Test Driving Statcast’s Newest Bat Tracking Metrics

On Monday night, just after midnight, Major League Baseball released a bevy of new bat tracking data. It was accompanied by the now customary combination of an explainer from Mike Petriello and a breakdown of the most extreme players from David Adler. Like many people, I’m still trying to wrap my arms around how these data work and what we might be able to learn from them. Bat tracking metrics are complicated because swings are complicated. The various numbers are interconnected, dependent on location, pitch type, and the batter’s tendencies and intent. There’s no one perfect way to swing, and it’s easier to draw inferences about individual players than overarching conclusions. My first takeaway was that something weird is going on with Leody Taveras. I’ll write about that tomorrow, but for now I’d like to take the new metrics for a test drive. We’ll look at two specific pitch archetypes to get a sense of what these numbers do and how they look in action.

Let’s start as simple as we can. I pulled the league-average numbers for swings against four-seam fastballs right down the middle in zone 5, but I split them up. The top row shows the numbers only for competitive swings on hard-hit balls. The bottom row shows the numbers only for competitive swings that resulted in whiffs. Let’s see how these two swings might differ.

Plenty of differences jump out at you here. When a hitter misses a four-seamer right down the middle, it’s usually because they’re behind on it or under it. All of the new metrics are telling us that in their own way. I’ll capitalize all the metrics in this article, just so we get comfortable with their names and definitions.

Take a look at Swing Length and then at Intercept Y all the way over in the column on the far right. Both numbers are significantly smaller during whiffs. Intercept Y tells us where the bat head is positioned relative to the batter’s center of mass. A bigger number means it’s closer to the pitcher and a smaller number means it’s closer to the catcher. As with all bat tracking metrics, the number used is from the video frame when the bat is closest to the ball. So on whiffs, the ball is closest when the bat head is 21.2 inches ahead of the batter. That’s 7.3 inches farther back than when the batter hits the ball hard, and it results in a swing that’s 0.3 feet (or 3.6 inches) shorter.

Attack Direction tells the same story. That’s the horizontal angle of the bat, where zero would be perfectly parallel to the front of home plate. On the hard-hit balls, the Attack Direction is two degrees to the opposite-field side, which is pretty close to zero. That makes sense because these are fastballs right down the middle; if you’re hitting the ball where it’s pitched, your Attack Direction would be completely flat, but on the whiffs, it’s 10 degrees farther back. Once again, the numbers indicate that the bat hasn’t gotten around yet.

Swing Path Tilt and Attack Angle are a little more complicated. Swing Path Tilt describes the angle that the head of the bat takes over the 40 milliseconds (or 0.04 seconds) prior to contact. Petriello referred to it as the shape of the swing. Every swing starts with the bat head up high. It drops down into the hitting zone, then finishes back up high again. Swing Path Tilt measures how steep the angle is as the bat approaches the ball. Attack Angle is measuring the actual vertical angle of the bat itself – how much lower the head of the bat is than the knob – at impact (or when the bat is closest to the ball, in case of a whiff). So on whiffs, batters are taking a steeper path to the ball, but when the ball arrives, their bat is actually flatter.

Once again, I think this has a lot to do with the fact that for most of these misses, the batter is late on the pitch or swinging under a rising fastball. Steeper swings tend to be longer, which would naturally result in the batter being behind. More importantly, because the ball is catching the bat deeper in the zone, and therefore earlier in the swing, the 40-millisecond window that Statcast is using will be heavier on downward movement and lighter on upward movement. For measuring purposes, this makes the bat path seem steeper than it really is. For practical purposes, it means that the bat actually is lower at the moment when the ball arrives. As a result, hitters try to avoid whiffing underneath the ball by raising the bat head – it’s easier to adjust your arms and wrists than your entire torso – which in turn raises the bat angle. This is just my inference, but it seems intuitive enough. That’s how you get a steeper swing but a flatter bat. Attack Angle is so flat because Swing Path Tilt is so steep.

It’s important to keep in mind that these are just averages. In fact, I went searching for swings that looked exactly like the averages that we saw in that table, and I couldn’t find a single one. So far, in roughly a season and a half of bat tracking, no one has whiffed on a four-seamer over the middle with a Bat Speed of 73 mph, a Swing Length of 6.9 feet, an Attack Angle of three degrees, an Attack Direction of 12 degrees toward the opposite field, a Swing Path Tilt of 35 degrees, an Intercept X of 37 inches away from their body horizontally, and an Intercept Y of 21 inches ahead of it. It hasn’t happened once! These are the two closest swings I could find. They certainly look like your classic whiffs on fastballs down the middle, with the batter swinging just plain behind and beneath the pitch.

There’s one metric I haven’t addressed, though. I haven’t discussed the fact that on our two average swings, the Bat Speed is nearly identical. As we know, Bat Speed increases throughout the swing, so because these whiffs are getting measured when the bat head has yet to really go out and meet the ball, we can safely assume that these swings are in fact even harder than the ones that resulted in hard-hit balls. I don’t know why this would be. It’s certainly not what we would expect. One of the things I’ve learned from breaking down previous bat tracking data is that adjustments are the enemy of Bat Speed. Unless you’re an exception like Shohei Ohtani, the more you have to alter your swing once it’s started, reacting to a location, pitch type, or velocity that you guessed wrong about initially, the more Bat Speed you’re going to leak because you’re shifting the swing mechanics that feed into your Bat Speed. As for why the Bat Speed is equal in our two league-average samples, I only have theories. We’re talking about samples of a few thousand swings each, so I don’t think we can write it off as randomness. For now, I’m just going to leave it as an area for further study and move on to our second pitch archetype.

So that was for four-seamers right down the middle. Let’s take a look at the opposite case. We’re going to look at curveballs down. If you want to get technical, we’re looking at the intersection of gameday zone 8 and attack zone 18, pitches over the middle of the plate horizontally, but at the very bottom of the strike zone. I picked that spot because I wanted pitches that were hittable, but not located abysmally for the pitcher. They’re not hanging curves, but they’re not buried in the dirt like the pitcher might hope. Over the past three seasons, batters have a wOBA of .294 against curveballs in that spot, close enough to the overall league wOBA of .314 for our purposes. Here’s the same table, but for our low curveballs.

This will be much quicker now that we’ve already learned the basics about all of these terms. As you’d expect, just about everything’s going in the opposite direction. The batters who whiff are no longer way behind and under the pitch. The Bat Speed is slower on the whiffs, because the batters are getting fooled and trying desperately to slow down their swings. The Attack Direction, Swing Length, and Intercept Y show that they’re way out in front of the ball. Both because they’re out in front and because they got fooled, their Swing Path Tilt isn’t as steep, so they’re compensating by dropping the bat head, which causes a super steep Attack Angle.

We can also combine our two examples of hard-hit balls to see how they differ.

The bat speeds are the same, but when batters square up low curveballs, they’re meeting them 6 1/2 inches farther in front of home plate, so their swings are longer and their bats are angled seven inches more toward the pull side. The two Attack Angles are almost identical, but in order to reach that lower pitch, the Swing Path Tilt is five degrees steeper.

I’m going to leave you with one video clip that demonstrates the new metrics we’ve been talking about in a fun way. This clip shows home runs from Corbin Carroll and Mookie Betts. They’re both on fastballs down the middle in nearly identical locations. Their Attack Angles, the vertical angles of the bats when they make contact with the balls, are also nearly identical. However, their Swing Path Tilts are very different. Carroll’s is 28 degrees. It’s subtle, but you can see him drop his bat head straight down in order to get on plane with the pitch and take a flatter angle. Meanwhile, Betts has a steeper 38-degree Swing Path Tilt. It’s a quicker swing. He’s taking the bat head straight down to the ball. Despite those different paths, they end up at the same place.

All of these new metrics can seem very descriptive rather than prescriptive, but there’s no doubt that players are thinking about this stuff. They might not be putting numbers to it, but they’re thinking about getting on plane with the ball, about finding a comfortable bat path, about letting the ball get deep or attacking it out front. We’ll learn more about what these numbers mean, but for now, it’s fun to see the way that the numbers line up both with how we think about hitting and about what our eyes can see when we break down a swing.