Monday, April 21, 2014

Quantifying the Slower Pace of Baseball


If you're a baseball fan, you've probably noticed that games are longer these days. Certainly the media has figured it out, as any relevant Google search will prove. Just a couple of days ago, new ESPN outlet FiveThirtyEight.com, devoted to the statistical analysis of all things sports, political, and even cultural, produced an in-depth article that showed the Yankees and Red Sox play the longest games.

More notably, the issue popped up when an unnamed MLB executive mooted the idea of shortening the game to 7 innings. That's obviously not going to happen, but the fact that such a silly idea has even made the news is telling. The problem though is not that games are getting longer, it is that they are getting slower.

Pace

If you've read this blog regularly, you'll know I use a statistic called Pitches Per Minute (PPM) to measure the pace of a baseball game. I originally developed this to compare MLB to the Japanese league, notorious for exceedingly slow games. Now, though, I am using it to look at games in MLB exclusively. This post will show that not only are major league games getting longer, they are being played at a slower and less exciting pace.

FanGraphs created a stat, also called Pace, which measures the average time between pitches. I prefer my PPM stat as it is more meaningful in the bigger picture and encompasses the entire game, including pitching changes, inning breaks, replay challenges, injuries, etc; all time we have to sit there while nothing happens. PPM can be used to represent the pace of a single game or an entire season.

Data Inconsistency

Before I get into the details, I have to note that the data I used is from three sources. First, I scraped  all box scores from ESPN.com for every season from 2009 to 2014. I quickly found that ESPN needs a data quality engineer because there is the occasional inaccuracy (check out Jim Johnson's pitch count in this box score, or the fact that every pitch was a ball in this boxscore for a couple of examples). I used Baseball Reference to correct these errors. So I have all game data for five full seasons and so far this year.

For seasons between 1998 and 2008, the game-by-game data was just not accurate enough, so I used season totals from both Baseball Reference and MLB.com. Again, there are some minor discrepancies, in other words, the total number of pitches for certain teams differs slightly between the two sites, but these are insignificant when dealing with such a large sample size. There is no reliable pitch count data before 1998.

Longer Games

Lets start with the basics - the time a game takes to be played. The chart below shows the average game time from 1998 through to 2014 (games until April 20th). This is using actual data from Baseball Reference and the numbers are different from what I have seen quoted elsewhere.


As you can see, there was a concerted effort to shorten games after the average time exceeded 3 hours in 2000, and this cut ten minutes off the average game by 2003. Since then, the pace has slowly crept up until it again reached the 3-hour mark in 2012. Now with seemingly every game enduring a replay challenge, that is up to 3:08 so far this season, 19 minutes longer than in 2005.

So yes, games are getting longer. But longer games are not necessarily a problem, if more is happening between the lines. For example, it seems like there are more extra inning games, and it is generally considered a fact that players are seeing more pitches and working longer counts. If these are true, we should see changes in the number of pitches per game.


Between 2001 and 2009, there was an upward trend (with the exception of an outlier in 2005), but since then the number has bounced around a bit. It's still too early to say this year's number (296) is meaningful as there are usually more pitches in the first month with pitchers still getting settled and colder weather. Last year, there were 294 PPG at the same point in the season but this dropped to 292 over the summer. The question then is whether these extra pitches are enough to account for the longer games. Let's take the total pitches per game and divide by the average time to get the PPM for each season from 1998.


So there you have it, a straight drop from 2009 and its 1.69 PPM to 1.585 last year, a 6.3% reduction in pace over those five seasons. So far this season, the number is even slower at 1.57 as replay challenges (and non-challenges where the manger comes out while his bench coach watches the replay) have entered the game. I should note that there is no appropriate value of PPM; I use 1.67 as a benchmark which represents 300 pitches over a three hour game. Faster than that and you shouldn't get bored, but when it drops below 1.6, the yawns start to set in.

For those interested, the quickest game so far was the opener in the day-night DH between Toronto and Minnesota on April 17th. There were only 4 pitchers, including R.A. Dickey who works quite fast, and they combined to throw 295 pitches in just 2:38, for a PPM of 1.87. The slowest was the 12-inning game on April 20th between the Yankees and Tampa Bay that saw 349 pitches over 4:23 for a PPM of 1.33. I'll update these when new records are set.

Postseason Results

Of course, as we all know, playoff baseball is when the games get even slower. Just looking at the 38 games from the 2013 postseason, the average time was 3:23 and the PPM just under 1.4, with the worst offender the deciding game of the ALCS, where a game with just 275 pitches took nearly 4 hours (3:52) to complete for a ridiculously slow PPM of 1.18. I realize it was a decisive game, but the Game 7 of the 1960 World Series (when the Pirates shocked the Yankees on Bill Mazeroski's 9th-inning homer) took only 2:36 for 253 pitches for a PPM of 1.62. There's just no excuse for such sloth.

Games Under 3 Hours

Another way to look at the increasing length of games is the percentage of games finishing in three hours or less. With my data, I can only do this going back to the 2009 season, although you can take it back farther by scraping game time data from ESPN.com or Baseball Reference. I'm also including games that finished in less than 2.5 hours for comparison.

Year    < 3H   < 2.5H  
2009    63.5%   15.2%
2010    64.0%   15.8%
2011    61.9%   14.0%
2012    55.6%    9.3%
2013    50.3%    7.8%
2014    45.8%    6.1%

In just six short seasons, the number of games under 3 hours has dropped nearly 18% and the number of games taking longer than 3 hours is over half for the first time this season. Only about 1 out of every 16 games finishes in less than 150 minutes. No wonder people are leaving early.

During the 2013 playoffs, only 8 of 30 games were under three hours, with the shortest coming in at 2:36. Not coincidentally that game, a 2-1 Cardinal victory over the Pirates in Game 4 of their NLDS, had the fastest PPM at 1.65.

So what have I shown here? Nothing new, I am sure. Baseball games are long, slow, and they get longer and slower in the playoffs. But hey, attendance is steady, so maybe games are more exciting.

Runs Per Hour

There are many ways of measuring excitement in baseball, including runs, homers, hits, and even strikeouts. Of all these, runs are the most important, so I decided to look at runs per hour, the average number of runs scored per hour of baseball. Personally, I love a 2-1 game but most fans seem to prefer an 8-7 slugfest, so I'll defer to them with this statistic. I used Baseball Reference data to get the average game time per team, total games played, and total runs for each season from 1955.

First, let's look at runs per game for reference purposes.


Most fans will recognize the pitching dominated 1960s, following by the lowering of the mound in 1969, after which runs mostly fluctuated between 8 and 9 except for the juiced ball year of 1987. Steroids appeared in the mid 1990s and runs per game were above 10 for a few seasons before measures were taken to eliminate the cheaters, and runs per game again dropped to between 8 and 9. In other words, pretty much the same as they were between 1975 and 1995.

Now when we factor in the increasing length of games, we really see an effect.



This year, runs per hour are coming in at 2.671, essentially the same as during the famed 1968 season, when they were 2.672. It is still early, with just over 10% of the season completed, but last year this number dropped as the summer progressed. In other words, the average fan is going to see less runs for every hour at the ballpark than any time in history. The media will never tell you this though, as their job is to get you to watch the games. But if value your time and enjoy an exciting, fast-paced ballgame with more than a few runs, you'll probably be better off spending your evenings at the minor league parks across the nation.

Conclusion

Does any of this even matter? Based on my observations, not really. Baseball attendance has been steady over the past few seasons, although it is still off the record number of 2007. I think attendance numbers are inflated and easy to manipulate, but it would be hard to claim that fans are staying away from the game. Case in point: me; I've already seen seven MLB games this season despite all my whining about the pace.

Where I differ is that I pay attention to the action on the field, while most fans don't really care much about the game that they paid to see. I attended the 14-inning game between the Braves and Mets at Citi Field on Sunday (good game but a PPM of 1.498 had me snoozing at times) and few of those around me paid more than scant attention to the goings on, other than to boo Curtis Granderson after every out. Many left before the 9th inning with the game tied at 3. Simply put, most sports fans are casual followers who see the game as nothing more than a distraction, not something to be taken seriously. When three hours are up, off they go to their next destination; sports doesn't dominate their lives. Only if the casual fan stops going because the games take too long and too few runs are scored will the powers that be institute changes. Until then, look for these trends to continue.

Best,

Sean

1 comment:

  1. Brilliant stuff as usual. MLB take note. I was getting the feeling that "baseball" and "short games" were becoming contradictions at 7:00 p.m. games drift to post 10:00 p.m. endings. As much as I love baseball, much past the two and half hour mark, I just don't want to be there anymore live or on t.v.. The dithering and putzing about by both pitchers and hitters is getting truly annoying. I wish they could clone Mark Buerlhe - take ball, take sign, throw....REPEAT. Yankees and Red Sox should be banned form playing one another and save on four and half hour foul ball fests and taking thousands of pitches a game to "work the count". More like working my patience to watch. Sharpy

    ReplyDelete