Nationals Baseball: Rumination

Monday, February 13, 2017


Baseball thinks it has a time problem. It thinks it is too long and too boring. In response it's coming up with potentially new rules to speed the game along. But what is exactly going on? Is the game getting longer and are fewer things happening? Or are the powers that be overreacting to a popular belief that's taken hold? I'm going to ramble so understand that going in.

The first question, are the games getting longer, has an easy answer. They are. Here you go.  The average time for a nine-inning game is just at 3:00 hours. Of course that's just a data point. What we want to know is the trend and what we see is a general increase in time of game. The most appropriate starting point is probably the TV age (mid 1950s) and games there sat around two and a half hours long. This length was pretty stable until the 80s when the time would rise from 2:31 in 1979 to about 2:45 by 1986 and never go under 2:30 again. It rose up to 2:50+ by the strike year, but would fluctuate from 2:45-2:54 through 2011.  In 2012 it crossed 2:55 which it had only done once before - and hasn't gone back down yet.

OK so it's been going up - however there's two points to be made here.  The first is it's been going up since the 1980s.  If you are looking for a 2:30 hour game to come back you are looking for something that is now over 30 years in the past. Second, it's gone up but is three hours fundamentally different than the low 2:50s of the steroid era?  Is a 6% difference or so that noticeable?

My guess is no, it isn't, at least not by itself.  But when coupled with the second issue - the idea that fewer things happening - the games as a whole feel different.*

Well are fewer things happening? Yes! Now by "fewer things happening" we mean that there are fewer balls in play. Home runs are fun but just involve watching one guy round the bases. Strikeouts can be fun, but again, no action is taking place on the field. Walks aren't fun. It's the increase of the three of these that has caused a drop in balls in play.

Here's a graph for you to see it through 2013. The issue has only grown since then and we're down to just about 2/3rds of all plate appearances ending up with a ball in play.

The thing is, for the most part this isn't driven by walks or home runs. Let's think about this mathematically for a second. PAs have been pretty steady (you can check it out here). It's been 38-39 batters a game since the dead ball era came to an end, with a brief dip here and there to 37-38. HRs are up but only to 1.0-1.2 per nine innings in the steroid era. Over the course of a season that matters, in any given game not so much. In comparison to the 0.7-0.9 range between strikes you are talking an extra HR every 4th game or so.

As for walks, they have actually been down since 2010, sitting in the 2.9-3.1 range in stark contract to the 3.3-3.5 that was in place from 2001-2010, and the brief run of 3.5-3.8 from the 94 strike to 2000.  Around 3.0 compares favorably with historic walk rates.  The six year average for walks per nine is the lowest since the lowering of the mound in 1969.

So it's not walks and it's barely home runs effecting the balls in play. Is it really all just strike outs?  Yep. Last year was the highest ever for K/9 at 8.1. Second highest was 2015, third highest 2014, fourth....well you get the point. Since 2007 each successive year has seen more strikeouts than the last.You can see a steady raise from around 1950 (under 4 K per game) to getting to about 6 a game in the late 60s. The 1969 changes took place and then it dropped back down under 5 until the 80s and it's been rising ever since. But the change recently has been crazy. A decade a ago you saw a one and a half fewer strikeouts a game than last year. It was in about 25 years before that where you saw the same rise.

This is telling. It's not just the rise in K's that's the issue, it's the speed in which that rise has taken place. Now baseball will try to change this by changing the strike zone. It's a brute force method of effecting change. Smaller strike zone leads to more balls thrown near where the batter can hit it well which leads to more swings that make contact which leads to more balls in play. This may or may not play out however.

Baseball has changed the strike zone five times before. In 1950 and 1969 they explicitly shrunk the strike zone. In 1963 and 1996 they explicitly increased it. In 1988 they by definition shrunk it with the hopes that they would effectively increase it**. Let's ignore that last one for these purposes. All we are looking for is simple patterns. Did similar changes create the same effects at the plate? The answer is sort of. Officially increasing the strike zone seemed to have the effects wanted. Strikeouts went up. Walks went down. But decreasing the strike zone had mixed results. If you are looking at decreasing strikeouts, which is the current goal, the 1950 change was not effective, the 1969 change was. But the 1969 change comes with a big caveat. They lowered the mounds as well. (plus there was an expansion which may have mattered but didn't seem to other times) So decreasing strikeouts by changing the zone is not assured. Plus as we noted there has been a huge change in Ks in the past 10 years with no effective change in the strike zone.

What does this really tell us? Well things like K's , BB's and HRs are subject to cultural baseball trends. That's the driver. The strike zone can fight that, but can't beat that. We have seen the general de-emphasization of contact. It is better to swing hard and miss than swing light and make contact.

How can you tell? Take a look at 2-strikes splits. The batting average for these PAs is down to .176. Lowest since we have split records (1988). Much like with K's the lowest numbers are all recent - in this case since 2010 on. (And you do have to remember that offense was much different in 1989 than today so their .182 with a .254 league average is worse than 2009's .186 with a .262 league average.) At the same time though the isoSLG is reasonable, better this year than last, better last than in 2014. Swing hard - get the big hit, don't worry about the K.

Is this for the best? It's hard to say. It's certainly not better in producing good at bats in a vacuum.  2016 was the worst year in hitting with two strikes in comparison to hitting in other situations. But it's not about vacuum hitting, it's about scoring runs. It seems to be an improvement in runs scored over say 2014 but not over 2010-2012. In fact it's much worse at runs per plate appearance than the early 2000s or post-strike 90s. But this isn't all - you have to consider how that approach effects the approach coming into the at bat. How easy is it to be all or nothing until two strikes than change things up? It'll take a lot more digging.

Anyway getting back to my point I guess. Baseball wants to put more balls in play because it feels it has an action problem. The slow growing length of games combined with the rapidly growing number of strikeouts has cemented this idea in people's heads over the past decade. To "fix" it they want to change the strike zone. Historically that approach has found mixed results so I think that it won't have as much impact as they think it will. Instead it has to attack the mindset that saysthat strikeouts are ok. A better way to do that may be focusing analysis on the two-strike outcomes where it isn't clear that an all-or-nothing approach, which works well early in the at bat, is the most effective one.

Years of baseball told us at two strikes we should choke-up and focus on putting the ball in play. A lot of statistical findings have come to find out "Hey, what was proven over years was in fact ground in reality". This might be another one.***

*Why have games been longer since the 80s? Lots of potential reasons. Right now we'll just accept that and move on 

** how does that work? Baseball felt that umpires had settled on a top of the strike zone much lower than the defined line because the defined line (armpit) was felt as being far too high. They didn't have confidence that telling ump to call old line would work so by setting a line a little bit lower than that, and pressuring umps, they hoped to get umps to raise the strike zone to that new line.

*** but it might not! I'm rambling here and this needs a lot more attention than I've given it. But it's worth a deeper look.


JE34 said...

Can we officially blame Tony LaRussa for being the first guy to use a million pitchers per game? If you want to speed up the game, find a way to ban the LOOGY. The proliferation of pitching changes is what kills the pace of the game, and ramps up the challenges to the batter... Hitters in late innings often face multiple fresh arms, which probably contributes to the lower numbers of balls in play.

JE34 said...

Dave Roberts-style pitcher management aside (I admit to yelling at the TV because of it during the playoffs), I would happily watch more baseball, games long or short. I think game length is a casual fan issue... is game length keeping fans away, or diminishing interest in the game? Probably a little, but I can't imagine it's something that has much of an impact on MLB bottom line$. A few minutes won't matter to those who aren't passionate about the game.

Robot said...

Without looking at any statistics or in any way trying to support what I am about to say, I agree with JE34. I think the frequent pitching changes, often a new guy per batter in late innings, has done a lot to impact both the actual time and the perception of the game's length. High K's may be a factor, but they don't disrupt the action the way multiple pitching changes do.

I'd also posit that multiple attempted pick-offs are likely a factor that contributes to the perception of tedium. Anecdotally speaking, Mrs. Robot gets exasperated when the pitcher keeps lobbing the ball over to first base.

JE34 said...

"We have seen the general de-emphasization of contact. It is better to swing hard and miss than swing light and make contact."

Perhaps the pendulum is swinging back to emphasis on contact... Chris Carter was tied with the NL lead for homers last year, right? And he got a weak one year deal. Granted that he can do basically nothing but hit homers (fielding and contact are huge issues for him), but GMs would not pay him for his power because of all those other negatives.

DezoPenguin said...

It's not just the lack of action on the field, either. While the K may not be the most action-filled result, at least it features the mental duel between pitcher, catcher, and hitter, tension as the stakes go up, etc. Maybe more balls in play would make for a more interesting game, or a faster game, but maybe not.

But what's *never* interesting is pitchers and hitters who won't step up to the plate/mound and play the game of baseball. Pitchers who stalk around the mound three times between pitches, rosin bag every three seconds, shuffle around without getting set. Hitters who step out of the box between every swing to spit, scratch, choke up, fiddle with their batting gloves, adjust their cup, step back in, then immediately call time and do it all over again. You watch a Yankees-Red Sox game and it's like every batter and pitcher are a couple of nervous teenagers getting themselves ready for prom night.

mike k said...

I generally agree with the comments here that the frequency of pitching changes contributes to the length of the game. I think that the culture of baseball, compared to the past, is this: strikeouts and weak contact are good for the defending team. Bullpen pitchers aren't the worst anymore. Therefore it's better to have your starter give more per pitch - increasing strikeouts and weak contact - and come out of the game in the sixth and then play mix n' match with the bullpen, than it is to have the starter give less per pitch - increasing balls in play - and have him finish the game. I don't think you're changing that. Harper, you're right that the culture will need to change, but I think (without doing any research on my own) that the data is set on modern day pitcher usage being better than the pitcher usage of the past. No way you're going to convince teams to change that.

Now, as far as hitting style with two strikes - maybe you're right. I'd be surprised if every team was going against the formerly accepted norms of how you hit with two strikes without proper data to back that up. But, if true, yea, you can change that aspect of the culture.

I do think, however, that the greater length of games and the less happening per game issues are more driven by the pitcher side than the hitting side. And that's not changing anytime soon.

How about less commercials? :P

Harper - is the difference in length of games large between the leagues? I'd figure the AL would have much more pitching changes in the middle of innings, which would make a difference.

Ole PBN said...

Looks like we're getting Adam Lind. Good move to spell Zimm and replace C-Rob and some LH pop off the bench. Could get some reps in LF when Werth needs to rest those tired legs, not that he'd be defensive upgrade. I like the deal.

Jay said...

I think a big part of it is catcher conferences on the mound. How many times do things grind to a halt in high leverage situations? I think a simple rule fix would be to limit the number of catcher visits to the mound just like coach and manager visits.

Anonymous said...

Does it really matter if the game is 10 min longer or 5 min shorter when the 'distracted generations' born after 1980 have their face in their smartphone the whole time? You can thank those distracted knuckleheads for the hideous little league chain link netting that ruins the view from 1b to the batters box.

The focus on the game should be to emphasize making the game more 'interesting' and fun instead of length of time, and I think a smaller / higher strike zone will lead to more batted balls in play or walks and more opportunities for something to happen of interest offensively or defensively.

And selfie sticks are gay.

Robot said...

tl;dr - can someone give me the gist of what Anon is saying, but in 140 characters or less? Thx!

Anonymous said...


I believe what he was saying is that he's currently cleaning his shotgun on the porch, listening to his transistor radio, yelling at kids to get off his lawn, and telling the paper boy that his childhood friend Gutenberg is the reason the boy has a job.

Anonymous said...

Have you tried Google translate or Wikipeadia? Or post it to your FB wall for comments.

Harper said...

mike k - no. In general AL is a couple minutes longer, that's all. And in a given year the NL can have longer average games. I am surprised too but I think what's happening in these cases is that the NL simply uses more pitchers per game with PH and warms up between innings are a little more stretched out for new pitchers. (and using a PH takes time as well).

Harper said...

I agree that it's probably pitching changes that are the biggest driver but a little more than one more a game in the past nearly 30 years doesn't alone explain a 20+ minute increase in time

Pitching changes / mound visits
+ More offense
+ Commerical time?
+ Pitches per at bat
+ Some cultural "take your time" pitching and hitting approaches
+ Replay!
+ etc. etc

no one change is going to get those 20 minutes back.