The Secret Benefit (And Cost) Of Sweeping Sliders
I, for one, am not a fan of the breakout of the “sweeper” in recent years. It’s not because I don’t enjoy frisbee-ish pitches that seem to pull out a map, ask for directions, and take a sharp turn on their way from the mound to home. A big part of my job is making GIFs of fun pitches, so I obviously love that part. Personally, I just don’t like the annoyance involved in classifying them.
To give you an example, I decided to do some research on sweeping sliders, or whirlies if you’re into weird nomenclature. In fact, that’s what this article is about. On my way to doing so, however, I had to spend some time getting obnoxiously technical. First, I downloaded all the sliders that right-handed pitchers have thrown this year. I separated them by movement profile, then started asking questions.
I asked a few people, “Does this scatter plot look like it separates out sweepers and non-sweepers to you?” It kind of did, and it also kind of didn’t. Are pitches that have 30% more horizontal break than vertical break sweepers? What about 50%? What about pitches that break a foot horizontally but move six inches downward? I sent several variations of that chart trying to nail it down, but nothing felt quite right – I’ll spare you having to look at the mess I ended up with.
The best definition I could find for a sweeper comes from Dan Aucoin, but even then, there are weird bright line issues. Should a pitch that breaks a foot horizontally and 1.75 inches downward be a sweeper, while another that breaks a foot horizontally and 2.01 inches downward be a regular slider? What about its velocity? And what about a pitch that’s a millimeter short of having enough horizontal break? It’s just downright hard to make definitions like this work, a good reminder that pitch classification is extremely difficult when a pitcher isn’t telling you specifically what they’re throwing.
That mini-rant aside, I did eventually end up deciding how to define a sweeper so that I could get on with my research. It’s been a revelation seeing the Dodgers, Rays, and Yankees tap into seam-shifted wake to reinvigorate various pitchers’ careers. Still, I had a sneaking suspicion that as neat as the pitch is, it’s not a panacea. I’ve seen enough games where a righty pitcher floats in a handful of big breakers to his left-handed opposition to wonder how effective sweepers are without the platoon advantage. For every perfectly located back-foot slider, I can call several no-chancers to mind, either crushed for line drives or comfortably taken for balls.
To measure this effect, I split all sliders into two buckets: sweepers and non-sweepers. I further bucketed them by handedness. Here are the physical characteristics of my two groups when thrown by right-handed pitchers against right-handed hitters:
| Characteristics | MPH | HMov (ft) | VMov (ft) |
|---|---|---|---|
| Non-Sweeper | 85.3 | 0.43 | 0.15 |
| Sweeper | 82.6 | 1.17 | 0.19 |
And here’s how both types of sliders have performed in those right-right matchups:
| Characteristics | SwStr% | Whiff% | GB% | PU% | BABIP | HR/CON | Run Value/100 |
|---|---|---|---|---|---|---|---|
| Non-Sweeper | 17.0% | 35.3% | 45.6% | 13.1% | .277 | 3.9% | -0.26 |
| Sweeper | 17.1% | 36.4% | 33.0% | 20.3% | .246 | 4.0% | -0.94 |
Holy mackerel, that’s a low BABIP. If you wonder why the Dodgers are allowing a .258 BABIP this year, their defensive positioning surely helps, but so does the fact that a pitch they popularized turns balls in play into outs at a heady clip. I included popup percentage (which I manually defined as batted balls hit at 45 degrees or higher) because I was reminded of Owen McGrattan’s research into that very phenomenon. Want some outs on balls in play? Get them in the air at a lazy angle. It’s not rocket science, though Los Angeles probably employs plenty of rocket scientists.
If you want a short encapsulation of the sweeper’s advantages, that chart is exactly what you’re looking for. The pitch misses just as many bats as a “regular” slider, and gives up just as many home runs per ball in play. But it gets more soft fly balls and popups, which results in a much lower BABIP and thus much better outcomes for the pitcher.
When you put it that way, it’s no wonder pitchers are trying to throw more sweepers. If your pitching coach came up to you and said “Hey, you have an average slider, let’s turn it into an average sweeper,” why would you say no? It’ll miss just as many bats as your old pitch, and also get better results on contact. If you’d prefer it in wOBACON terms, sweepers allow a .325 wOBA on contact, while “regular” sliders allow a .357 mark. Easy peasy.
How does the sweeper hold up against lefties? Bad news:
| Characteristics | SwStr% | Whiff% | GB% | PU% | BABIP | HR/CON | RV/100 |
|---|---|---|---|---|---|---|---|
| Non-Sweeper | 15.6% | 31.8% | 38.4% | 14.6% | .268 | 5.2% | -0.35 |
| Sweeper | 12.9% | 28.5% | 32.8% | 16.1% | .284 | 4.4% | -0.05 |
In essence, that contact advantage goes away. Lefties hit more line drives against the pitch, and don’t pop up at the same elevated rate. In wOBA terms, regular and sweeping sliders both allow a .370 wOBA on contact, but sweepers are at a disadvantage when it comes to missing bats. That makes intuitive sense to me. Pitches that break towards a batter are easier for them to see; that’s why pitchers use changeups to fight opposite-handed hitters.
If your sweeping slider doesn’t have an edge on contact, that’s a problem. That’s the whole reason you throw the pitch; if it doesn’t do better on contact, and it’s breaking towards left-handers’ bats, you should really shelve it. Pitchers more or less know this; they halve their sweeper usage to opposite-handed batters.
Maybe that’s not a huge problem, because pitchers halve their overall slider usage when they don’t have the platoon advantage. I’d argue that they don’t need to, at least not as much – bullet sliders demonstrate smaller platoon splits, although not always to the extreme that they have this year. Even though the term “sweeper” has only entered baseball vernacular recently, people still threw them. I used the same criteria to hunt down platoon splits for the two slider variations in the past four years:
| Year | Regular vL | Regular vR | Platoon Gap | Sweeper vL | Sweeper vR | Platoon Gap |
|---|---|---|---|---|---|---|
| 2019 | -0.26 | -0.54 | 0.28 | -0.39 | -1.59 | 1.2 |
| 2020 | -0.32 | -0.59 | 0.27 | -0.75 | -0.34 | -0.41 |
| 2021 | 0.07 | -0.19 | 0.26 | -0.3 | -1.26 | 0.96 |
| 2022 | -0.35 | -0.26 | -0.09 | -0.05 | -0.94 | 0.89 |
I don’t want to act like this is some unassailable logic, proof that I’ve cracked the mysteries of this pitch variation. There are huge population issues; the pitchers who throw sweepers aren’t the same as the ones who throw regular sliders, at least not in every case. The average pitch quality might be different; the average opposition might be different, too. The sample sizes aren’t even huge; there have already been nearly twice as many sweepers thrown in 2022 as were thrown in 2019, and the season isn’t over.
But even taking that into account, I have some bullet points for you the next time you watch someone throw a big, sideways-breaking slider with a bit of velocity. One: it’s going to get a ton of popups and suppress BABIP, even if it doesn’t miss extra bats. Two: it has a meaningful weakness against opposite-handed pitchers. For what it’s worth, the left-handed pitcher sweeper data looked similar, but the sample was so small that I just discarded it.
What does that mean? First, I think that sinker/slider pitchers should mostly only throw sweeping sliders if they’re relievers. Relievers get the platoon edge far more often, and they also don’t have to see the same batter twice; you can get away with four straight sinkers to a lefty, but probably not eight. If your job is to get righties out, and you can learn to throw a sweeper instead of a tighter slider, that seems like a good tradeoff to me.
Second, if you’re a Yu Darvish-esque pitching savant, capable of throwing anything and everything, the kind of guy who Pitch Com doesn’t work for because it doesn’t have enough buttons, why not add a sweeper? If your arsenal is already so deep against opposite-handed pitchers that you don’t need a slider there, it seems like something worth toying with. That contact suppression looks real; across all four years of data, right/right sweepers have allowed a .248 BABIP.
I don’t think this will continue indefinitely, for the record. It’s still a competitive advantage for some teams, but more teams are teaching it, and there’s a natural cycle to pitches like this: more and more pitchers throw it until too many pitchers throw it, which affects the aggregate quality. Then its popularity wanes, and only the experts are left throwing it. Sinkers went through a similar pattern; I think splitters might be next. Sweepers are neat, and they have a lot working in their favor, but don’t expect them to replace tight, bullet-style sliders overnight.
Ben is a writer at FanGraphs. He can be found on Bluesky @benclemens.
The broom of the system
Good article. It appears to confirm most ideas I instinctively have regarding sweeping sliders. They have larger platoon splits, the increased horizontal movement improves chances for weaker contact, and likely get punished more if the pitcher misses over the plate with them.
So, as I’m getting downvoted I wanted to clarify. I didn’t call it a good article because it agrees with my preconceptions. I thought it was a good article due to Clemens detailed methodology while finding a way to make that methodology engaging to readers. I would have found it to be a good article even if it completely when against my instincts.
I think it’s another response to the league-wide shift toward elevating and looking to hit for power. Just my hypothesis, but I’d imagine that an earlier era of contact-oriented hitters with flatter swing paths would have little trouble with the sweeper. Much like the elevated fastball, it seems like a pitch that only becomes difficult when you are teaching swing mechanics designed to get more loft.
A flatter swing likely has a better BABIP against sweepers. The consequences of a flatter swing is that homers drive scoring with as good as pitchers are currently and pitchers have more pitches than a sweeping slider.
Baseball is a game of constant adjustments and I think what we’ve seen over the last 8 or so years is that it has as much of a ‘metagame’ in terms of popular pitches and swings that respond to them as many online competitive games do, due to the availability in both areas of comprehensive data.
In the early-2010s, the 2 seamer was all the rage with many teams (like the Pirates!). Players who had swings that did well against this pitch started to flourish (whether naturally or by intention). This lead to the 2016-2018 rise in the high 4-seamer which those swings did not handle well. Many of the players who performed well in the previous area suffered rather sudden drop-offs in performance as the ‘metagame’ shifted away from them. At times teams were honest about this: I remember Joe Maddon and the 2017 Cubs commenting after a sweep at Dodger Stadium that they were had not yet figured out how to adjust to the constant diet of high 4 seam fastballs. The 2016 Cubs were very good at the previous meta; not so much at the new one. Many of those players never adapted and were never that good again. As the 4 seamer ascended, we saw new players debut and excel with swings that were well suited to attack that pitch.
In 2020 we saw the pendulum start to shift back the other direction. The Rays and Dodgers both made the World Series with bullpens full of guys who threw 2 seam fastballs again. Having recognized that the league had started to adapt and rosters were now full of guys who were prepared for the high fastball, those teams focused again on attacking the lower part of the zone. This worked. As this movement has proceeded, we’ve seen guys who were excellent during the 4 seamer period fall off – some rapidly – while we’ve seen others rebound after years of struggle who were originally good at this (see: the old guys on the 2021 Giants).
There’s a level of star player who can handle everything. And there’s a tier of player whose swing is adaptable enough that they can respond to the constant shifts. But many players are not. They get to the majors because they’re good against what pitchers are throwing now, and then they struggle or fail to live up to that initial performance when it shifts away from their strength. I do not think it’s remotely coincidental that guys like Brandon Crawford and Brandon Belt who excelled in the previous 2-seamer era experienced significant rebounds as pitching staffs shifted back towards the sinker and away from the high 4 seamer. They simply are better at one pitch than the other, and managed to stay in the league long enough for the cycle to return to what they were best at.
I think the ‘sweeper’ is just another instance of that – this 2 seamer era has featured a lot of horizontally moving 2 seamers (think Dustin May) and not the emphasis on the ‘sinker’ 2 seamer (think Dallas Keuchel) that characterized the 2012-2015 era of 2 seam popularity. Horizontal movement is a path to weak contact, especially against guys whose swings excel upwards in the zone and perhaps do not as much lower in the zone. As teams debut more guys who are good at handling that pitch, we’ll see bullpens adapt again by returning to the high 4 seam riding spin guys.
Starting rotations can’t adapt that quickly because the number of pitchers with the ability to start is too limited. But bullpens are full of fungible arms, and you can see them rapidly adapt to attack in whichever way their organization thinks is the best way to attack. That’s the sweeper and the 2 seamer today, but I’d guess it’s different a year from now, and definitely in 2024.
Evan Phillips has been amazing. It works great now for the Yankees. In 2 years, the teams that aren’t on the cutting edge will be trying to have guys who throw sweepers, and hitters will have started to figure it out, and the Rays, Dodgers and Yankees will have moved on to whatever comes next. It will be an endless game of discover-and-catchup, and the only really interesting question is if and when another organization will manage to join those 3 at the front of the line in innovation, or whether one of those 3 will fall out of that group.
Very cool. So how did you come up with your movement cutoffs? Do the opposite of normal data cleaning (so looking for high z-scores instead of smaller)?
Is the measurement of horizontal movement in “Slider Shapes” supposed to be feet?
Whoops, should be, edited to reflect that.
I feel like splitters are already in the “experts only” bucket already — and have been for a number of years. There’s only 25 guys in the league who have thrown 100+ splitters this season — and only 51 who have had ANY PA end with a splitter. That’s not a lot.
Looking at the data, the sweeper just looks like a much better pitch overall – it outperformed 3/4 years vs righties and 3/4 years vs lefties, often dramatically so. Pitchers are apparently catching on as well, given “there have already been nearly twice as many sweepers thrown in 2022 as were thrown in 2019”.
This was my takeaway from the final graph as well, only in 2022 has the sweeper not outperformed the regular slider when at a platoon disadvantage. It’s just that it typically way outperforms a regular slider when there is a platoon advantage. Not sure I see that as a negative the way the author does.
Now maybe with it being more common in 2022 batters have adjusted, but we shall see if the trend continues.
The sweeper is better *right now* against the players in the league *right now*, who over the past 3-4 years have been heavily selected for their performance against high 4 seam fastballs and pitches moving vertically.
If this remains popular, along with the horizontally running 2 seamer, we will start to see different players who are stronger against those pitchers succeed, and the gap between the sweeper and the ‘regular slider’ may invert when we reach the point where players are more prepared for horizontal movement than vertical movement. Right now, most rosters are adapted to attack the verticality that has been the mainstay of pitching staffs for the past 5 years or so.
Star players are adaptable enough to succeed against anything. But average and bit players are not. The difference between a star and a league average player in most cases is that the first one has shown the capacity to adjust to cover holes in their swing and the second one has not, and teams simply exploit their weakness for their entire careers. Of late, there’s been heavy pressure on that demographic of players to select those whose strengths come against vertical change-of-direction (4 seam/curve/trad slider) and not horizontal (horizontal 2 seamer/sweeper/cutter). If the horizontal movement becomes prevalent enough, the league will adapt to it, as the constant churn of non-star players on MLB rosters will start to prioritize guys who can handle what’s most common.
Most average or even below average MLB players aren’t consistent in that level of performance. They have areas and pitches where they excel and they have weaknesses they’ve never figured out how to address that opponents try to exploit constantly, and that prevent them from being great (the greats more or less uniformly being highly talented people who *successfully adjust*). This is why there are so many ‘huge talent, never quite turned out’ prospects and players: whether they’ll be able to constantly adapt to what happens to them in the major leagues is a question that can’t be answered before they’re in the major leagues. Until then, they simply don’t face the scouting, analysis, and opponents capable of consistently exploiting their individual weaknesses enough to punish them for having them.
When you are thinking about whether a particular pitch or swing or what have you is ‘clearly better’, keep in mind the time horizon is almost always ‘right now and in the near future’, because what is the best for pitchers will constantly change as hitters attempt to adapt and catch up with what *was* best for pitchers. And those who fail to adapt will just be replaced with other, equally flawed players whose particular mixes of strengths and weaknesses are better suited for the status quo of pitching.
Statistics across baseball aren’t driven by the performances of stars. They’re driven by the body of all players, and as a unit those players are highly interchangable and fungible and are constantly being replaced by new, equally (but differently) flawed guys from the minors who their teams think are better suited for the status quo.
Thanks. I think that’s what I was saying with the last sentence albeit a little more succinctly.
No one has ever accused something I have written (especially when it’s 5 paragraphs long) of being ‘more succinct’
The Dodgers have allowed the lowest BABIP in MLB for the last 5 years, so there’s more to it than this, but it’s certainly not hurting.
Good stuff
This data supports the reason most Dodgers with sweepers also have a cutter for lefties, Philips, Treinen, May, Buehler, Almonte (probably in the future if he didn’t get hurt.)
Ever since I first watched Tanner Houck embarrass the Braves in 2020 I have been addicted to the sweeper. Then this year John Schreiber, a organizational filler for most of his career, showed up and continued to make hitters look foolish waving at his sweeper. The next thing to watch is whether hitters will make successful adjustments as more pitchers throw it. It is a different pitch than a slider, actually more like a horizontal curve, and needs to be seen as such.
Such an awesome analysis. My stats profs used to freaking love that “four-fold” table like you used in the two dimensions (speed and break). Then the prof lost me with his multidimensional, multivariate regressions. Great job, bud!
The sweeper’s performance on pops and by handedness seems intuitive. When a batter slightly misreads a sweeper in a same-handed matchup, they’re liable to reach. Reach = pops, especially if they’re also out in front. The bullet slider requires less reach to contact.
For an opposite-handed matchup, the “fooled” hitter ends up getting the bat head out a little early and the ball pretty much breaks right into the barrel. So you get a pull-side laser show. Classic “hit it out front.” Or, if the pitcher tries to use it away, it’s still breaking into the barrel in a way that can be lasered to any field.