By Brendan Kent
A while back, I wrote a piece on travel distance in Major League Soccer. It got me thinking about which aspects of travel affect a team’s performance. It would seem that as a trip progresses, travel might take an increasing toll: the more days on the road, the more days away from home, the greater the impact. Essentially, the question boiled down to: do teams perform worse as a road trip progresses? While this analysis could be run on the NBA or the NHL as well, I decided to start with baseball, where long road trips are the norm.
Using data from the 2014 MLB regular season, I examined the effect of days away from home on wins, runs scored, and runs allowed.
Using logistic regression, with wins for the away team as the dependent variable and cumulative days on the road as the independent variable, I tested several models before finding that a cubic model was statistically significant at the 90% confidence level (and almost significant at the 95% level) for the variable. The trend is interesting…
The model shows a very slight increase in win probability initially, followed by a decrease, followed by another increase. This trend is somewhat similar to what Matt Swartz reported in a 2009 piece for Baseball Prospectus. Swartz calculated the winning percentages of away teams by cumulative road trip games, the results are below:
While the numbers don’t line up exactly, in both studies away teams seem to win most frequently in the 2nd, 3rd, and 4th road games. As Swartz notes, “It does seem that road teams may do a little better early in a road trip than late, but only by a small margin.” While my model shows an increase in win probability for the 11th game of a road trip, an 11-game road trip is relatively rare (only 12 times in 2014 did an MLB team have an 11 game road trip) so it’s difficult to trust the model at that point.
Runs Scored and Runs Allowed
A natural question is whether offense, defense + pitching, or a combination of both is being affected by the longevity of the road trip. To answer this, I made runs scored the dependent variable and again fitted a model using cumulative road trip days as the independent variable.
A simple linear model fits very precariously with a p-value of .118 for the variable. This model tells us that on average, away teams will score .038 fewer runs for each additional day on the road. However, we can’t take much away from this, as the r-squared value is .001.
Again, I attempted to fit a model with cumulative road trip days as the independent variable, this time using runs allowed as dependent variable. None of the models had a p-value lower than .2, suggesting that defense is not affected.
What can we take away from this? Well, not much at all. It seems that cumulative road trip days perhaps have an extremely slight effect on away team performance. And this seems to be caused by offense. But this explains very, very little of the variation in away team performance. The length of a road trip may take a toll on players, but on a game-to-game basis, the players barely show it.