For a better browsing experience, please upgrade your browser.

FiveThirtyEight

Sports

Major League Baseball is slowing down at the same time that its formerly slowest team is speeding up.

The New York Yankees haven’t dominated the majors since 2002, winning just one world championship over that period. But they’ve dominated MLB rankings for length of game. From 2002 to 2012, the Yankees’ average length of game was in the top four of the 30 major league teams each season, including five league-leading performances. Yet lately they’ve gotten faster: The average Yankees game has been shorter each year than the year before since 2009, culminating in last year’s 15th-place showing of three hours and five minutes — just half a minute longer than the league-average figure.

The slowest team in baseball is now — wait for it — the Boston Red Sox. New York’s archrivals have ranked in the top three in average game length each year since 2003. Last year, Boston’s fourth straight season as the slowest team in baseball, its average game took three hours and 15 minutes.

Boston is part of a broader trend that the Yankees are defying: Baseball games are getting longer. Games averaged two hours and 56 minutes in 2011 and three hours flat in 2012. They tacked on four more minutes last year — even as runs per game fell, to 4.17 from 4.32 the year before. And games this year, with the advent of instant replay, have added another three and a half minutes on average, through Sunday.

bialik-time-of-game-1a

When the two slowest teams of the modern era meet, great baseball often results. Slow baseball almost always does. The three Yankees-Red Sox games last weekend each took over three hours — and each was faster than the average Boston-New York meeting since 2002.1 Those games have averaged three hours and 23 minutes — 15 minutes longer than the next slowest matchup with at least 100 games during the period (Yankees-Angels).

bialik-time-of-game-2a

When staging such a great rivalry, former Boston closer Jonathan Papelbon asked in 2010, what’s the hurry? “Have you ever gone to watch a movie and thought, ‘Man, this movie is so good I wish it would have never ended’? That’s like a Red Sox-Yankees game. Why would you want it to end?”

Many factors fundamental to baseball influence game length: Batters’ rate of getting on base safely, their plate discipline, and how long they take to round the bases after a home run. Boston’s David Ortiz regularly tops the leaderboard for longest tater trots, which contributes to the glacial pace of an average Red Sox game.

Another influential factor has to do not with the action, but with the time in between plays. Pitchers stare down base runners or wave off signs from catchers. Batters step out of the box and adjust their helmets and other gear. The time between pitches is getting longer. We know because of a byproduct of the sophisticated PITCHf/x tracking system installed in every major league park that stamps a time on every pitch. FanGraphs turned those timestamps into a stat, called pace, measuring the time between two consecutive pitches in the same plate appearance.2

The pace is slowing. The average break between pitches was 21.6 seconds three years ago, 22.1 seconds two years ago and 22.6 seconds last year, according to FanGraphs. Taking into account the number of pitches per game, the slowing pace could account for five of the eight minutes tacked on to the average game between 2011 and 2013.

Boston slows things down when at bat and on the mound. Last year, Red Sox batters ranked second in time between pitches, with 23.5 seconds. (The Yankees ranked first.) And Boston’s pitchers ranked seventh at 23.4 seconds between pitches. No individual player was particularly responsible for the Boston slowdown; only reliever Junichi Tazawa ranked high on the pitcher or batter leaderboard. It was a team effort.

The reasons for lengthening game times have been a longstanding baseball preoccupation. “No one element accounts for the delay of game. Some developments, such as extended commercial breaks and the tendency to use more pitchers, are more obvious,” the St. Louis Post-Dispatch reported – in 1991.

Just how much does any one team, or any one attribute such as pace, affect game length? I ran a linear regression to check. ESPN Stats & Info provided detailed data on more than 29,000 regular-season games, between 2002 and this past Sunday. I checked how the teams involved contributed to game length, while controlling for year, number of innings, number of relievers used, runs per inning, plate appearances per inning, pitches per plate appearance and closeness of score.3 (I didn’t control directly for other factors, such as the number of pickoff attempts or pinch-runners, or the big increase in defensive shifts in recent years.)

The result: The Yankees added 13.9 minutes to the average game, relative to the fastest-paced team, the Oakland A’s. Runner-up Boston added 10.5 minutes.4

bialik-time-of-game-3a

These, again, are the results after controlling for attributes of the game. But those play a role, too. Each extra inning in a game makes it about 18 minutes longer. Each additional call to the bullpen adds a little over two minutes. Every additional plate appearance per inning tacks on nearly 21 minutes, and an increase of one pitch per plate appearance increases game time by half an hour.

The more time passes, the more time passes: The length of games has grown 0.7 minutes on average each year since 2002, after controlling for other changes in the game. And all else equal, close games take a little longer: A minute more time for every run shaved off the final margin.

All of these factors are highly statistically significant,5 which isn’t hard to accomplish with a data set this large.

The pace stat is available, and reliable, going back only to 2008. So I also took a closer look at the games between 2008 and 2013. The overall trends, after running the same regression, were similar: The Yankees continued to rank No. 1 despite their speed-up toward the end of the period, adding 12.8 minutes per game relative to the fastest team in the period, fellow AL East team Toronto. The Red Sox ranked second, adding 11.8 minutes per game.

These teams’ slow pace of play explained some, but not all, of the time their presence added to games. I added another variable to the regression: an estimate of how much time the two teams in the game were adding to it based on their average pace, at bat and on the mound, during the six-year period.6 Controlling for that, Boston was just three minutes faster than Toronto, which remained the fastest team. The Yankees added about six and a half minutes, which is a lot of dead time, but a lot less than 12.8 minutes.

Major League Baseball has a rule to keep the game humming along: Official Rule 8.04, which gives the pitcher 12 seconds to pitch when no one’s on base. It’s rarely enforced — about 15 to 20 times in 2009.

Without stricter enforcement, baseball might get slower. Any new innovation by one team that proves successful is copied by rivals — hence the growing price tag for on-base percentage. And the Yankees and Red Sox, the two slowest teams since 2002, are also by far the winningest teams in that era. The Dodgers and Angels, next in the rankings, both are in the top eight in wins. Slowing down the game from pitch to pitch, for whatever reason, is correlated with winning games.7 Correlation isn’t causation, and there’s no particular reason slow play should equal good play. But Yankees fans might not want to take any chances: Their team, in its fastest-paced season of the last 13 years last year, had its lowest win total of the period.

Footnotes

  1. The first game of their four-game series last week, on Thursday, took a brisk two hours and 55 minutes. ^
  2. Baseball Prospectus introduced its own version of the stat this week. It differs slightly from FanGraphs’, mainly because it omits pickoff attempts. 

    Papelbon did his part to make Yankees-Red Sox games seem endless: He’s second, behind Rafael Betancourt, on FanGraphs’ pitching leaderboard for slowest pace of play since 2008. ^

  3. Pitch data was available only back to 2002. As I wrote last week, closeness of score doesn’t have a big impact on raw time of game, but I included it to see if it does have an influence after controlling for other factors. ^
  4. These results apply whether the team was at home or away. Preliminary analysis suggested which team was at home made little difference to the results so they were treated the same, which doubled the sample size for each team. ^
  5. p <10^-15. ^
  6. For each game in the data set, I took each team’s batting and pitching pace over the period. I divided each figure by the average pace during the period of 21.8 seconds between pitches. Then I multiplied each team’s number of pitches by its relative pitching pace and by the batting team’s batting pace. That’s roughly how much time elapsed between that team’s pitches during the game.

    To understand why, suppose one team’s pitching staff played 5 percent faster than average during the period, and its opponent’s batters played at an average pace. Then the time taken before pitches by this calculation is (1*0.95)*(number of pitches*average time taken between pitches), or 5 percent less than the average time. Then I subtracted from the result, the number of pitches multiplied by the average pace. The result is an estimate of how much time was added or taken away relative to average by the time before that team’s pitches.

    I did the same calculation for the opponent’s pitches, using its pitching pace and the first team’s batting pace. Then I added the totals. This gives a rough relative measure of how much the pace of the two teams involved in the game affected its length. ^

  7. The correlation coefficient for the 29 teams that have been in the same city since 2002 — excluding the Washington Nationals, né Expos — between wins and our regression output for time added to games is 0.55, a moderately strong positive relationship. ^

Filed under , , , , ,

Comments Add Comment

Never miss the best of FiveThirtyEight.

Subscribe to the FiveThirtyEight Newsletter
×

Sign up for our newsletters to keep up with our favorite articles, charts and regressions. We have two on offer: The first is a curated digest of the best of what’s run on FiveThirtyEight that week. The other is Ctrl + , our weekly look at the best data journalism from around the web. Enter your email below, and we’ll be in touch.




By clicking subscribe, you agree to the FanBridge Privacy Policy

Powered by WordPress.com VIP