It Remains a Game of Constant Home Runs
A few minutes ago, I turned on a game between the A’s and the Twins, which is a sentence no one has ever written before. I turned the game on just in time to see Jharel Cotton pitching to Danny Santana. You might not know very much about Danny Santana. In fact, you probably don’t know very much about Danny Santana. That’s okay. Context clues. Relatively speaking, he’s a little dude. Hits a bunch of grounders. Utility player; reserve; Minnesota Twin. Here’s what Santana did:
Off the facing of the second deck. Okay. Good one, Danny Santana.
You know that, last season, home runs spiraled out of control. It all technically began around the 2015 All-Star break, and I don’t need to go back over this, because it’s been the subject of dozens, if not hundreds, of analytical articles. At this writing, on May 4, 2017, Yonder Alonso is second on the A’s in home runs. The following players all lead their teams in home runs, either in a tie or by themselves:
Home runs have been hit by home-run hitters. Home runs have been hit by non-home-run hitters, such that they’re turning into home-run hitters. We probably need to adjust our baselines, because, in short, home runs are everywhere. This isn’t a new article. It’s not fresh, I don’t think. It’s just an article updated with the newest data. The home-run spike is continuing. Whatever’s driving it hasn’t yet been phased out of the game.
In this plot, you can see league-wide home runs per batted ball, going back to 1954. Sometimes people prefer to look at home runs per batted ball in the air, but it doesn’t matter what you do. The trends are the same.
So far this year, the rate stands at 4.5%. That would be the highest rate in the observed window, topping last year, which was the highest rate in the observed window. Although the increase over last year is slim, it’s still +3%, which isn’t nothing. And, importantly, last year’s data point was extreme. Generally, extreme data points subsequently regress. This one has anti-regressed, because, probably, the mean to regress to is shifting.
Now, one thing that’s obviously true about 2017 is that it’s far from complete. So why don’t we look at something like the above, except only for April statistics?
Home runs this April were up almost 12% over last April. We have an overall sample of 64 Aprils, and this April’s rate of 4.50% stands as the second-highest, behind only April 2000’s 4.53%. In total, we have a sample of 377 individual months, and this April’s homer rate stands as the fifth-highest. It’s way too soon to conclude anything, but I will point out that based on the extremely early May 2017 data, we’re looking at perhaps the biggest homer month ever. It makes sense that May would pick up where April left off. Why wouldn’t it? Nothing is changing. Except air’s getting warmer.
It’s not that hitters are hitting dramatically more fly balls. They’re just leaving the yard more often. Oddly, according to Baseball Savant, league-wide exit velocity is slightly down, even when you narrow to only balls hit in the air. Look, I’m not going to solve anything. I can just share with you what I know, and what I know is that the basic numbers are saying.
This is an era of rising strikeouts and rising home runs. Unsurprisingly, then, home runs are playing a bigger and bigger role in offensive productivity. Here, you can see what some people refer to as league-wide Guillen Numbers, over the course of the wild-card era. This shows the rate of total runs scored that score on home runs.
Between 1995 – 2015, on average, 35% of all runs scored came in on dingers. Last year, that number jumped to 40%; so far this year, the number has jumped to 42%. To try to split things out, here is average runs scored per game, again during the wild-card era. This shows home-run runs per game, and non-home-run runs per game.
It’s nothing you wouldn’t expect, but it’s a pretty stark visual display. Right now, in the average game, you’ll see 1.86 runs score on dingers. That number hasn’t been so high since 2000. But you’ll see 2.61 runs score not on dingers. That number hasn’t been lower in these 23 seasons. Compared to 2000, it’s a drop of 19%. The drop here is gradual, so it’s nothing you’d notice if you just took in a single game at the ballpark, but it’s the easiest thing in the world to explain. Pitchers are better than ever. Strikeouts are higher than ever. It’s tougher than ever to string together walks and hits. Offenses increasingly need those home runs to score guys from first and second, to say nothing about situations when the bases are empty. This is how the game has evolved.
It’s not like every single player has adopted the same playing style. We’ve still got Ichiro Suzuki and Billy Hamilton and Ben Revere and, today aside, Danny Santana. There are still productive outs and there’s no pervasive lack of throwback fundamentals. Everyone is amazing. But as Dave wrote about several weeks ago, we’re seeing more of the three true outcomes. And if you thought for a second over the offseason that something was going to change, well, nope. Doesn’t look like it. For whatever reason, the ball is still flying. I don’t know why it wouldn’t. 2017 looks like it’s going to look a lot like 2016, if it doesn’t turn into something even more extreme, and I’m not going to tell you how to feel about that. I’m sure you’ve already made up your mind.
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.
Only one obvious solution:
Gravity has been tampered with.
Look at me in the eye and tell me you are lighter than you were two years ago.
For whatever reason, this is the ad that is running on this page just above your comment: http://i.imgur.com/4BonghD.gif
“A few minutes ago, I turned on a game between the A’s and the Twins, which is a sentence no one has ever written before”.
You had me at hello, Jeff.
Let the players go back to juicing. Looks like it was helping the pitchers more than the hitters.
(note: comment may not be a serious suggestion for more reasons than I can count)
Thnx. Was wondering when one of you guys were going to tackle this. It was a crazy month for the long ball with so many games having 4 or more HRs hit.
I don’t know it reflects a lack of contact skills in favor of the all or nothing approach, or FO’s are seeking this out by design. Whichever it is, there always seems to be some team(s) that will exploit any extreme tendency by filling the void that it creates, so I’m curious to see where this goes… my sense is a growing demand/premium for players with great contact skills.
The Rawlings factory that produces the balls used in major league baseball is located at the foot of a volcano which began a recent period of activity in October of 2014. You can look wherever you’d like, but I blame the new “explosiveness” of baseballs on a thin layer of Turrialba volcanic ash.
I thought this was a joke but it turns out you’re serious about where they are made and the volcano suddenly becoming active (although I think it became active in 2015).
http://www.stuff.co.nz/world/americas/73481751/amazing-footage-of-a-volcano-in-costa-rica-erupting
This is going to be my go-to theory for this phenomenon too.
This is… perhaps more compelling than it otherwise should be, if only because we’ve ruled out basically any other possible cause anyone can think of
Oh, I’m definitely serious about the factory and the volcanic ash – it gets on everything in the central part of the country – though the effect that it could possibly have on baseballs is just a joke. I live in Costa Rica and have driven past the factory, I have friends that live just a few blocks from there, but I’ve been told they don’t give tours.
If you’re really looking for a conspiracy, though, in mid-2015 they laid off about a third of the workforce there. They cut all of the non-baseball employees, moving the apparel division to another plant. I don’t know if a such a huge cut or change in the plant may have effected the baseballs being produced.
http://www.ticotimes.net/2015/08/04/rawlings-lays-off-200-employees-costa-rica-operation
A couple of questions:
1. All of this talk about exit velocity, flyballs, and etc. but you point out that EV is slightly down. Is that slightly down from last year from when the spike in homeruns was noticeable? How does batter EV compare to 3, 5, 8 years ago especially post-steroid era?
2. Danny Santana may be undersized, and underproductive, but what about the fact that we are talking about exceptional professional athletes? Why is it hard to believe that organizational focus on adapting to harder hit flyballs at an optimal launch angle is impacting the Danny Santana’s of the world? Is it truly crazy for Mookie Betts or Rougned Odor to hit 30 home runs at their size if they’re actively developing their game to hit them?
3. What about the relative lack of dramatic change in total runs scored relative to the change in home runs? Just going all the way back to 1980 here, 2016 ranks 14th in terms of runs scored and 14th also in Slugging %. Even shrinking the limits to go back to 1990 those ranks do not change.
4. To the juiced-ball theorists out there, wouldn’t a tampered baseball leading to more homers increase slugging % in general? A change in the baseball wouldn’t change contact rates (which has decreased) or walk rate (which has increased recently) but wouldn’t it still result in a higher percentage of extra base hits and/or isolated slugging %? Obviously with so many home runs iso is way up, but compared to other years with 5000+ home runs it doesn’t look abnormal. Even taking a more general approach, the extra weight towards extra walks and home runs at the cost of contact and ground balls hasn’t effected .wOBA (.318 league average in 2016) like one would assume. For extra comparison in 1997 about 1000 less home runs were hit yet slugging percentage was actually higher that year (by about .002) and league .wOBA was an excessive .013 higher than last year.
So I guess what I’m really saying is that relative to the last 5-ish years of weak offensive baseball (especially in comparison to the steroid era) the recent surge in home runs doesn’t seem to be “abnormal” in effect on the general game. Isn’t asking “why is this happening” simply answered by “conscious organizational development to turn Khris Davis into a a legitimate baseball asset instead of a fringe platooned power hitter?”
It’s harder to believe with Danny Santana because he’s nowhere near the hitter of Mookie Betts.
Is there an organizational philosophy of this sort on the Twins? Did every organization in baseball suddenly adopt this philosophy during the AllStar break in 2015 without telling anybody? Yet they somehow told the players, and the coaching staff, but except in a couple of cases they’re all downplaying it?
I think it has more to do with the rookies coming up then at 2B/SS/CF. Those are very different player types than the players they replaced – who grew up only knowing what SS do during the steroids era – and the MLB adjustment time was shorter. Just looking at 2014-2016 qualified PA for SS – Santana, J Ramirez, W Flores, E Suarez, Bogaerts, J Baez, Lindor, Kang, Correa, Russell, Marte, Seager, Story, A Diaz, Anderson, G Garcia, Peraza, Polanco, O Arcia. That’s a lot of turnover at one position.
Maybe. But that hypothesis was floated last year, and people tried to prove it in a variety of ways, with no luck. In the Hardball Times Annual, Jeff Sullivan specifically looked at rookie production, trying to see if he could account for it, and he couldn’t (this is mentioned in the 538 piece, which tries to prove it in another way with no luck).
Again, this is what Sullivan meant by “dozens, if not hundreds, of analytical articles” — all sort of theories have been mooted to explain the home run spike, and all of them have been investigated. If somebody had found a satisfactory explanation, we wouldn’t still be talking about it.
Seriously, people have rigorously investigated a lot of ideas. If you’re some random commenter on a baseball blog, and you think you have the answer, odds are somebody already thought of that a year ago and tried to prove it, and failed.
I’m not talking about generic rookies. I’m talking about the positions historically valued for defense (C/2B/SS/CF) rather than offense – and for speed not power (so exc C). And it wouldn’t just be the rookies. Vets at those positions would be very aware that a different type of young player is trying to take their job. They’d be nuts not to adjust to that – esp since speed and contact hitting (and defense for that matter) is devalued now.
From BRef:
Def/Spd Posn (2B/SS/CF) – 1217 HR’s in 2015 and 1624 HR’s in 2016. 407 HR (33%) difference.
Everything else – 3665 HR’s in 2015 and 3962 HR’s in 2016. 297 HR (8%) difference.
Don’t have the month-by-month but its prob similar.
Alan Nathan article on 2016 data: looked like exit speed had increased specifically for air balls in the HR launch angle range. So didn’t really look like the ball there.
http://www.hardballtimes.com/exit-speed-and-home-runs/
You are misrepresenting his research and his comments. All of the data support the notion of a “livelier baseball” except for the fact that a small band of launch angles in the “line drive range” curiously show no increase in EV.
However, that anomaly has a good explanation which comports with the “juiced ball theory.” It didn’t get much press, but it’s out there. You can find it if you look hard enough.
I didn’t misrepresent what he wrote then. If he’s found a more conclusive interpretation later, please do link to it.
I will agree with you that you did not “misrepresent what he wrote or concluded.” Let’s just say that you didn’t tell the whole story or that there really isn’t a story to be told other than reading the entire article.
The pertinent part of Alan’s article is this:
“And I have to admit that I am very puzzled by this plot. While there is a clear increase in mean exit speed for launch angles in the home run sweet spot, 200-350, the mean exit speeds are essentially identical for line-drive type angles, 00 to 100. This is not what I would expect to find if the ball were indeed juiced. If anything, I would expect the coefficient of restitution to have a greater effect on the line drives, which are generally more “squared up” than fly balls, as one can see from the higher average exit speed. This does not bode well for the juiced ball theory.
My conclusion that the higher exit speeds account for most of the increase in home runs still stands. However, as much as I hate to admit it, the exit speed versus launch angle plot really puzzles me. Sometimes it happens that the answers are not as clean and crisp as we would like; this seems to be one of those times. Looks like I’ll be spending more time on this one.”
I don’t think he phrased his conclusions very well. I will phrase them better:
The data clearly suggest a juiced ball. Except that a juiced ball theory would result in greater EV for line drives as well as for other LA, so that throws a wrench in the “juiced ball” theory, although it by no means repudiates. As Alan does say correctly and with clear articulation, sometimes there are results that are puzzling and do not seem to comport with ANY accompanying theories.
But wait there’s more…
I don’t know whether Alan wrote about this or not, and as I said, it has not been widely reported by anyone, but the prevailing theory on why there was a large increase in fly ball EV but not on LD EV is this:
Early on in the Statcast data dump, they were missing lots of low LA balls. Most of these were soft batted balls with low EV. Therefore the numbers you see in 2014 or 2015 for low LA balls are artificially high since many more low EV balls are missing than in 2016. THAT, if true, would explain the conundrum that Alan found.
If we accept that, all of the data perfectly support the “juiced ball” theory. Basically at every LA, the EV has significantly increased from pre-ASB 2015 to 2016. The only explanations for that are either faster swings from the batters due to them all of a sudden getting stronger, or a significant change in equipment.
Now, I don’t know whether Statcast has added or changed data, but downloading the first month of the season, from 2015 to 2016, I get this for LA of 10-20 degrees:
2015: 92.7 (N=2644)
2016: 93.4 (N=2602)
So, based on the N alone, there do not seem to be missing data in 2015 compared to 2016, and the EV has increased by .7 mph, which is half of what Alan expects in order to account for the HR increase. I disagree with Alan’s expectation of a 1.5 mph increase to account for a 25% HR increase, because I think that once batters change their approach by hitting more FB (increasing their LA), which they did, we don’t need an extra 1.5 mph in EV. We need less, because when the EV increases it becomes correct for batters to hit more FB which “compounds” the increased HR rate. Alan’s estimated 1.5 mph increase (in order to account for the HR surge) assumes no change in batter approach.
In any case, I think Alan’s conundrum of the “no increase in EV between 10 and 20 degree LA” has been solved one way or the other (either missing data in 2015 or he erred with the data), which leaves us with the conclusion that a juiced ball is the likeliest explanation for the increased HR%/scoring.
That exit velocity thing makes me wonder… If we think this isn’t just a result of the players playing differently somehow, e.g. the launch angle revolution (and the lack of more fly balls tends to suggest that’s not it, though far from conclusively), then it seems like balls are flying better through the air, rather than coming off the bat harder. That’s not remotely implausible: the golf ball industry, for instance, has known for over a century that the texture of a ball’s surface can affect how well it flies. So I wonder if they’ve changed something about how the balls are interacting with the air.
Well, yeah, that is one variation of the “juiced ball” theory: the NCAA explicitly did something like this just a couple of years ago. Trouble is, people have tried to find a difference (both in terms of bat-ball collision and aerodynamics) between pre-HR-spike balls and the current ones, and haven’t turned up anything significant. (There are plenty of articles like that written since that one — this is part of the “dozens, if not hundreds, of analytical articles” Sullivan mentions)
There’s not much you can do to a ball that makes it fly further through the air, but there are some things, as your NCAA link explains (seams).
However, no one that I am aware of has addressed these things. Your other link, to the 538 article, definitely does not speak to anything other than what would produce an increased EV.
Not to start anyone on a political bender, and I hope it doesn’t devolve to that, but climate change maybe? Temperatures are rising year after year, and the ball flies farther when it’s hotter.
I would hope that an average temperature that’s just a degree hotter than a couple years ago* isn’t suddenly making baseballs fly 20 feet further.
* I have no idea what the actual difference is, but I’m pretty certain it’s relatively minute.
Even though there’s no reason to think a trend that amounts to fractions of a degree per year would create a sudden spike like what we’ve been seeing, this theory seems to come up often enough that Alan Nathan felt obliged to write a paper about it.
Thank you very much for linking that article. I was thinking more in terms of day-to-day fluctuations than average changes, but this still takes the wind out of my suggestion (har, har).
No, a degree here or there will have only the tiniest impact on HR rates and run scoring. Run scoring has risen 1/2 run a game since 2014, the year before the HR/run scoring spike (after the ASB).
Run scoring goes up around .03 runs per game for every extra degree in temperature. Unless we suddenly move very much closer to the sun and everyone starts dying, we’re not going see much of an effect on HR’s and run scoring from a rise in temperature.
And OF COURSE if one ever suspects something like that, simply look at HR rates and run scoring in the dome stadiums. You will naturally find the same HR and run scoring spike as you do everywhere else.
How many parks have moved their fences in over the past 20 years? Does not having less outfield grass to cover, and less distance needed to hit a HR over the fences not contribute to more homeruns and less non-HR R/G? This is exactly what the owners and sponsors want anyway: More homeruns, less annoying things such as strategy, more pitching changes and therefore more commercials. These changes are very well liked by Progressive, StateFarm, Nationwide and yes, football fans who have to ‘endure’ baseball games at bars. I would also guess that the rate of triples and ITPHR are also down because they are two of the more exciting plays in the game.
A) Park sizes have not changed to any considerable degree over the last 50 years, let alone 20. We can easily check to see whether that is true or not. And it is.
B) We are mostly talking about a huge spike in runs and HR since the ASB in 2015. Unless the teams secretly all moved their fences in around 10 feet in the middle of the night, this is not an issue.
Whatever is causing it – and honestly I suspect it is a consequence of the saber emphasis on what happens in the batters box and on individual players – I really don’t like it at all.
It kind of reminds me of the rise of the 3 in the NBA over the last couple years.
In theory, more three true outcomes would lead to less value being placed on defense, which would lead to more offense-first players, which would lead to even more homers.
Less value on fielding, on baserunning, on speed, and on contact hitting. Those positions which have historically emphasized those (SS/2B/CF) are the ones which make the biggest fastest adjustment as their fielding/speed value declines. But yet – WAR still pretends that those are premium positions – because of fielding/speed/etc.
What baseball needs now is to play Dan Vogelbach at SS. Or maybe some true lefties.
In looking into this issue last year, IMHO, it is advanced metrics have changed the market which has been a powerful force. Power is being democratized across a larger group of players.
The majority of the increase in HRs last year were hit by players in the two smallest size (physically) quintiles. They didn’t hit the ball further, but take a look at the increase in the number of well hit fly balls (90MPH+, 15deg+):
Smaller than average players +1,323 (423 More HRs in 2016)
Average or larger -96 (282 More HRs in 2016)
Younger players also showed a significantly higher contribution:
HR Change (2015 to 2016)
21-26 yrs +729
>26 yrs -24
Both of the above would support a market driven change – that the players are responding to the difference in how they are valued. It certainly can’t be the ball, atmosphere or anything else that just changes for the smaller and/or younger players. Now that would be a conspiracy! More if you want to read here (http://www.fangraphs.com/community/the-home-run-conundrum-part-ii-less-is-more/)
Whenever an extrinsic force is applied to increase HR numbers it will always affect the players with less power more than the players with more power. So that really has nothing to do with anything. Why is that?
We can answer that with an extreme example (which is often a good way to answer these kinds of questions). What if a player can’t generate enough power to ever hit a home run? The best he can do is hit some warning track fly balls? Juice the ball or move the fences in and he hits some HR’s. His HR rate goes up by infinity (from 0 to some positive number). Say there’s another player who hits the ball 450 every time he hits a HR. How much does his HR rate go up with a juiced ball or shallow fences? None. That’s sort of the way it works. Not exactly like that, but sora.
Is WAR adjusted to account for the fact that we’re seeing more three true outcomes?
That is irrelevant to WAR other than the fact that the values of the individual components can change slightly with the run environment and the way runs are scored. Slightly.
I don’t understand why it would be irrelevant to fielding components of WAR or position valuation. SS are already less involved in the fielding game than they have been in decades (since the early 1950’s). A TTO focus exacerbates that long trend. At what point does SS stop being valued for defense – or an Ozzie Smith-type simply is overkill on defense?
It is “irrelevant” because the fielding metrics already account for the number of balls in play. For example, if we got to the point where everyone either K’s, walked, or hit a HR, everyone would have a UZR of 0 and defense would not be part of WAR. You would also be surprised at how little defense IS effected by a small change in the percentage of balls put into play.
Well, I am at least glad that we got rid of the steroids that were causing all those home runs.
The Home Run spike is like a biotech stock that has a huge price spike and channel upwards. Some event, like only one, occurred to break the previous trend and since then the players have seen the higher HR rates and have done things to make the rate go even higher, albeit at a slower rate compared to the spike.
In this case, if the ball is juiced, it behooves players to try and hit more fly balls (again, think of the extreme, where the ball is so juiced all fly balls are HR or so dead that no fly balls are HR, and then think of the proper strategy) such that the effect of a juiced ball is indeed compounded.
Indeed if the ball were all of a sudden de-juiced by a lot, we would not see quite the lower run scoring that we would expect if we modeled the game with a de-juiced ball. Players would change their approach to mitigate the run=depressing effects of a de-juiced ball (hitting fewer fly balls, trying to hit more line drives, and walking more).
I like the home runs but if the league is going to make an adjustment towards more contact like some suggested (maybe a smaller zone) I think they would need to counter that with something that depresses power a little (different ball maybe? or make ballparks bigger or at least walls higher if you can move back the fence?).
I think 4-5 runs per game and 3-4% HR on contact is good but if a contact promoting adjustment would raise the HRs even more it would not really be good for the game.
fans love HRs but part of its appeal is that it is rare, nobody wants to see a slow pitch softball game.
I don’t think HR should go down but if starlin castro or freddy galvis can hit 20 it means that HR are already losing some of its appeal so it at least shouldn’t go up much more from now.
baseball needs heros and part of this is HR hitters separating from the back. but what we see is the opposite, everyone now hits 25 (I think like 60 guys or so last year) but still nobody hits 45+. the field is much closer together and that adds to the HR appeal loss. for example yasmani tomas hit the same number of HRs like mike trout last year (yeah I know different park etc. but still).
that is why people found the steroid era HRs more exiting. the overall HR number was the same but the gap was larger, as some guys hit 60. I think the league should give miguel sano a better chance to separate himself in HRs from starlin castro and guys like that.
it is probably a good sign for the integrity of the sport that nobody hits 60 anymore (steroids…) but from an exitement standpoint it is not great if the best guys hit 40 and the average joes hit 25 (slight exaggeration but you get the point).
Re: the distribution of HRs — I understand the perception, but I think the recent relationship between top HR guys and the average player is actually quite normal.
For each season, I calculated the difference between
(a) the average of that year’s top 20 in HRs, and
(b) that year’s overall HRs per 600 PAs.
Omitting the strike years of ’81 and ’94:
— +20.9 is the average gap for the DH era.
— +20.7 is the average for the expansion era.
— +20.0 is the average for the post-WWII era.
The past 2 years fit right in there:
— +20.6 was the average for 2016 (38.8 for top 20, 18.2 MLB avg per 600 PAs).
— +21.7 was the average for 2015.
Of course, this is just one way to measure the distribution. But in this measure, at least, it’s the steroid era that was abnormal.
regarding the EV statcast might have gotten better at recording slow rollers and pop ups which previously were often missed by the system.
I think we are overlooking something pretty basic. What if pitchers are just missing more over the plate? I’m a Yankees fan, and a lot of the homeruns hit by Judge are meatballs and hangers. If you throw 95+ it speeds up your bat. Hangers get hit harder, and meatballs get hit harder because bat speed is faster. Does Statcast track bat speed? I would love to see the changes in bat speed.