Home Runs and the Middle Class
Yesterday, with the help of a much smarter friend, I dug a little bit into the home-run surge that we’ve all observed. There was room for some more examination, though, so this is to be considered a brief follow-up. Toward the end of that post, I engaged in some speculation. Based on evidence below, I believe I was wrong!
To re-cover some territory real quick, we do have this fact, established in the original post: Home runs are being bunched unusually tightly. The distribution of home-run rates around the game is closer to being even than ever before in at least the game’s modern version. That’s interesting! More and more hitters are getting into the dinger spirit, such that the landscape isn’t so dominated by a handful of elites. That now being given, let’s move on.
I speculated in the post that the relative rise of the middle class could be an argument against the different-baseball theory. There’s a very compelling and convincing line of thought that the baseball has changed, leading to more homers, but I guessed that any such change would affect all hitters equally. That’s not what the data seemed to show. Now I want to argue against myself.
For this research, I made use of Baseball Savant, which provides all that beautiful Statcast data. I wanted to compare 2016 to 2015, so I grabbed all the player-seasons with at least 200 plate appearances, and this time for each of them I calculated home-run rate per fly ball or line drive. In the post yesterday, I calculated home-run rate per batted ball, which is a little different. Once I had all the rates, I divided them into quartiles again, in order of power output. Then, from Baseball Savant, I pulled a bunch of average fly ball/line drive distances. I understand that using the average isn’t the absolute best way to go about this, but it should do most of the job.
Here is a plot of average distance by quartile, for each of the last two years:
This is why I think I was wrong. That top quartile is showing a boost, here. Every quartile shows some kind of increased distance. Now, I get that Stacast hasn’t captured every ball in play. That’s a potential source of error. But if we take this for what it is, there’s an increase across the board, that increase most favoring hitters who should want balls in the air. This is what we would expect if the ball were changed. Everyone gets to hit the ball a little farther. I think that pulls the rug out from under what I speculated.
The most powerful quartile shows the biggest increase in distance. To sort of switch gears, here are the home-run rates themselves:
You see an increase in the most powerful quartile, but you see an identical increase in Quartile 3. Further, you see a similar increase in Quartile 2. Quartile 1 hasn’t budged much. The batters in Quartile 4 have increased their average distance by almost five feet. The batters in Quartile 3 have increased their average distance by three feet. The batters in Quartile 2 have increased their average distance by less than two feet.
A few of you got this right away, in the comments and on Twitter. And it seems to be true: The strongest hitters just don’t stand to benefit so much from a distance increase, because they would hit relatively few “almost-homers.” The weakest hitters also wouldn’t benefit too much, because they would seldom challenge the warning track. But those guys with warning-track power would benefit the most, because then it’s just a matter of a few feet turning an out into a dinger.
The marginal value of an extra few feet for a player with average power surpasses the marginal value of an extra few feet for a player with tremendous power, or for a player with zero power. Think of it as operating a lot like the win curve we’ve talked about a million times. An 89th or 90th win is far more valuable than a 108th win or a 56th win. Related to this, here’s a raw curve, showing home-run probability by fly ball/line drive distance, calculated in increments of five feet:
It’s not a straight line, and it steepens around 380 feet, before flattening out again around 420 feet. Any movement for a player around the middle could be hugely important. Around either end, it wouldn’t matter so much.
This all feels like something very obvious, but it’s good to have evidence, even if it could stand to be more robust. I’ll leave further research to other people who are better at this than I am. They number in the, I don’t know, thousands?, but even with where we are today, it’s of interest to go over a couple implications.
For one thing, we still can’t disprove the idea that the baseball has physically changed. They keep saying that it hasn’t, and I even want to believe them, but there’s just no other reasonable explanation I can think of for what’s been happening league-wide. I thought I might’ve been able to make one point, but then I went and argued against myself, and I’m mentally back where I started. The baseball is flying, for all hitters. The weaker ones and the stronger ones. I would really, really love to know why.
And for another thing, because the baseball is flying, that seems to have most strongly benefited the middle class of would-be power hitters. Those are the hitters for whom a few extra feet of distance would have the greatest marginal value. More distance helps every hitter, but it doesn’t help them all the same. So teams might be unsure how to proceed. Should they invest in average power, or not? Will this end up a mirage, or should roster-building start to favor the warning-track types, since they’re suddenly becoming second-row types? It’s almost impossible to know what to do without knowing the reasons behind the homer surge in the first place. That’s one of the reasons this is such a big deal for everyone.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
Thanks for this article. Mentioning a “juiced ball” regularly gets you downvoted to all hell on here but to me it has seemed like the most logical explanation for awhile now.
Agreed. I think the game has been changing for the last decade or so, but for it to just “click” for such a large group of guys in the same year seems extremely unlikely. Something else has to be going on.
Couldn’t it be that the physics of hitting a ball are more accessible than ever before and hitters are adjusting their swings to maximize their exit velocities and launch angles?
So over the span of one year, players have universally used this data and successfully changed their swings?
That graph of distance vs HR probability and the steep slope from 380-410ft makes the argument clear as day. Unfortunately it still won’t help with the “It’s PED’s!!” crowd since arguments like marginal utility and steepness of slopes don’t tend to hold much sway with the less math inclined 🙁
They’re two separate arguments. Jeff is pointing out the data suggest that hitters with less power benefit more from about an equal increase in FB distance. He’s not addressing the question of why the FB distance is greater.
Valid point. I find the suddenness of the observed league-wide EV jump around mid season 2015 and this flattening of league wide HR totals rather than records once again being threatened both fly in the face of juiced players.
Another commenter put it well in yesterday’s article that it would be weird if a new undetectable PED became widespread overnight midseason 2015. In addition I believe Rob Arthur recently found that the same HR surge is not happening in the minors where presumably the same PED would be in use.
Between those solid arguments and my entirely unscientific observation that we just aren’t seeing the sorts of physiques that predominated “the steroid era” the juiced ball seems much more plausible.
The marginal utility of 5-10 extra feet being much greater for all the guys with “warning track power” just doesn’t lend itself to succinctness in a theoretical bar debate.
What would even be some other arguments? They all sound crazier than the juiced ball theory.
1) New cutting edge steroid that can’t be detected and went from nobody taking it to almost the whole league on it within a year
2) Weather related. I’d think we’d be hearing about the most unusual wind patterns or air pressure on record in the US if it was this.
3) Swing for the fence mentality taking off. This could explain the change in middle class. Anecdotally, we hear of guys like Daniel Murphy changing their swing for more power. This would be a stretch explanation for the middle class as a whole and wouldn’t explain elite power guys and slappy hitters distance increases.
IN terms of #2. Is it possible that global warming could have any impact on this? I know the average temp of the planet continues to rise every year. Is it possible that some sort of thresh hold has been surpassed that is now affecting the ball? It seem s ridiculous but it is something I’ve wondered for a while.
Regarding point 2 (weather), we’ve been consistently seeing reports like “hottest July on record” this summer. These warmer average temperatures are largely reflected in heat that lingers on into the evenings rather than hotter maximum temperatures. We know baseballs fly farther in warmer temperatures. I don’t think weather patterns can explain anywhere near the flight differences noted in this article but something like 20% weather related seems plausible.
I’m no genius or anything, but can an outside party get access to a couple dozen game balls (including some that were HRs), and then cut them open and examine them against previous years balls?
Seems pretty convincing that the ball is juicy. I, for one, like it. Kind of offsets all the strikeouts and makes the game more fun. It’s not like it’s somehow “unfair.” Anyone else?
For the record – HR/G are up to 1999/2000 levels. Runs/G are around 2009-ish levels. R^2 relationship between HR/G and Runs/G is pretty strong (0.66), but this year’s datapoint actually sticks out like a sore thumb on the graph.
http://i.imgur.com/5INbmDC.png
Without this year included, R^2 for HR/G and Runs/G is 0.81. Potentially suggests that something was done that didn’t change approach or strike out rates, but changed the rate at which home runs happen.
Putting the graphs back in fangraphs: Yes! Awesome addendum comment to an excellent article.
So strikeouts are still up, now home runs are up… More three true outcomes! I wonder if ERA is correlating with FIP better since (I assume) the TTO are becoming a larger % of events?
HR’s per game are at their 2nd highest rate in Major League history. A couple weeks ago the average had surpassed 2000’s record. http://www.baseball-reference.com/leagues/MLB/bat.shtml#teams_standard_batting::12
“You see an increase in the most powerful quartile, but you see an identical increase in Quartile 3. Further, you see a similar increase in Quartile 2. Quartile 1 hasn’t budged much. The batters in Quartile 4 have increased their average distance by almost five feet. The batters in Quartile 3 have increased their average distance by three feet. The batters in Quartile 2 have increased their average distance by less than two feet.”
Wrt HR/FB rates, Q1 has actually increased more than Q4 as a % of the former rate:
Q1 15.3%
Q2 21.8%
Q3 18.75%
Q4 12.1%
Yesterday’s article looked at the % increase in HR over the previous rate, so this should be compared to the % increase in FB distance over the previous distance. The % increase is roughly the same for all four cohorts.
Also, it’s a little confusing to reverse the quartiles. Yesterday, Q1 was the cohort with the highest HR rate, whereas in this article, it represents the weakest cohort.
Yeah, that quartile thing was just sloppy. Don’t really know how or why that happened! My bad.
Do we have launch angle data from last year or is this the first year we’ve had it?
Exit velocity and launch angle data are available from last year, at least to some people. There is an article in HB Times posted here right now that discusses it.
The timeline, as far as I understand it, is that mostly-reliable Statcast data started become available at the start of 2015. Then half-way through 2015 and continuing to the present, the home run rate spiked.
Is it possible that Statcast itself is responsible? In the same way I’ve seen in theorized that better data and pitch f/x helped pitchers dominate batters, what if Statcast has allowed hitters to improve their approach? I’d mostly expect to see this in terms of optimizing launch angle rather than exit velocity, but the two could be related.
Not saying the ball’s not juiced, but it does sound like there’s an emphasis on optimizing launch angles and hitters changing their approaches to generate more high-quality fly balls rather than groundballs. That would lead to fewer hits and more home runs.
That’s an interesting thought, and I’m sure there’s data somewhere that could tell us one way or the other. My thought is that it’s VERY hard to change approach/trajectory of the ball at will, otherwise the shift wouldn’t be as effective as it is. But I’d be interested in following your line of thought and seeing what we could find.
Interesting read over at HBT.( http://www.hardballtimes.com/barrels-normative-analysis-and-the-beauties-of-statcast/ ) Suggesting that the stat “barrels” correlates strongly with ISO; HR/G and; SLG. Maybe Statcast is influencing this trend? At least some what.
Another thing that could be interesting to look at is where those home runs are happening. From the graph of flyball distance vs. likelihood of home run, if the biggest impact is in the 380-420 range, wouldn’t that mostly help in centerfield as opposed to the corners? Are doubles down at all this season?
Going back to 1999, doubles/game peaked in 2007, but has been holding steady since 2012. Makes sense. If we are to assume the “juiced ball” theory affects all batted ball events, then we’ll lose some doubles to home runs, but something that would have been a long single is now a double.
Slow to age speed players such as Gyorko, Horner, and Bruce have appeared to have converted long doubles to homers- absolutely.
I feel like this is about to be incoherent and/or poorly worded but I wonder how much of a factor (if any) the influx of young hitting talent the last could years could impact this, especially at premium defensive positions. I think of rookies like Seager, Story, Sanchez, Turner, Contreras, Diaz as primarily up the middle guys not to mention Schimpf, Mazara, Kepler etc. Plus there are younger guys who are going through their 2nd and 3rd seasons and likely making adjustments and learning such as Addison Russell, Semien, Bryant, Gregorious, Correa, and Lindor among others who may be displacing aging players hitting for less power. I would be interested if the data from the other home run post and this post could split by position or tenure to see how premium defensive positions are performing home run wise as well as 1st, 2nd, and 3rd year players who have been making strides in the power department as they become more experienced. I am just thinking anecdotally it seems like we would see an increased home run floor among lesser tenured and up the middle players compared to previous years. Maybe we are seeing less inequality by position as well as the quartiles you mentioned previously.
Has there been a change to the bats being used. Perhaps a shift to more maple bats, axe handles, etc ?
Curious thought. Perhaps it’s as simple as supply and demand. Ash species of trees have been decimated by a borer insect the past few years over it’s range, thereby decreasing supply dramatically. Maple use is going up in part to this epidemic. In fact, it will be very difficult to purchase an ash bat once the borer has completed it’s march through the range of ash. The exotically introduced pest is predicted to virtually extinguish all mature ash trees to extinction, similar to the fate of the American chestnut.
The switch to maple may be that we just passed an economic threshold, and inadvertently increased production in homeruns league-wide.
I’d venture that an increase in broken bats would go up at a similar rate to homeruns, since maple is more susceptible to splintering and shattering.
A lot of people on here have speculated that homerun surge has come from guys changing their approach, but I wonder if the inverse is true (i.e. players are changing their approach because they’ve noticed it’s easier to hit homeruns).
Jeff, why are you comparing 2015 to 2016 when presumably “the change” started in the second half of 2015? That is only going to dilute any differences (obviously).
For the record, you would need an average temperature change of maybe 30 degrees to account for the offensive surge. IOW, weather is NOT a factor. For the record, the average game time temp in 2016, according to retrosheet data, was 74.0 degrees. In 2015 it was 73.6. So let’s stop with the weather stuff.
If players can be grouped into upper, middle and lower classes based on their ability to hit HR, I guess those of us reading and responding to such articles are sleeping on the street in third world countries.
I’d be at the bottom of the “untouchable” caste.
This system is rigged for the Mark Trumbo’s and Edwin Encarnacion’s of this country. If you need proof, just look at how Daniel Norris lives–in a van! We, the Anthony Rendon’s and Nick Markakis’ and Munenori Kawasaki’s of this country, need to stand up to the 1% and tell them that all players deserve a living wage–$1.5 million per year with a partial no-trade clause
Jeff, I posted the following on a message board on June 3rd regarding speculation on this trend, based on anecdotal examples at the time (Betts, Altuve), increased exit velocity across the board, and eyeballing the full homerun rankings list.
“Roll back to the juiced ball season of 1987- the leaders weren’t too too high, but the average HR total per player skyrocketed- long fly-outs became homers- Wade Boggs had 24, Dale Sveum had 25, Juan Samuel had 28, Mattingly 30, Wally Joyner 34, Mike Pagliarulo 32, Keith Moreland 27, Alan Trammell 28, Larry Sheets 27, Dwight Evans 34, Brian Downing 29, Brook Jacoby 32, Brett Butler 9, George Bell 47, Ozzie Virgil 27, Will Clark 35- 500 more than were hit in 1986- perhaps we are seeing 1987 redux this year with these Punch and Judy homers just making it over the fence?”
Thank you (and the many commenters) for cementing this hypothesis with a boatload of hardcore analysis.
-Craig
“Altuve and Betts happen to be poster children for the homerun curve flattening in 2016, but you could pick 50-100 guys as examples. That will be the story next year for drafts”
I next want to see a follow-up regarding my August point about 2017. If the trend is trusted as predictive, power-only stars could fall like a lead balloon in drafts, due to their marginal HR utility. If the trend does NOT continue, they will be bargains.