Sources of the Home Run Surge
Baseball Dynamics, Climate Change, and Ballparks, oh my.

Justin Verlander, no doubt annoyed that his home run rate has increased along with the rest of baseball, tweeted a fascinating graph. It shows that, given the same exit velocity and launch angle, the probability of hitting a home run increased significantly in 2017 compared to 2014, and it did so for every conceivable vector (except low line drive homers).

The guy reaction is to return the discussion of the homer spike to questions about changes in the ball. And, indeed, there have been changes to the ball. One study by the nerds at 538 found, for example, that the average baseball was half an ounce lighter in 2017 than in 2012.

https://fivethirtyeight.com/features/juiced-baseballs/amp/

It's a good read, so check it out.

So let's consider all of the factors that could theoretically impact the physics of projectile motion for a batted ball. Those factors are:

  1. Baseball material physics (how the ball compresses, recoils, and spins)
  2. Ballpark effects
  3. The atmosphere (air density, primarily)
  4. Trajectory
  5. The quality of pitching (not just as it pertains to launch vector but the batters' ability to get the optimal clean contact to produce a favorable type of spin that would reduce drag)

I won't conduct a science class in projectile force equations because that will include differential equations once we get to air resistance and recoil/deformation...but we can summarily rule out two possible sources of the home run spike. The homers can't be caused by a change to more favorable trajectories (at least not in total) because launch velocity hasn't significantly increased overall and, even if it had, the same velocities are leading to more home runs. And it can't be caused by the ball's mass alone. Projectile motion is not impacted one iota by the weight of the projectile.

The reasons the mass of the ball matter are twofold. First, a lighter ball should have an increased exit velocity on average if it's compression and recoil dynamics are similar to the heavier ball because momentum must be concerned and if you shuffle the batting momentum away from the mass, you get more velocity. Second, a lighter ball is not lighter because it's smaller. It's lighter because its core is different. The precise nature of that difference determines how the ball responds to the impact of the bat.

HOWEVER, it is not as simple as a lighter ball is springier (or the reverse). If a ball is lighter because its core is wound more loosely, that will cause the ball to be 'mushy', meaning it will compress more when struck by a bat than the tighter core. A more compressed ball takes a less spherical shape during and just after launch, which increases the drag of air resistance and causes the ball to lose more of its initial velocity early in flight (and, thus, to fly significantly shorter distances). If the ball is lighter because the material is less dense due to being altered to make the core 'springier', the compression issue neutralizes faster (springy materials spring back to normal shape faster) and the springiness increases the recoil jolt (jolt is the physics term for the rate of change of acceleration), leading to increased launch velocity.

But, we just established that increased velocity isn't causing all of this home run spike. And, when you take apart a modern baseball, the core material isn't significantly different.

I posit, then, that some other factors are at play beyond the weight of the ball.

So let's look at my first component above: baseball dynamics. What are the elements of a ball that alter its flight? We ruled out springiness and mass, generally. That leaves the skin of the ball and the seams. The smoothness of a surface alters the turbulence created by the air rushing past the irregularities and creating little eddies. More turbulence = more drag. A shiny ball (one with more of a glossy surface) will fly more easily than a scuffed up ball, or even a ball with a suppler skin. Lower seams will create less drag than higher seams.

So what are pitchers saying about the skin of the ball?

https://www.usatoday.com/story/sports/mlb/mlb/2017/10/29/world-series-sl...

This is just a comment on the World Series - the pitchers thought the world series balls were slicker than normal, but subsequent analyses suggested that breaking pitches broke less than normal (slightly) even as the league used fewer fastballs than ever in the history of the game. Ergo, one possible factor could be that regular season baseballs are subtler in their slickness, but still slicker than balls from 2014.

How about the ballparks? Have any of the parks change dimensions between 2014 and 2017? Yes, in fact. There have been several changes. Minute Maid Park (Astros) deleted Tal's Hill prior to the 2017 season, shortening center field by 35 feet and the power allies by five feet on either side. Safeco Field pulled the left field wall in by 10 feet in the power alley and five feet down the line. Oriole Park increased the height of the left field wall by several feet. Pittsburgh shortened LCF slightly. Atlanta got a brand-spankin' new ballpark in 2017...one that played more toward hitters than the last park did. On net seven ballpark dimension shifts of note and six of them favored hitters.

There's no evidence that the quality of pitching has gone down aside from the homer spike, but that would be circular reasoning. K rates are up, BB rates are down, BABIP is down and guys like Lynn and Cobb can't find a job because the young talent in the league is so deep that teams feel they can avoid signing expensive veteran stopgaps.

That leaves the weather. And here, my friends, is the great, overlooked factor in this debate.

Air density (the primary driver in air resistance changes) depends on altitude (hello, Coors/Chase), sea-level pressure (highs and lows that create our sensible weather changes), temperature and dewpoint. Being a meteorologist particularly skilled in medium and long-range forecasting and seasonal forecasting, I am always aware of general trends over the US in temperature and precipitation. So I knew that 2016 and 2017 were much warmer than 2014 before even looking at the data (and wetter, BTW). But here is that data:

CONUS average temperature by year:

2014: 52.8 F (+0.8 over 20th century averages)
2015: 53.9 F (+1.9)
2016: 54.9 F (+2.9)
2017: 54.6 F (+2.6)

Urban areas have warmed more than rural areas (up 2.4 degrees vs. up 1.5 degrees) in the last four years, continuing a trend of urban heat island expansion and magnification. Nighttime low temperatures have increased drastically more than daytime highs (for example, in 2017, CONUS average high temperatures were 0.5 F warmer than normal and low temperatures were 4.7 F warmer than normal). And high latitudes have warmed considerably more than subtropical areas, meaning stadiums in the north that are exposed to open air will feel the effects more than their southern neighbors (relatively speaking).

Now, some fun facts about MLB:

There are nine stadiums in the southern U.S., 9 in the middle latitudes, and 12 at high latitude.

Of the southern ballparks, Chase Field, Minute Maid Park, Tropicana Field, and Marlins Park have some degree of weather modification for temperature and humidity. Of all other stadiums, only Rogers Centre can alter temperature and dewpoint (by closing the roof).

Baseball games are played 72% at night.

Only five ballparks that aren't weather modified are placed outside areas that would be called urban. Kauffman Stadium, Target Field, the Oakland Collesium, The Ballpark in Arlington and Angels Stadium.

As for humidity, rainfall has increased over CONUS every year (year over year) from 2014 through 2016 and then declined very slightly in 2017 despite all the hurricanes. Rainfall and humidity aren't perfect proxies for each other, but it hints that vapor pressure was higher in 2017 than 2014.

So, if the average major league game was played in weather between 2.5 and 4.5 degrees warmer and slightly more humid in 2017 compared to 2014, that's a *significant* contributor.

I believe the home run spike is a confluence of several factors. The average exit velocity may be slightly higher, the ball may be slightly smoother, the weather is definitely significantly warmer and slightly more humid, and the parks tilt more toward hitters than they used to.

Complaining about the ball is, I believe, fruitless because the ball might well be the smallest factor.

Comments

1

A great read.  

Interesting stuff about nighttime temperatures.  I would jump to the conclusion that increased cloud cover would likely be the reason, but if rainfall was down, that might be a big reach.   Your evaluation here?

2
tjm's picture

This is absolutely fascinating stuff coming completely out of the blue (get it?) to me.

Thanks Matt. I think you and Doc could do a roadsow. Doc does his bouncing off walls schtick then here comes Science Man brings the facts to excite the crowd. It'd be like Bobby Bland and BB King from back in the day. Way back in the day. Bobby oozing his smooth, beautiful intoxicating love vocals then BB's guitar hits you with the hard truth.

3

...at least over the last several years, although there is little evidence that the longer-term climate shift has caused an overall increase in rainfall.

However, earth experienced a major satellite dimming during the eighties and nineties (fewer clouds made the planet darker, as in...lower albedo) and warmer. And that cloud trend has started to reverse of late.

5

Well done.

I'd be interested in a data set that included only Apr-Oct, since a warmer winter (or colder for that matter) doesn't matter at all and could potentially skew the relevant averages.

6

And it does count against my argument above, as well. Most of the warming we've experienced in the last few decades has been in the winter months.

However, here are the summer (June-August) averages for those seasons:

2014: 70.7 F (+0.3)
2015: 72.8 F (+1.4)
2016: 73.5 F (+2.1)
2017: 72.7 F (+1.3)

Spatial distribution matters as well. In 2014, it was fairly cool over much of the center of the country (not extremely so) and hot in the SW US with normal conditions elsewhere. In 2015, after a cold winter across most of the eastern 2/3 of the country, summer was fairly hot everywhere except the western Great Lakes and the northern Plains States. 2016 was very hot everywhere...it's still the hottest year in CONUS recorded history and the third hottest summer on record. And in 2017 it was nearly as hot everywhere except the southern plains and Gulf Coast states due to tropical cyclone activity and excessive rainfall (which made for a below average summer).

As for Precip, the summer was normal to slightly above normal for rainfall everywhere except Texas and Oklahoma in 2014. In 2015, the country ran wetter than normal except in New England/New York and the cotton belt, which ran a bit dry. In 2016, it was very dry throughout the western third of the country and in Florida and Georgia as well as southern New England. but very wet in the central Great Plains, the Midwest and Tennessee/Ohio River valleys. In 2017, it was a classic El Nino summer - wet across the southern states, even including California, and up into the eastern states, and dry in the Northwest and northern Plains.

On net, 2014 would bias toward more humidity where we have baseball teams in 2014 and 2015, normal humidity on average in both 2016 and 2017.

7

This is some superior shtick Matt.  I Bet you Rob Manfred added pop to the core of the ball to make the game more exciting.  He  has the means, motive and opportunity to pull such a caper.  His name is on the ball and he talks of nothing else.

Also, I suspect that steroids aren’t caught any more.

I always thought humidity was a great enemy of  ballistics.  What gives?  Good Friday Frappachino.

8

Humid air is less dense than dry air because water vapor molecules (H2O) have an atomic weight of 18, whereas N2 molecules have an atomic weight of 28 and O2 is 32. In short, water vapor is lighter than air. As the air takes up more water vapor, it displaces a bit of the air if the atmospheric pressure does not increase. In other words, if you take a box of air and pump in water vapor, the air inside will become denser (because you're cramming more stuff into the same volume), but if you do that in an open atmosphere, the result is frequently a movement of air to equilibrate the pressure and a less dense air mass where it is humid.

Lower density = better ball flight. The moisture, if it accretes onto the projectile, can act to increase drag, but normal amounts of humidity in the air don't deposit enough water on the ball to significantly impact drag in the five seconds it's flying (give or take 3). Now...when firing an ICBM that flies for 30 minutes before reaching the target, humidity (especially super-saturated cloud layers) might be a problem.

10

Relentlessly logical, data-driven, and a pleasure to read.  Thanks for the analysis Matt!

11

Vey good explanations of complicated interactions.

One thing not fleshed out is the increase in breaking and offspeed pitches.  I can't find any info on League HR/FB by pitch type per year.  Just that and pitch type per year might explain a decent chunk as well or might only show symptoms instead of the cause. 

13

I'd like to see whether home run rate by visiting clubs has changed in climate controlled stadiums by the same rate that all other stadiums have seen. There are four or five stadiums that can either fully control the climate (Rogers Center, Tropicana) or cool the air despite some open access (Marlins Park, Chase Field, Minute Maid). That makes for an interesting control group to attempt to isolate the weather impacts.

Add comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd><p><br>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

shout_filter

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.
  • Web page addresses and e-mail addresses turn into links automatically.