Should baseball hitters with two strikes change their swing?

Baseball’s newest tracking tools have given one old dugout argument a faster swing.

When a hitter comes up with two strikes, fans, coaches and former players often call for the same adjustment: cut down the swing, make contact, keep the at-bat alive. It sounds like common sense. The harder question is whether it actually helps a batter produce better results, or simply trades one kind of failure for another.

A new study in The American Statistician finds that the old advice contains some truth, but not the whole truth. Hitters who shorten and slow their swings with more strikes do make contact more often. They also give up power, and in the data that tradeoff mostly cancels itself out.

“What we found is that there’s really a tradeoff,” said Scott Powers, assistant professor of sport analytics and statistics at Rice University and a co-author of the paper with Ron Yurko of Carnegie Mellon University. “When players slow down their swings, they do make more contact, but they also sacrifice power.”

That balance led the authors to a restrained conclusion. “The takeaway from this is that it’s not clear batters get better results by slowing down their swing speed as the number of strikes increases,” Powers said.

For Powers, the question has followed him from front offices to academia. Before joining Rice, he spent five seasons with the Los Angeles Dodgers, including three as director of analytics, then served as assistant general manager for the Houston Astros. He won a World Series ring with the Dodgers in 2020 and another with the Astros in 2022.

During those years, he said, the two-strike debate never really disappeared.

“When I was working for the Dodgers, every time one of our batters would strike out in a big situation, I would get texts from my friends complaining about the approach of the batter,” Powers said. “But I never understood. Is the batter actually trying to swing for the fences?”

Public bat-tracking data finally offered a way to test that. In 2024, Major League Baseball released pitch-by-pitch bat data that included swing length and bat speed. Powers and Yurko built their analysis from 685,143 pitches from the 2024 regular season, using publicly available Baseball Savant data downloaded with the R package sabRmetrics.

The new information matters because it captures swings one pitch at a time, not just as player averages. Bat speed is measured as the linear speed of the bat’s sweet spot near contact. Swing length measures the distance the end of the bat travels from the start of the swing to contact, or the point nearest contact on a miss.

That makes it possible to ask how a hitter changes from count to count. It also creates a trap.

At first glance, the raw data seemed to suggest something that did not make much baseball sense. Faster, longer swings appeared to produce better contact.

That early pattern turned out to be a lesson in measurement.

“When we first glanced at the data from a high level, we observed that faster, longer swings appeared to have higher quality contact,” Yurko said. “But this ignored the context of how MLB records the swing length and bat speed measurements, along with various confounding factors such as a batter’s ability to recognize pitches and timing.”

The problem is that swing length and bat speed are recorded at contact, and contact itself depends on timing. An early swing and a late swing can reflect the same mechanics but produce different measured values because the bat is captured at a different point in its path. In that sense, the numbers are not just inputs to an outcome. They are partly outcomes themselves.

Powers said the key was recognizing that the pitch, and a batter’s reaction to it, shape the reading.

At the center of the study is the idea that hitters are not simply choosing a swing speed in a vacuum. They are reacting in real time, often after guessing right or wrong about pitch type and location. A batter ready for a fastball may take a freer swing if the pitch matches the expectation. A batter forced to adjust mid-swing is more likely to lose bat speed and less likely to square the ball up.

To get closer to intent, the authors focused first on a narrower set of successful swings: squared-up contact against pitchers’ primary fastballs. They then used Bayesian hierarchical skew-normal models to estimate intended bat speed and swing length, rather than treating the observed measurements as direct expressions of a batter’s plan.

Their skew-normal models outperformed simpler Gaussian versions in out-of-sample testing. The difference was modest in scale but consistent, with the skew-normal approach showing better expected log predictive density for both swing length and bat speed.

Once the researchers modeled intent, the two-strike picture looked more familiar.

As strikes increased, average intended swings became both slower and shorter. The fixed-effects estimates showed expected bat speed dropping by about 1.13 mph per added strike and swing length shrinking by about 0.14 inches per strike. More balls in the count produced the opposite pattern, with longer and faster swings.

The next question was what those adjustments actually bought a hitter.

Using instrumental variables regression, the authors estimated how count-based changes in swing style affected three outcomes: contact, whether contact became a fair ball, and the expected run value of fair balls. Batters who reduced bat speed more with additional strikes made more contact, but they also lost power when they did put the ball in play. Shortening swing length helped contact too, though less dramatically, and did not appear to reduce power in the same way.

That made one result stand out.

“The best type of two-strike approach is to shorten your swing without slowing it down,” Powers said. “Julio Rodríguez stood out as somebody who does that particularly well.”

When the team translated the contact-power tradeoff into run value, the gains and losses stayed small. Across approaches, the spread from the best to the worst was about four runs per 500 plate appearances, roughly half a win over a season. The most valuable approaches were those that cut swing length while preserving bat speed.

Table-based rankings in the paper put Matt Chapman, Matt Olson and Mark Canha among the best examples of those higher-value approaches, while several others graded worse because they gave up too much bat speed relative to the contact they gained.

Yurko said one of the most striking parts of the study was how neatly the evidence lined up with what baseball people had long suspected.

“Batters can reduce their strikeout rate by changing their swing length based on the count, such as choking up on the bat with two strikes,” Yurko said. “As statisticians working in sports, we usually do not see such clean results.”

The study does not settle the two-strike argument so much as refine it. Telling hitters to simply “swing easier” misses the central tension. A more conservative swing can help them make contact, but lost bat speed can take away much of the reward.

That matters for coaches, analysts and fans trying to interpret the flood of new swing data. The paper argues that bat-tracking metrics should not be taken at face value, especially when they are measured at contact and shaped by timing and pitch recognition. It also suggests that the best two-strike adjustment may be a precise one: trim the swing path, not the force behind it.

For future analysts, Powers framed the lesson more broadly. “The thing I hope students doing baseball analytics research in the future will take away from this study is that you can’t just take swing-tracking metrics at face value,” he said. “You have to think carefully about how these things are being measured and what they mean.”

Research findings are available online in the journal The American Statistician.

The original story "Should baseball hitters with two strikes change their swing?" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.