What's in a swing? A metrics explainer

We don't have hitting figured out, but Freddie Freeman might

If you were to ask a bunch of different professional scouts and/or analysts how they go about evaluating big league hitters, you’d get a bunch of different answers.

My preferred evaluation structure includes two phases:

• 1) Decision making: I find it helpful to try to understand a hitter’s approach before diving into his swing. Here, I’d be focusing on some general questions:

  • How aggressive is this hitter?

  • How strong are his pitch recognition skills?

  • Where does he like the ball?

  • How much does his approach vary situationally?

• 2) The swing: When a hitter decides to “go” on a pitch, what does his swing look like? There are a ton of variables to consider at this stage (a hitter’s body movements, the path of the bat, the speed/acceleration/angles created by each of these parts). Each variable is likely not one static measurement, either; a hitter’s decision making process informs the distribution of his recorded bat speeds. A 2-strike swing might look different than an 0-0 swing, for example.

I introduce this structure because I think it can be a helpful starting point to understand some of the new hitting metrics MLB released last weekend. We’re about to see an explosion in public research on hitting. And because hitting evaluation is a lot more complicated than pitching evaluation, the potential for distraction is high.

But before diving in, a quick recap on what’s been released.

What did MLB roll out?

A suite of new metrics that are rolled up from the Hawk-Eye camera tracking systems installed in MLB parks in 2020. This weekend’s release focuses on (a) partial/check swings, (b) bat speed metrics and (c) one metric on the path of the bat.

Are these metrics new to front offices?

No. Front offices have had access to the raw data behind these metrics (and additional ones yet to be released by MLB publicly) for several years. But because they’re all relatively new (and dense to wade through), some teams probably are holding competitive advantages on how they’re applying them in 2024.

Ok, background covered. Let’s jump in to process these new metrics and what they tell us. I’ll start with decision making and then cover the swing.

Check swings tell us something about decision making

On the surface, it may seem like data on the path/speed of the bat relates more to the second part of the hitter evaluation (the swing). But this new release does serve as a reminder that we should consider a bit more nuance when we evaluate a hitter’s swing decisions.

Here’s why. Until now, we’ve treated hitter swing decisions as yes/no questions. Did a hitter swing at that nasty breaking ball, or did he stop himself? Simple.

Is it, though?

This is a swing, officially - because the home plate umpire said so. But it was a close call.

To the dismay of broadcasters who already had their own names for half/check swings like these, MLB has termed these as swords. The name is meant to carry a shameful connotation for the hitter.¹

Swords shouldn’t be looked down on, though. I’d argue to the contrary. While the sword might sometimes look awkward, it signals restraint. Where some hitters would accelerate through a full swing, others (the sword-ers) are changing their minds mid-swing. The prior yes/no framework to conceptualizing swing decisions should actually be thought of as more of a spectrum. Something like:

• No swing → Stride, but no swing → Check swing/sword → Full swing

When we score hitters based on their decisions made to swing or not swing at a pitch, shouldn’t we penalize the sword less than the full swing at a chase pitch?

If you don’t buy my point, take a look at the leaderboard:

If I’m a big league hitter, I wouldn’t mind being on a list with these names. All but two names at the top of this list are currently posting above-average offensive lines². You’ll also note that the hitters with the most swords tend to swing hard. Swords are really soft swings, though - so they’re actually bringing down the “average bat speed” number for these hitters.

You know who has zero swords in 2024? Giancarlo Stanton. Swings like this one aren’t swords because Stanton’s pitch recognition skills are not in a great place right now:

Because Stanton’s hard swing here is really more of a decision making issue, that “all swings” bat speed column in the table above is a bit misleading - after all, do we really care how hard a hitter swung and missed at a pitch he was fooled on? For hitters at the top of this leaderboard, we see that their bat speed when making contact in hitters’ counts (when a hitter is likely to “let it eat”) is a degree higher (the last column in the table above).

To summarize: to “sword” means to cut off a swing that probably wasn’t a good idea in the first place. The sword is a stat that doesn’t tell us much, but it should prompt us us to think about (and model) hitter decision making in a way that is a little more nuanced than a “did/did not swing” framework allows for. We don’t have enough data yet to do that, but I would expect that MLB’s next phases of metric releases will allow us to get closer.

What can we learn about the swing?

MLB released two other types of metrics last weekend: a batch related to bat speed, and one metric on the length of the bat path.

Let’s start with bat speed. As we covered briefly above, a hitter’s bat speed distribution is probably more interesting than the number itself. Let’s consider two extreme examples:

Source: Baseball Savant

The blue distribution is Stanton, and it’s very tight. He always swings hard. Some might say he has no “B swing,” or that he never “gets cheated” on a swing.

I instead would suggest that Stanton’s bat speed distribution highlights the connectedness between the swing and decision making. Stanton is not making mid-swing corrections to make contact with a pitch that he’s mis-judged. If he were, we’d see some softer swings in his distribution. He also isn’t making late decisions to put a modified, softer swing on a pitch that he originally didn’t want to swing at. I’d call his approach something like “low swing adjustability.”

And then there’s Freddie Freeman. He’s able to meld strong decision making with an adaptable swing in ways that no other hitter can. You can see this ability in his wide bat speed distribution.

An example: here’s Freddie tracking a Yu Darvish sweeper for what seems like a few minutes. He sees this breaking ball come out of the hand, get its hump, and start breaking toward the ground. Freeman wants to know: is this pitch going to bite down below the zone, or will it remain a strike? Most hitters would be making a decision in that second frame. Freeman waits until the third frame to make his call:

The video allows you to fully appreciate how long Freddie waits to swing:

This was a 69 mph swing. It’s a softer swing than any swing Stanton has taken this season. And the result, because of Freeman’s clean contact between the barrel of the bat and the pitch, is a double that left the bat at 94 mph.

I’m cherrypicking, sure. Generally, harder swings result in better contact and better results for hitters. But that doesn’t mean a higher bat speed always signals a more effective hitter³.

On bat path and length

And now, the metric that seems most intriguing on first glance: something called swing length. This one measures the distance the bat head travels (in all three dimensions) from the start of a swing until contact. This is the only path-related metric MLB has made available.

Unfortunately, I think swing length isn’t usable without a lot more contextualization. The examples above hopefully suggest that hitter intentions and decision making processes vary greatly across the game. And even if you could hold hitter intent constant, here are some of the other factors that screw with the interpretability of the swing length metric:

The location of the pitch. Swings on pitches that are closer to a hitter (pitches up/in) are naturally going to be shorter than ones that are further away (pitches down/away).

To this point, a hitter’s position in the box also matters. Justin Turner has the 4th “shortest” swing length per the metric; he also crowds the dish.

How “deep” contact is made. Others have pointed out over the past few days that pitches that are pulled tend to come with longer swings. That’s because the bat travels further when a hitter pulls the ball - those swings typically make contact toward the front of the plate, whereas opposite field swings typically make contact a little deeper. Decision making factors in here, too. If a hitter misreads a breaking ball, he’s more likely to leak out in front of the plate to hit it.

How long a hitter’s arms are. Wait, what? Because the arms are the levers that move the bat through the hitting zone, they matter, too. I don’t think it would be possible for Aaron Judge to have a shorter swing (per this metric) than Steven Kwan.

Having shorter arms is a prerequisite for a short swing length (again, the way this metric is defined). The five hitters with the shortest swing lengths are 6’0” or under. On the other hand, two of the five longest swing lengths (Judge and Stanton) are 6’6” and 6’7”, respectively.

How flat a swing’s attack angle is. The shortest path to a pitch is the flattest one. But one way to create loft is through a low-to-high bat path. Kyle Schwarber is a classic example of a steep attack angle swing:

Schwarber doesn’t have long arms, and yet the route he takes to this pitch is longer than the path taken by 98% of other left-handed hitters who hit a fastball in this area of the zone⁴. This had a swing length of 8 feet.

I can’t resist another Freeman comparison. Here’s a similar pitch (fastball, same location):

Swing length: 6.4 feet (shorter than 62% of other left-handed swings in that spot). It’s not hit quite as hard as Schwarber’s (99 mph exit velocity to Schwarber’s 110 mph), but it still travels 379 feet. Freeman is more direct to the ball; his loft doesn’t come from a low-to-high path as much as Schwarber’s does. Where does Freeman’s loft come from? That’s a question for another post - one that covers bat angles.

What we’re actually looking for in the swing length metric

Further complicating the swing length conversation is that ranking high or low on this metric, even if it were appropriately contextualized, couldn’t really be called a “bad” or “good” thing. The example illustrates that longer swings are generally harder swings. But they are also associated with more swing-and-miss (Freeman: 19.4% in 2024; Schwarber: 31.2%⁵).

The ideal metric to describe a short or a long swing should probably focus on the efficiency of the bat path; one that can identify how short a hitter can be to get the barrel to the pitch, but also how fast the barrel is moving once it gets there. The ideal bat path metric should also consider the angle of the bat head (steeper resulting in more loft). I’m not going to put all of those three factors together in a metric today - I couldn’t if I wanted to. But I will leave you with an image of what I think it looks like:

By the new metrics, this swing doesn’t register as exceptional. It wasn’t the hardest (77.5 mph). It registered an above-average swing length (7.4 feet).

But damn, it was efficient. And even though the metric doesn’t say so, this swing was short - we just have a long way to go in defining what short means.

MLB deserves a huge thank you for putting these metrics out for public consumption. Even if this batch can’t tell us much about what marks a good hitter (or swing), it can at least allow us to better appreciate how hard hitting is. Our ability to evaluate will come in time.

Have a great week!

1. MLB’s explainer on swords (including why that name exists). I’m not a huge fan of a metric designed to embarrass.

2. per FanGraphs’ wRC+

3. If you need more proof, check out the weak connections between bat speed and offensive production identified by Ben Clemens yesterday at FanGraphs.

4. Schwarber’s swing length on this pitch was 8th highest among 454 balls put in play by left-handed hitters on pitches up and over the the middle of the plate this year (per Baseball Savant).

5. per Pitch Info