Time of Game and Instant-Replay Review by Jeff Sullivan April 16, 2014 There was a variety of reasons for why certain people were opposed to instant-replay expansion. It was certainly an affront to baseball purists, who’d already had to deal with replay on boundary calls. Replay reviews would serve to disrupt the flow of the game, irritating observers and players alike. But maybe most importantly, replay threatened to slow down a slow game. Baseball doesn’t exactly fly by at a dizzying pace at the best of times, and the game hasn’t been in need of additional minutes of nothing. Baseball was thought by some a boring sport before agreeing to sometimes spend several minutes stopping the game to look at the same play over and over. It’s the middle of April and we have some early results. There have been nearly 70 challenges, and those games have lasted an average of 197 minutes. If you’re not super good at mental math, that’s three hours and more than a quarter of another hour. That’s too long, considering baseball is supposed to be maybe three hours of programming. The easy assumption, then, is that replay is to blame. But while the current replay system could use a little bit of polish, there’s also a lot more that needs to be said. For one thing, it seems like replay is going to be selective, in part, for longer games, since longer games include more events, and more events means a greater overall likelihood of there being a challenge or two. Maybe what we should care about more is the overall league average. This year, games without a challenge have lasted an average of about 184 minutes. Games overall have lasted an average of just over 188 minutes. A season ago, games overall lasted an average of just over 184 minutes. Right there, you see an increase of four minutes, and the temptation is to blame replay. Without question, replay has been a part. But then, how do you explain what came before? Between 2012 and 2013, average game duration increased by four minutes. Between 2011 and 2012, average game duration increased by nearly four minutes. Between 2010 and 2011, average game duration increased by nearly three minutes. A quick aside: a week ago, I was watching a game between the Dodgers and the Tigers. The Dodgers were starting Josh Beckett, and at one point early on he locked horns against Victor Martinez. Martinez and Beckett squared off for a nine-pitch plate appearance, and in total the plate appearance lasted something like eight minutes. So it was nearly a pitch per minute, and though there was a runner on base, that’s not a full excuse. It was the first inning, and it’s April, and both Beckett and Martinez were wasting an awful lot of time. Below, a chart, covering the last 20 years of major-league baseball. You’ll see the average game duration, in minutes, as well as the average number of batters per game. One of those has barely changed, and one of those has fluctuated. In 2000, the average game lasted 181 minutes. From there it went down, reaching a minimum of 169 minutes in 2005. Since then, the average game duration has increased by 11%, while pitches per game has increased by 3% and batters per game hasn’t changed. If you ignore this season, then between 2005 and 2013, duration increased by 9% while pitches per game increased by 2%. The message is simple: replay has contributed to baseball slowing down, but baseball was already slowing down before expanded instant replay. By coincidence, Ben Lindbergh wrote today about a similar thing. He mostly focused on Pace, which we have available here on the FanGraphs leaderboards. We only have reliable Pace data going back to 2008, and since then, starters have slowed down by a second between pitches, while relievers have slowed down by about a second and a half. I don’t know the explanation for this, but there are a few added minutes right there, considering the number of pitches thrown in your average ballgame. It’s cute and colorful that some batters are superstitious, and it makes sense that some pitchers might work more deliberately than others, but when you put it all together the players themselves aren’t helping the game speed up. The players make it slower. The replay makes it slower. Commercials and assorted other breaks make it slower. How do you speed the game back up a little bit? Figure there’s nothing much you can do about ads. Replay isn’t going away, and it can get only so much more efficient. Remember, the goal is just to get the calls right, and correcting a wrong call is worth a few minutes’ delay. You essentially have to focus on the players, perhaps by enforcing current pace rules, and perhaps by introducing others. Baseball doesn’t have to do anything dramatic. It’s perfectly fine as a three-hour ballgame, but it would be a good idea to start enforcing some things to try to offset the replay delays. It would be easy enough to trim 5-10 minutes. It would be harder to trim another 5-10 minutes, but those minutes don’t necessarily need to be trimmed. And, of course, some people don’t think pace of game is a problem in the first place. Those people are the easiest to please. It was inevitable that replay was going to slow things down, and that’s what we’ve observed in the first few weeks. The system could be made to run smoother, and there are improvements I expect to be made in time. But the game hasn’t slowed dramatically because of replay, and the slowing down we’ve seen only fits a pattern that, before, had nothing to do with replay at all. At issue is a question of efficiency. Instant-replay review is new, and it could be more efficient. The same could be said of the gameplay, which goes on a lot longer than a challenge review does.