Parsing a Pile of Confusing Data About Sliders by Ben Clemens May 18, 2021 What’s the most important characteristic of a slider? Let me show you a table: Slider Value by Location, 2020-21 Attack Zone Run Value RV/100 Pitches Heart -138.9 -0.7 Shadow -785.0 -2.5 Chase 163.8 0.8 Waste 465.2 4.5 There are two things that might need explaining in here. The attack zones are Baseball Savant’s way of cutting the strike zone up into granular pieces, and I think they’re neat. They look like so: Run values are from the batter’s perspective, so that -785 runs in the shadow zone means that batters have been 785 runs below average — what they’ve done on all pitches across the whole year — when they faced sliders in the shadow zone. In other words, sliders on the corners of the plate have been excellent — not really a shock. What’s the most important characteristic of a slider, then? Well, allow me to show you a different table: Slider Value by Speed, 2020-21 Velocity (mph) Run Value RV/100 87+ -149.3 -0.8 85-87 -80.6 -0.4 82-85 -43.8 -0.2 <82 37.2 0.2 This is hardly surprising; you don’t need me to tell you that it’s harder to hit faster sliders. The surprise was the magnitude, but maybe it shouldn’t be. Jacob deGrom’s hard slider is obviously excellent, but so is Joe Musgrove’s mid-speed offering. Steven Matz throws one of the hardest sliders in the game, particularly for a starter, and it’s always been a middling pitch. Velo matters, but not as much as you’d expect. Let’s continue the gimmick: what’s the most important characteristic of a slider? Here, take a look at this table: Slider Value by Movement, 2020-21 Horiz. Break Run Value RV/100 7.8″+ -107.8 -0.6 4.9-7.7″ -30.4 -0.2 2.5-4.8″ -67.9 -0.4 <2.5″ -30.4 -0.2 Finally, we seem to have found something that matters less. Just one problem: it seems like it should be the thing that matters most. Big-bending sliders look so cool on TV, and batters look foolish against them. The group with the least break are mostly backup sliders — with minimal glove-side movement or even arm-side movement. These are the sliders that hang up and make announcers gasp with disappointment. Those mistakes perform just fine! Clearly, my lazy one-factor description of sliders isn’t going to work. Let’s go to the next level down and combine two of these indicators to see if we can learn anything more. Here’s a grid of velocity crossed with location: Slider Run Value Per 100 Pitches, 2020-21 Velo/Location Heart Shadow Chase Waste 87+ -0.94 -3.41 0.19 4.24 85-86 -0.90 -2.76 0.86 4.65 82-84 -0.49 -2.24 0.91 4.22 <82 -0.54 -1.85 1.55 5.18 Throwing hard is obviously great, but it’s not a miracle cure. The shape of each velocity bucket is the same: over the heart is okay on average, the edges of the plate are wonderful, and nothing else is particularly good. Faster is better more or less across the board, but variation by speed is much smaller than variation by zone. Hit the shadow zone and win a prize, regardless of your slider’s zip. What about horizontal movement crossed with location? This one is quite weird, but it’s telling: Slider Run Value Per 100 Pitches, 2020-21 HMov/Location Heart Shadow Chase Waste 7.8″+ -2.51 -2.53 1.08 4.63 4.9-7.7″ 0.21 -2.32 0.62 4.43 2.5-4.8″ -0.63 -2.60 0.58 4.37 <2.5″ -0.01 -2.65 1.04 4.56 If you’re throwing over the middle of the plate, it pays to have movement. To some extent, that’s obvious — batters swing when you attack the middle of the strike zone, and it’s harder to hit pitches with big break. I’m less sure about the second group of pitches — with above-average but not extreme horizontal movement. In a statistical sense, that poor result is significant; there are more than 4,000 pitches in each bucket, so a difference of 0.84 runs per 100 pitches (relative to the third bucket) is large enough that it’s extremely unlikely to occur due to chance. But I wonder about other factors involved in both, enough to say that the only clear takeaway is that huge movement is great in-zone. Leave the zone, and the opposite appears to be true. Big-breaking sliders have the worst results in the chase and waste zones, and they’re roughly average in the shadow zone. That agrees with a recent finding by Tom Tango that pitches with lots of break perform better in the zone. His finding was different — he was looking for pitches that had a lot of break for that particular pitcher rather than overall — but the general findings point in the same direction. Lastly, here’s one more combination, velocity crossed with movement: Slider Run Value Per 100 Pitches, 2020-21 Velo/HMov <2.5″ 2.5-4.8″ 4.9-7.7″ 7.8″+ 87+ -0.16 -1.07 -1.03 -2.50 85-86 -0.40 -0.36 -0.10 -1.45 82-84 -0.05 -0.24 0.01 -0.35 <82 0.29 0.71 0.99 -0.30 I don’t know, this one doesn’t speak to me. Slow sliders are not good at all unless they break a ton; that checks out. Fast sliders are better if they break a ton; yep, I buy it. Mid-velocity sliders are… Basically, they’re a whole mess. Again, sure, I buy it, but this finding is less interesting to me than the zone-based ones. There’s even an increased complexity in the fact that some of the outlier observations are subject to severe survivorship bias. If you throw a 75 mph slider that doesn’t break, you’re either doing something else great or you aren’t a major leaguer. It’s a complex problem to analyze, is my point. What do all these grids mean? Partially, they mean I love to make grids. It’s a very satisfying way to present a big pile of information all at once, and boy is there a lot of data to consider in this particular instance. That aside, I think there are several major takeaways from all of this. First, it’s clear from every slice that getting your breaking pitches near the strike zone pays off. Some of this is obvious — the edges of the plate are always a great place to live as a pitcher — but I was surprised by just how true that was, even for sliders. There’s a sharp count effect here, though, one that isn’t immediately evident in the data. With two strikes, the pattern looks completely different: Run Value/100 with Two Strikes, 2020-21 Attack Zone RV/100 Heart 0.94 Shadow -3.62 Chase -1.54 Waste 4.36 That’s not a surprising finding. It’s something baseball players and analysts have known implicitly for decades. Batters defend the plate more with two strikes, so swing rates go up. Given that plenty of the value in throwing over the heart of the plate lies in called strikes, it becomes relatively less attractive; in fact, leaving one over the heart as compared to throwing it in the “chase” zone goes from 1.5 runs better per 100 pitches to 2.5 runs worse, a huge swing. So, heart early in counts and chase late — shadow all the time, of course, but everyone was already trying to do that. That’s obvious, but my favorite takeaway has to do with the intersection of movement and location. Big, sweeping sliders excel over the heart of the plate. Garden-variety sliders aren’t so good there; leave it over the heart of the plate, and you might get tattooed, whereas it’s hard to hit a pitch that’s breaking eight or nine inches sideways on its path home. Which pitchers are accounting for this by throwing pitches with more movement in the strike zone? There are no pitchers in the game who have thrown at least 10 sliders in both the heart/shadow and chase/waste buckets, get 7.8 inches or more of horizontal break over the plate, and get less than 7.8 inches of break when they’re off the plate. Dallas Keuchel, however, comes close; his sliders that are around the strike zone have a horizontal break of 10.3 inches on average, while his pitches in the chase and waste zones break only 8.5 inches. Perhaps more importantly, pitchers with big sliders should flood the zone when they can. Forget modulating your slider based on location — if you have the ability to throw a big slider, focusing on the zone should be priority number one. Collin McHugh, Yu Darvish, Shohei Ohtani, and Caleb Thielbar all excel at getting their big-movement sliders into and around the strike zone. They’ve all had success with the pitch this year. I’m certainly cherry-picking results, but the theory makes a lot of sense to me. There’s plenty more to unpack here. The data suggests that pitchers with middling sliders might want to attack the strike zone with it but only early in counts, relying on something else for a putaway pitch. On the other hand, nothing appears more important than hitting the corners of the plate. Perhaps pitchers with middling-movement sliders should focus less on changing the shape of the pitch and more on repeating their delivery to hit the corners. John Means doesn’t have a fearsome slider; with an average horizontal break of 3.5 inches, it’s in the bottom third of all sliders. It’s also been a tremendously valuable pitch this year after being so-so before. Why? Well, he’s hitting the shadow zone with 55.4% of his sliders so far this year, and that’s the best thing you can do to add value. Before this year, he’d only located 39.6% of his sliders at the periphery of the zone, and the pitch was worse. More corners, better results: it’s easy math. I’m sure I’ll come up with more takeaways from this mountain of data, but I’d love to hear what you have to say, too. Are pitchers unconsciously taking advantage of this already? Are teams teaching it? I have no idea! But it’s a strong effect, and I’m anxious to hear more theories about why it exists and what pitchers (and hitters!) can do about it.