How to Approach Power After 2016’s Historic Spike

Mark Trumbo was pretty great in 2016, but what about 2017? (via Keith Allison & Joon Lee)

Mark Trumbo was pretty great in 2016, but what about 2017? (via Keith Allison & Joon Lee)

Editor’s Note: This is the first post of “Fantasy Baseball Preview Week!” For more info, click here.

Explaining last year’s historic home run surge is an arduous and dubious task. Understanding it, or at least acknowledging it, however, is imperative to crafting your approach to the 2017 season. Never before had major league players hit home runs in at least three percent of plate appearances. (The previous record, set during the heart of the alleged Steroid Era, clocked in at 2.992 percent.)

Many have tried to solve the puzzle, including the Ringer’s Ben Lindbergh and Rob Arthur, to little avail. They outlined the prevailing hypotheses as of July 10, 2016, both their own and of others, both debunked and outstanding; there are so many that you’re better off climbing down the rabbit hole and getting lost in the prognosticating for yourself. But the results, by and large: nada.

I’ll re-investigate these claims briefly in an attempt to understand the inputs to 2016’s outputs. But I’ll devote more digital breath to simply digesting the outputs, understanding the implications as they stand, and evaluating how they might, or perhaps should, affect how you coordinate your strategies for your upcoming 2017 drafts.

Due Diligence

Composition of the League Batted Ball Profile

Major league batters hit 701 more home runs than they did in 2015. Hitters accrued more PAs, but only 951 more — not enough to plausibly explain the difference, proportionally speaking. Due to the ever-increasing strikeout rate, hitters actually put fewer balls in play, and the number of outfield fly balls — calculated as all fly balls plus line drives minus infield fly balls — dipped as well. Alas, in terms of percentages, 2016’s home run surge is even more dramatic.

For a hot second, it seemed the influx of ultra-talented rookies, and the sophomore growth of 2015’s similarly talented rookie class, might have impacted offense in a disproportionate way. Home runs per outfield fly ball (HR/OFFB) indeed made a large jump relative to pre-2015 levels, yet the rest of HR/OFFB increased, too. Even the notion that talented rookies replaced past-their-prime old guys doesn’t work out: rookies do routinely hit more home runs than the outgoing/retiring cohort, but the old guys also saw their collective home run rate spike.

Since Baseball Info Solutions began tracking batted ball data, major league-wide HR/OFFB rate has hovered around 7.21 percent:

hr-offb1

Prior to 2014, the farthest the HR/OFFB rate strayed from the mean was 6.8 percent and 7.8 percent. Since 2014, we’ve seen a range of 6.3 percent (2014) to 8.5 percent (2016). Perhaps the 2016 season was the baseball gods’ way of making up for a relatively anemic 2014 season; alas, the rolling average HR/OFFB rate finally rebounded from its three-year low.

Is there anything to this tidbit? Maybe, maybe not. Anecdotally speaking, I don’t remember much, if any, discussion about the lack of home runs in 2014 — a deficit almost equal-and-opposite to this past year’s surplus. It’s not as if the 2014 season heralded the fall of power hitters or ushered in a new dead-ball era. No fuss was made, and everything returned to normal the next year. Perhaps this simply is regression to the mean. Perhaps.

Composition of Home Runs, Specifically

Ultimately, the HR/OFFB rate increased by more than one percentage point from 2015 to 2016. BIS offers no real resolution; fortunately, the introduction of more granular batted ball data via Statcast may cast light on our bounty of remaining shadows.

Statcast data, while awesome, are not infallible; they have failed to track roughly one of every nine batted balls. While most of those batted balls are likely weak pop-ups and ground balls, neither of which have a chance of leaving the yard, Statcast did miss 97 home runs in 2015 and another 25 in 2016. We’re talking about mere tenths of a percentage point worth of home runs, and video evidence of all of the game’s biggest blasts assures us we aren’t missing outlier data points that would otherwise dramatically skew the results. But it’s worth acknowledging. Due diligence, and all that.

An evaluation of Statcast’s data on home runs reveals the average exit velocity, distance and launch angle on such hits barely budged from 2015 to 2016:

STATCAST HOME RUNS (2015-16)
Season HR EV (mph) Dist. (ft) Launch Angle* Barrels/HR
2015 4,812 103.22 399.62 27.82 71.9%
2016 5,585 103.36 399.49 28.08 74.4%
SOURCE: Statcast
*Vertical angle

Barreled balls per home run increased despite nearly identical stats across the rest of the board. (It is worth noting that Statcast’s raw data do not indicate which hits are barrels. It’s not quite a black box calculation; MLB.com’s glossary vaguely describes the methodology, providing enough of a hint that this author attempted to approximate barrels by linearly extrapolating the EVs and angles using the parameters MLB.com specifies.) On first glance, this coincidence seemed to imply that, while the simple means from one season almost perfectly mirrored the next, their distributions varied. But home runs didn’t get better — they looked almost exactly the same, simply becoming more frequent:

normdist

It’s hard to tell, given I’ve presented the data in discrete mile-per-hour bins, but each season’s distribution of home runs is almost perfectly normal. It’s a relatively unsurprising development given what we know about skills, contact quality, the intersection of the two, and how all have a distribution of possibilities centered on a commonly expected outcome. But it tells us virtually nothing about the increase. Even the marginal increase in home runs (per mph bin) follows a cumulative distribution function. In other words, the increase in home runs itself follows a normal distribution, suggesting it occurred naturally:

cumudist4

A more critical dissection of this information — especially the increase in the barrel rate, all else equal — suggests to me the intersection of exit velocity and launch angle played a more critical role in 2016 — and by launch angle, I mean both vertical and horizontal (lateral). A 370-foot bomb pulled to the left-field foul pole will clear fences infinitely more often than the exact same batted ball to dead center. Perhaps we have right-handed hitters to thank, then, because Statcast indicates that they hit home runs to the pull side three percent more frequently in 2016:

CHANGE IN HOME RUNS BY OUTFIELD THIRDS
Handedness LF CF RF
L +77 -4 +118
R +498 +30 +64
SOURCE: Statcast

Right-handed hitters pulled almost 500 more home runs in 2016. That’s a big deal! Except it’s less of a big deal knowing home runs in all directions increased, obscuring the meager three-percentage point increase in home runs to the pull side by righties (and slight decrease in the percentage of pull-side home runs by lefties). It’s easy to conflate correlation and causation here, and while there’s some merit to this theory, it probably only explains about one-fifth of the surge. The other four-fifths are anybody’s guess. Also, why did the power gains (in relative terms) go almost exclusively to righties?

Approaching the 2017 Season

If we can’t understand why the home surge transpired, we can at least characterize exactly how it manifested itself.

Identifying the 2016 Beneficiaries

I binned hitters by their home run outputs and compared each of the last three years. The lines in the following graph represent the number of players who hit at least X home runs in a given year:

hr-threshold1

More hitters hit home runs at every count, even at the lowest thresholds up until the 35-homer mark. (Note how the 2014 season lagged. While it hugs the 2015 curve more tightly, it lacks spectacularly of power hitters — only one hit more than 37 home runs, as opposed to 11 in 2015 and 13 in 2016.) That much we know and is well-documented by now. The more pertinent question is who hit all those extra home runs.

Given the number of 35-homer hitters didn’t change from 2015 to 2016, you might expect the power surge to be concentrated in the middle tiers of hitters. You’d be right. I calculated HR/OFFB for all hitters who recorded at least 100 PAs in 2016 and followed by each of those hitters’ cumulative HR/OFFB rates from 2012 through 2015, given they each recorded at least 100 PAs in that time span as well. (Hitters who debuted after 2012 are included in these counts and calculations as long as they recorded at least 100 PAs prior to 2016.) I rounded each of their 2012-15 HR/OFFB rates to the nearest percentage point and binned the hitters accordingly. I then calculated the average HR/OFFB among the players in each bin:

hr-offb-changes

The dashed line depicts the break-even point, where a bin’s 2016 HR/OFFB rate equals its 2012-15 rate; the maroon line depicts the actual 2016 rate; and the blue line depicts the number of hitters in each bin. The game’s best power hitters actually experienced a slight decrease in production per OFFB, whereas lesser sluggers achieved a nice boost. It seems as though Mark Trumbo enjoyed all the luck — the rest of the hitters in his 14 percent HR/OFFB cohort collectively lagged behind their previous four-year rates. If the ball was juiced, wouldn’t you expect to see everyone benefit? Or am I misunderstanding this?

Handling Specific Players

Many name-brand players broke out in big ways or revitalized their careers last year as part of the power surge. How should a fantasy owner treat the following five players, all of whom hit many more home runs than the year prior in roughly equivalent playing time? I’ll dissect each one with references to FanGraphs’ FANS projections, which are aggregate projections generated by FanGraphs readers:

  • Mark Trumbo: Trumbo barreled up almost twice as many batted balls as the year before. His average EV remained virtually unchanged, testifying to his increase in well-hit balls as well as poorly hit balls (the latter as evidenced by his decreased GB EV). The FANS expect only 30-homer power (33 homers), which is reasonable enough, but NFBC ADP shows he’s being valued far more bullishly than other one-dimensional sluggers.
  • Brad Miller: I always hoped Miller would fulfill his power-speed dual-threat promise, but I don’t think any of us expected this. It doesn’t seem like he sold out much for his power, and Statcast doesn’t see much of a change in his EVs. That his power doubled overnight is suspect. Jeff Zimmerman noted Miller made an attempt to sell out, too, but you figure you’d see more of it in his peripherals. I’ll take the under on the FANS (24 homers) — by probably half a dozen.
  • Ian Kinsler: Kinsler’s average EV, both in general and specifically on fly balls and line drives, improved slightly in 2016. But he actually barreled the ball up less than in his anemic 2015 campaign, and his barrels per batted ball ranked in the bottom third of hitters. It didn’t do much better in 2015, either, making last year seem especially fluky. The FANS are sold, though (21 homers), so there seems to be an evident fade opportunity here.
  • Brian Dozier: Dozier’s power kept improving, but nobody saw this. Statcast shows a decline in EV and the exact same number of barrels, leading this author to think there’s some luck involved. The FANS agree, although projecting 30 home runs would still make it his second-best season in terms of home runs.
  • Adrian Beltre: Beltre, entering his age-37 year, broke out of a two-year “slump” in which he failed to break the 20-homer threshold. Like most others, his Statcast metrics improved, albeit marginally, but his ratio of home runs to barrels improved disproportionately. The last time his fly ball rate spiked (in 2011), it immediately crashed back down to earth; I’d expect the same to recur. Beltre’s a legendary player, but it’d probably be unwise to bet on anything close to a repeat. The FANS generally agree (23 homers).

There are so many more we could discuss — many had career power years (Jonathan Lucroy, Rougned Odor, Freddie Freeman, Carlos Santana, Jedd Gyorko, etc.) and many more returned to form (Miguel Cabrera, Robinson Cano, Victor Martinez, Mike Napoli, Evan Longoria, etc.) — but it would take all day. A thorough inspection of a player’s underlying data will help inform your judgment of the legitimacy of his 2016 gains. However, without a concrete explanation for the astronomical spike in home runs, our faith in evidence crumbles and steers us toward the uncertainty of belief.

A General Approach to Home Runs in 2017

At this point, you belong to either one of two factions: the Believers and the Nonbelievers.

The Believers think last year’s gains will carry over into 2017. The guys who had career years have tapped into previously untapped potential, establishing new levels of talent and dancing on the graves of their former selves. The Nonbelievers expect the power to regress — the same way all metrics regress every year.

Count this author as part of the latter group. If the ball was juiced, who’s to say it will still be juiced in 2017? And if the ball wasn’t juiced, what evidence is there to believe hitters did anything dramatically differently to achieve such a dramatically different outcome than the previous year? I’m intrigued by the notion that 2016’s home runs are the baseball gods’ recompense for a lackluster 2014 season. If we assume each hitter has a particular distribution of outcomes, isn’t it possible, albeit not highly probable, that an abnormal number of hitters outperform their expected value, thus skewing an entire season’s distribution of performance? Isn’t that exactly what happened in 2014? No extenuating circumstance existed for us to expect power had cratered for good. (Indeed, it hadn’t.)

No matter your perception, I leave you with these parting thoughts. If you’re a Believer, power will run deeper than usual. The evidence I presented suggests the game’s best sluggers didn’t benefit from the 2016 power surge, making their contributions marginally less valuable than they would have been otherwise. It’s likely, then, that the game’s best sluggers are being overvalued in this context. If you’re a Nonbeliever, power is not as plentiful as it seems. The game’s best sluggers — the downtrodden Chris Carter’s of the world — will have their marginal values restored, and all (or at least most) of baseball’s biggest power surprises will return to form, lest their peripherals veritably demonstrate change.

Ultimately, you shouldn’t approach your draft too differently, fundamentally speaking, from how you normally would. There will be busts at the top of draft boards and lottery tickets at the bottom, just as there always is. But if you fail to hedge your bets and invest too heavily in last year’s surprise breakouts and bounce-backs, you may wind up empty-handed if (when) the power dissipates.

References & Resources


Currently investigating the relationship between pitcher effectiveness and beard density. Two-time FSWA award winner, including 2018 Baseball Writer of the Year, and 8-time award finalist. Previously featured in Lindy's Sports' Fantasy Baseball magazine (2018, 2019) and Rotowire's Fantasy Baseball magazine (2021). Tout Wars competitor. Biased toward a nicely rolled baseball pant.
newest oldest most voted
Ryan
Guest
Ryan

The reds had a historically bad pitching staff, giving a boost to home run production. A mix of better hitting and worse pitching?

Ross
Guest
Ross

“If the ball was juiced, wouldn’t you expect to see everyone benefit? Or am I misunderstanding this?” Sure everyone would benefit, but not equally. Wouldn’t less powerful hitters enjoy more of the benefits of a juiced ball, as the “Changes by HR/OFFB” chart seems to indicate? Since a smaller percentage of their OFFBs were going over the fence previous to 2016? (I.e., someone who hits it 350 ft can now hit it 365. Someone who hits it 400 can now hit it 415+…given equal batted-ball profiles, the first hitter will see a larger increase in HR/OFFB than the second.) In… Read more »

crew87
Guest
crew87

This is a good comment.

Rob
Guest
Rob

One could argue that given the physical properties of a juiced ball there is a minimum threshold of latent power talent that needs to be crossed before large gains occur. For example, I don’t think a juiced ball would affect Dee Gordon’s home run totals all that much. Overall I think you comment is spot on, but I would caveat that the very weakest power hitters might not benefit either.

dominik
Guest

I think the reason why the elite sluggers might benefit less is that their good OF flyballs leave the yard anyway. There are no extra points for a 450 hr. Of course they also hit flies that don’t leave the yard but those are mostly non ideal contacts. Some of those non ideal contacts might go out now but most probably still not. The guy who benefits is the guy who hits his best flyball to the warning track (and occasionally out in a small park or with tailwind). Actually now making contact and hitting a lot of balls well… Read more »

dominik
Guest

Likewise the zero power guy doesn’t benefit either because his best flyballs now land 90 feet short of the warning track instead of 100 feet (dramatization).

Michael Bacon
Guest
Michael Bacon

Still waiting for an explanation for the huge increase in balls the chicks love in 1987…

dominik
Guest

I think the league overall got closer together. The number of 20-30 hr guys went up but the number of 40 hr guys not really and the number of 50 hr guys is non existent. If it is the ball I think it is mostly the fringy power guys (low double digit guys) who benefit. The real sluggers hit it out anyway and the zero (0 to 5 hr) still cant hit it out. So who benefits is the guys with mostly doubles and warning track power who now can hit it over the fence more often especially if they… Read more »

Ryan Brock
Guest
Ryan Brock

“If the ball was juiced, wouldn’t you expect to see everyone benefit? Or am I misunderstanding this?”

IMO, Andrew Dominijanni provided the perfect explanation for how a juiced ball could have uneven benefits – http://www.fangraphs.com/community/kinda-juiced-ball-nonlinear-cor-homers-and-exit-velocity/

TL;DR, the coefficient of restitution of a ball is nonlinear. Changing the slope of that nonlinear COR would mean lower EV balls get a bigger boost than already-high EV balls.

Andrew Dominijanni
Guest
Andrew Dominijanni

Hey, thanks for the shout-out. Indeed, although I’ve received some advice on how to further refine and improve my batted-ball model, I have no reason to rule out my fundamental suspicion.

I don’t think we’ll ever be able to get independent corroborating evidence (e.g. impact measurements with the ball), and I’m sympathetic to arguments about a perfect storm of lateral angle, vertical angle, and exit velocity that produced this spike without change in the ball. I think this is one of the most fascinating things to watch early in 2017.

Fake Yeezy Boost 350 V2
Guest

This is a mystery release, with the sneakers coming packaged in a black bag inside their box that hides which colorway is inside. Shoppers won’t be able to pick between the black and brown pair, and won’t know which they’ve got until they purchase the shoes and open the bag