do with the NL having more parks and more extreme parks. It’s also interesting the extent to which home field advantages have risen pretty much across the board in comparing these two three-year periods. I won’t venture to guess why at this time.
The groundball advantages – as well as the advantages for extra base hits, triples, and foul outs – can be explained pretty intuitively, as fielders from the home team are more familiar with the environment of the ballpark. Since there does not appear to be a substantial or universal trend in the number of airballs that are caught or go for home runs, it’s unlikely that the issue is that fatigued visiting ballplayers aren’t hitting the ball as hard. This is probably an issue of familiarity.
The balls and strikes, on the other hand, are more difficult to explain. Are umpires really helping out the home team? Are pitchers more familiar with the pitching mounds at home? Are visiting batters having a harder time seeing the ball or reacting to it? Is fatigue affecting the results?
It may be significant here that the total number of balls in play (bip) per plate appearance (pa) does not differ much – the home/road factor for bip/pa was 1.0079 in the NL and 1.0002 in the AL. That leads me to believe that the major change in walks versus strikeouts is related to judgment of whether a pitch is a ball or strike – either the batter’s poor judgment, induced by lack of familiarity with the hitter’s backdrop, or the umpire’s judgment, giving better calls to the home team. Whether that speculation is accurate, I’m not in a position to say.