April 13, 2016

Prospectus Feature

MLB Can't Beat the Weather

by Samuel Mann

On Friday afternoon, I attended the Mets home opener at Citi Field, undoubtedly a great and celebratory day for Mets fans (as a Phillies fan, it was just another day in the countdown to J.P. Crawford, Nick Williams, et al.). But it was also mid-40s with a stiff breeze and plenty of cloud cover. Saturday’s forecast called for 38 and rain, so even though I planned on going to the game, I ended up staying home. As these weather patterns are not uncommon for April in the Northeast or Midwest, the friend I planned to go to Saturday’s game with asked why MLB even schedules early-season games in northern climates rather than just starting the season in warm weather stadiums or domes. We were not the only ones wondering. So let’s look at whether it would even be feasible to avoid scheduling home games for cold climate teams and reasons why it likely will not happen.

Depending on how one chooses to define “cold climate,” there are 12-14 teams for whom April weather poses a significant risk to home games. I defined “cold climate” as any city with an average high of 60 or below or an average low of 45 or below, and then added Washington to the list out of an abundance of caution (Saturday’s game in Washington was postponed, so perhaps this is recency bias). Admittedly, it’s not a perfect sample and this weather data contains averages for the whole month, whereas for our purposes we are probably only worried about the first two weeks of April, but it provides a decent starting point. Over a third of the league would be affected by such a proposal.

City	Average April Temperature (High/Low)	City	Average April Temperature (High/Low)
Baltimore	65/43	Detroit	59/39
Boston	56/41	Minnesota	58/37
Chicago (Cubs)	59/39	New York (Mets)	61/45
Chicago (White Sox)	59/39	New York (Yankees)	61/45
Cincinnati	66/42	Philadelphia	64/44
Cleveland	59/40	Pittsburgh	62/40
Colorado	61/34	Washington	67/47

Temperature statistics from 1981-2010 climate data compiled by NOAA National Climatic Data Center, available here.

Any realistic proposal would call for the above teams to play at least the first two series on the road. But would one week really make a significant difference? Three series would seem to be a more effective proposal, delaying the first cold weather games to mid-April most years. This year, of the 14 teams listed above, five began the season at home, five played their second series at home and the other four played their third series at home. Seven of 14 played two of their first three series at home. Such a schedule would require changing at least 10 matchups in the first week and at least 10 more the next week to accommodate a three-series delay.

Assuming the proposal is three series, cold climate teams would be on the road for at least 10 days to begin the year. The Home Opener is a major event almost everywhere—how would Mets fans have reacted to having to wait 10 days into the season to see the NL Champs banner raised? Asking 12 to 14 teams to open up with a lengthy road trip year after year is probably not something cold weather teams would be willing to entertain. Nor would warm weather owners want to have over 10 percent of their home games come before Tax Day, when attendance everywhere is suppressed by a variety of factors.

The other uncertain factor is whether there would be a competitive balance impact. Some studies have shown slight but mostly insignificant fluctuation in winning percentage over the course of long road trips, but it is not difficult to believe there could be a negative effect to a scheduling arrangement that increases the number of lengthy homestands and road trips.

The most significant factor to any schedule change would be attendance and the resulting revenue impact. Games during the summer months are simply more valuable for a number of reasons: better weather, less competition from other sports, pennant races beginning to emerge and school out of session, making it easier for families to go to weekday games. Texas, for example, would presumably not like to trade six July games for six April games (last year Texas drew 5,000 more fans per game in July than April). Not every team has this significant of a discrepancy, but almost all teams draw better in July or August. April, conversely, is on the lower end.

	2013 Average per Date	2014 Average per Date	2015 Average per Date
March-April	28,822	28,911	29,302
May	29,763	29,378	29,502
June	31,328	31,128	31,125
July	32,399	32,168	32,882
August	31,501	31,312	31,637
September	29,269	29,926	28,423

2013 and 2014 figures courtesy of Number Tamer 2014 MLB Attendance Report. 2015 figures from 3^rd Line Grind team-by-team infographics.

Of course, there are some team-by-team exceptions that demonstrate the multitude of factors that affect attendance. In 2015, the Dodgers were the only team for which April was the highest drawing month, but the Dodgers averaged more than 45,000+ fans every month so the difference was insignificant. Oakland and Philadelphia, two teams who suffered over 10 percent declines in total attendance from 2014, achieved their highest gate receipts in May. Two others locked in pennant races, St. Louis and Toronto, saw their highest attendance in September. But for most teams, either June, July or August serves as the highest attendance month. Though ticket-based revenue only accounts for about 30 percent of team revenue in this era of big local television deals, the revenue impacts of an April schedule change are likely to defeat any proposal that takes away home games for warm climate teams in those months.

Last year, September was the lowest month across the majors at large, likely due to significant declines for teams well out of the playoff hunt (Baltimore, Chicago (AL), Cincinnati, Colorado, Milwaukee, Philadelphia, San Diego and Seattle all saw sizable decreases in September compared to other months). Several teams in pennant races saw attendance gains in September, but not enough to balance out the non-contenders. The first year of the second wild card, 2012, also featured September attendance below that of April. The data are inconclusive so far, but wouldn’t it be ironic if the second wild card, implemented to make more teams relevant later in the year, was negatively impacting September attendance? In any event, it is doubtful owners would want to make up for April changes in September, given the wide disparity in September attendance that appears to be largely driven by the standings.

This leaves May as the only feasible time to make up for lost/gained April home games. So it would be possible to make changes to April and May, but doing so would only further increase the number of long homestands early in the season. This would ensure just about every team goes on an additional lengthy road trip within the first two months of the season. And again, May is generally more conducive to higher attendance than April.

The other question we have not yet explored is which problem we would be trying to solve—postponements or games that are played in less than ideal weather conditions? Either way, absent a reduction in the length of the season, I’m not convinced either one is worth solving given the cascading effects of shifting two or three series of road games for more than a third of the league.

The impact would probably not be that significant on any of the factors mentioned above, but there are definitely negatives to any proposal, and it’s not entirely clear what the benefit would be outside of increasing the comfort for fans in northern cities like myself. In other words, April scheduling is unlikely to change much in the near future, so might as well get your hat and gloves out and hope it doesn’t snow.