The Unexplained Home Runs
On Saturday and Sunday I attended my first Saber Seminar, along with a number of other members of the FanGraphs team. It was a great weekend and Dan Brooks and Chuck Korb were excellent hosts and organizers. In a department of chemistry lecture hall on the campus of Boston University, a space where periodic tables flanked the stage, we saw a number of interesting research presentations.
One of the lectures I was most interested in attending was one of dealing with the physical sciences, a presentation by Dr. Alan Nathan, who has of course become a superstar in the field of baseball analysis and research. It is Nathan who, along with others, has helped writers like Ben Lindbergh and Rob Arthur investigate whether the ball is juiced.
Much has been written here and elsewhere about the home run surge that began in the summer of 2015 and has shown no signs of waning, but rather keeps increasing. In part due to the suddenness of the game changing from an extreme run prevention environment to an extreme power environment, it sure seems like the ball has something to do with it — if it was not the primary culprit. Whether MLB and Rawlings are consciously or inadvertently altering the ball — and it would seem like it would have to be a material (and/or size) change since the major league ball is hand-stitched — the ball is playing differently.
But while Nathan has assisted in other research, I was curious to learn the findings of his own study, which he presented at Saber Seminar. I am not going to focus too much on the details or process of Nathan’s study. He may publish an in-depth piece soon himself for those who could not attend his talk. Instead, I want to focus on one aspect of the presentation: the unanswered question — the home runs that the ball cannot account for.
I will share some background of his study. Nathan’s data was taken from games played at Tropicana Field, which is the most pristine testing environment in the majors as it rests at sea level and is of course a climate-controlled dome. Nathan also was only studying balls hit there over a three-year period that had a 90 mph or faster exit velocity and launch angles between 25 and 30 degrees. In other words, Nathan was concerned with balls most likely to become home runs.
(Edit: Nathan was studying how the ball carries from year to year. COR was not studied).
I shot the following photos of the presentation with an iPhone 7:
What Nathan found was that the ball did change, it had less drag, in comparing 2015 to 2016. But what was most interesting to me is the ball has apparently not changed from 2016 to 2017.
But HR/FB rates continue to increase — 13.7% this season compared to 12.8% in 2016 — even as the ball is playing similarly in 2017 compared to 2016, according to Nathan’s research. We can deduce there are reasons behind the spike.
I’m convinced beyond a reasonable doubt that the ball has a role in this surge and perhaps even a leading one, but there are more elements involved and that was a takeaway, for me, from the presentation.
And the other missing variable I strongly suspect continues to be those joining a growing band of air-ball revolutionaries, which can have strong convictions in their swing plane changes thanks to the data tools now available. While pitchers and defenses first enjoyed data-based advantages to understand their own performance and influence strategy, hitters can now easily understand and quantify launch angles and exit velocities, thanks to Statcast and similar tools.
After all, it was not until heat maps based and batted-ball data were widely available — hard evidence in easily digestible forms — that many coaches and players accepted the findings and began to play defense in non-traditional ways. It was not until then that shifts proliferated throughout the game. It was data that reinforced belief with defensive positioning, and I suspect it is doing something similar with hitting.
This is a fun subject for me, and for many us, because there is science involved, but also an element of mystery. And what Nathan, I think, demonstrates is there is not one smoking gun explaining the surge, but multiple smoking guns.
The ball is juiced, hitters are launching more of them into the air, and their unconventional approaches and practices are being backed by hard evidence from Statcast.
The ball is perhaps no longer changing, but perhaps hitters are.
A Cleveland native, FanGraphs writer Travis Sawchik is the author of the New York Times bestselling book, Big Data Baseball. He also contributes to The Athletic Cleveland, and has written for the Pittsburgh Tribune-Review, among other outlets. Follow him on Twitter @Travis_Sawchik.
I imagine some, if not all of the continued surge has been cohort replacement, as the upside to pull-heavy three-true-outcomes hitters increased in the juiced ball era at the cost of contact-heavy spray types. When more balls go out, the average value of attempting to pull a ball in the air (at the cost of striking out more) increases and the average value of simply putting the ball in play decreases.
I think if you look at changes hitters have made, rookies coming into the league or who is receiving playing time in positional battles, the pull-heavy power hitters are on top in all three areas. Once the ball was changed, teams have been responding to the new incentives by acquiring and playing players who fit the new ball, driving the game further in that direction.
Yeah this is my thinking as well. More power hitters combined with the noted propensity for most/all hitters to increasingly swing for the fences, results in HR/FB not really being comparable over time. It’s not like it’s some magic number that numbers must revert to – strength, effort and swing plane (which could quite possibly involve more efficient and powerful biomechanics, especially if you watch Josh Donaldson talk) are not measured by it.
I’d love to see a plot of HR/FB rates over time vs strike outs, exit velocity and/or launch angle (or some single batted ball metric that combines probability of a home run). Hell, even average player weight would be interesting to look at. I think those would show a partial explanation for this.
As a BU alum, very disappointed that I never knew stuff like this was going on on campus.
I suppose if reduced drag is what he came out with, and we have a sudden blister problem for pitchers of sorts, there are two pieces of evidence that some attribute of the outside of the ball has changed.
I think enough people have looked at this issue that it’s logical to conclude that there is not A reason for the home run surge, but an aggregate of smaller factors – the players, the ball, the environment, and even just some fluke performances, could add up to an extreme outlier. Sometimes a bunch of unrelated variables add up to a perfect storm.
There was a piece (I think here) where they interviewed a pitcher who says that the seams are actually pretty different from ball to ball. Given how difficult it is to repeat pitching mechanics, I wouldn’t be surprised if part of the issue is just the difficulty of pitching with varying balls.
there you go.
another factor would be the bats. Not long ago many bats were splintering… not much these days.
Hitters aren’t launching THAT many more balls in the air. Since 2014: FB% – 34.4%, 33.8%, 34.6%, 35.5%
Fun fact: FB% from 2006-2011 – 36.7%, 37.9%, 35.9%, 37.8%, 37.5%, 36.0%
Yep, every year from 2006-2011 there were MORE fly balls than during the alleged “fly ball revolution”. And the inter-annual variation shows that the recent trend could be no more than just background variation and not an effect of hitters trying to do this (which remains a narrative that isn’t data driven).
But way more of those balls are leaving (2014-2017): 9.5%, 11.4%, 12.8%, 13.7% (From 2006-2011 those numbers ranged from 9.3-10.6%)
It seems that the majority of the effect is the ball, any remaining effect could very likely just be normal variation associated with changes in player identity, weather, ball park factors, etc.
I think the idea of a “fly ball revolution” is poorly named, and rather misleading. A lot of the supposed adopters of this approach have not seen their fly balls increase. However, they have seen their exit velocity increase by quite a bit.
This says to me that the so-called “fly ball revolution” is a more subtle “swing mechanics change”. Some people wind up increasing their launch angle, some their exit velocity, and most don’t do both…but the underlying change is something else being done with the rest of their swing.
Of course, only a few players are supposedly making this change. Again, that points back to the ball as the primary cause.
Right, this may just be a “swing harder revolution”.
“Factors that Affect Drag. Geometry has a large effect on the amount of drag generated by an object. As with lift, the drag depends linearly on the size of the object moving through the air. The cross-sectional shape of an object determines the form drag created by the pressure variation around the object.”
If the seams are higher, than that makes the cross-sectional shape of the object different. Its starting to make sense…
Is this where the juiced-ball story ends? I am betting on different cores, but that would be be hard/impossible to study. Anyone who has played a lot of baseball will know what I am getting at.
All balls are not created equal. Some balls are harder and fly a lot further than others. At the point where a different plant is doing the seams differently, it is certain that the cores are also different. If the superficial appearance and feel of the ball is different, imagine how different the insides are. The inside is literally hundreds of feet of tightly wound yarn. Any changes in humidity or materials would have a huge effect on the ball. This is not the kind of mystery that is solved by a computer model though…
What I find most interesting about the whole thing is the mandatory progressive data-driven narrative that accompanies the juiced balls. We could either cheapen all of the players accomplishments, like we did in the 2000’s or we can praise the data-driven, progressive approach of modern players. Players would have figured this out without Statcast, I am 100% sure of that. You notice when you hit home-runs more frequently. Perhaps MLB learned a lesson from what Bud Selig did as they ruined the marketability of the game through their PED witch-hunt. Is it coincidence that Statcast became public as juiced balls were rolled out? Personally, I am sick of hearing about the fastest/youngest player to hit X home-runs in his first Y games. This should all be asterisked… or maybe MLB learned that asterisks are for suckers! Asterisks honor the past, which is not particularly marketable. This time we are not going to call them cheaters, we shall call them progressive! Personally, I don’t call anyone a cheater – I don’t care what pro athletes do and I like offense. I just think the narrative seems condescending of past players, which I am not down with. Players would have realized that balls were juiced back before the first computers were invented. In fact there is a word for it – the live ball era! You don’t need graphs and reports to realize when balls fly over fences.
> Players would have figured this out without Statcast, I am 100% sure of that.
I am much less certain, the difference is so small. People are really bad at noticing small changes, especially in something statistically driven.
Put another way — for a full season’s worth of PA, for one player, this is like a coin going from 50/50 to 51/49, very slightly biased in favor of the home run. Humans are not skilled at determining these types of odds changes, because of the small magnitude and because variance swamps everything.
Players are narrative/emotion driven, same as the rest of us. It’s only in the huge aggregate that we can see the difference, and that’s only by using statistics and maths, something most players are not known for.
I hear your point, but I think you are underestimating people who do this for a living. This is why it took a few years to really take hold. A few guys noticed that things have changed, then a few more, then a few more…
Guys can tell from the first well-struck batted ball how well it is carrying on a given day, which is not all that far from what we are talking about. When you simply realize that you are hitting a bunch more balls out of the park, you are going to notice that. When you hit a ball that you know was warning tracked that goes over the fence, you notice that. You would notice if some Neifi Perez character started hitting more homeruns – especially if they are not scalded out of the park. For a ballplayer, it is really clear when a non-well-struck ball goes further than it should. Baseball players make a living by hitting balls – it seems random to those not involved, but they know exactly how they struck a ball and what it should do. I think they would notice. I could be wrong about players noticing on their own, but I want to give them credit for not being morons. It is odd to me that people want to give credit to technology over the players, but it is a common phenomenon in modern society.
small adjustment to your statement.
13/25 MLB ballplayers make their living hitting baseballs.
Fair enough. The other half makes their living managing the outcomes of those batted balls.
good point concerning all balls not created equal… or prepared equal I would add. Teams in higher elevations are using humidors (per MLB requirements now). Each team prepares the balls with the umpires. I recall from tending goal for ice hockey games different pucks kept colder flew faster and harder(ouch!).
Sounds right. Hell, every slow-pitch softball player will know what I am talking about. They are literally marketed according to the density of their cores – different associations tolerate different balls. The among the different densities, some make harder balls and they cost more because everyone wants to hit home-runs.
I agree. I think the ball is the explanation for the initial surge, but hitters changing is the primary reason why the surge has sustained itself/increased.
I’m curious to see if power numbers go down in the next couple years as pitchers adjust.
Thanks for sharing this for those of us who couldn’t attend the event, Travis.
It would serve MLB well to further juice the balls. Just saying! At some point all the records will be broken. Or MLB could take the air out of the balls, and pitching could undergo a saber-metric revolution…
like it… mostly.
wonder why if Mr. Nathan’s study was concerned with “balls most likely to become HR’s” you think looking for the unexplained ones could be revealed?
Your idea of the shifting offense via a higher swing slope is spot on. and to add the lag time is also a key factor. Seems to me that offense always follows defense/pitching… could be as simple as play starts with the pitcher.
My own most recent anecdotal observation is that batters (and perhaps managers in setting their lineups) have now completely adjusted to (and rendered worse-than-useless) the old pitching saw about ‘establish the fastball early’. And where I’ve noticed it is the number of pitchers who are having trouble getting out of the 1st/2nd inning when they are supposedly establishing the fastball by throwing too many of them.
Hard to support that with data – though there is a huge tripleslash rampup – and mainly the power component – since 2014 in the 1st inning outcomes (more so than other innings) – from 256/319/398 to 266/335/453 – with slightly more PA in that inning (4.229 to 4.327) and a lot more runs (.4866 to .5796). And fastball has always had the most extreme bip outcomes
I felt last year there was still a mix of dead balls, especially in the first half in certain parks . For example the ball didnt carry as well in YS3 in the first half as in the second half.
Also, the claim that the ball has not changed in 2017 from 2016 seems unverified. The Lindbergh and MGL tests were on the 2016 and pre 2nd half 2015 balls. Perhaps the only change is the seams as more pitchers seem to be complaining about that than in 2016.
It may well be only the strike zone. The strike zone is noticably smaller in 2017 than 2016 leading to more hittable pitches which translates to more optimally hit FB
But the meme that quantification and tech are somehow teaching hitters what we all knew 50+ years ago seems to be held dear by many