The Reasoning Behind Austin Barnes’ Bunt
The Dodgers cruised to a 6–2 victory over the Rays, thanks to the masterful pitching of Walker Buehler and a revving offense that bombarded Charlie Morton, who before the game had a shining postseason ERA of 1.45 across nine starts. It was unexpected, and it was also a testament to how incredible the Dodgers have been all year.
There was another, albeit minor bit of unexpectedness in this game. In the top of the 4th inning, Dodgers’ catcher Austin Barnes executed a move that would have had old-school advocates everywhere screaming in ecstasy:
A bunt! A squeeze bunt! What made this extra-surprising is our awareness of how analytically-orientated the Dodgers and Rays are. Bunting in a playoff game seems antithetical to the sabermetric mindset championed by Andrew Friedman, who embedded it first in the Rays organization, then went on to lead the Dodgers as president of baseball operations.
On Twitter, I noticed a back-and-forth between Devan Fink and Max Greenfield, two young baseball minds. Max wondered if the Dodgers had built a model that determines whether or not a bunt is justified, with Devan replying it was likely.
We don’t have access to proprietary data, but I wondered whether I could deconstruct the elements of the Dodgers’ model — what contexts combined to give Austin Barnes a green light?
From strictly a run expectancy standpoint, all forms of bunting are not a good idea. In Barnes’ situation, successfully executing a squeeze bunt does score a run, but the benefit it provides is marginal at best. In 2019 (I couldn’t find 2020 data), with runners at the corners with no outs, teams scored an average of 1.758 runs. With a runner on second with one out, however, they only mustered 0.696 runs. Do the math, and 1.758 minus 0.696 is actually greater than one!
What’s worse, a failed bunt results in that drop without the consolation of an added run. So if you lose more than you gain, why ever bunt? Well, things could go worse. If Austin Barnes grounded into a double play instead, the Dodgers’ run expectancy would plummet to––wait for it––zero.
This is where the risk calculation comes into play. Given the numerous variables — the pitcher, the batter, the infield alignment, etc.––is bunting a better alternative than a regular at-bat?
Let’s start off with the pitcher, Charlie Morton. The count was 1–0. His pitch was a well-located sinker. Is there any relationship between the count and the pitch he throws? Turns out, there is. In early counts (0–0, 1–0, 0–1,), Morton shows a preference for his fastballs:
This is beneficial for a would-be-bunter. Fastballs are easier to execute a bunt against––if you tried it against Morton’s curveball, which has over 50 inches of vertical drop, you’d get lost just trying to track the ball’s movement. So from the get-go, Barnes had a strategy: squeeze early.
What if Barnes had swung instead? Unfortunately, he has an above-average ground ball rate––of the 77 fastballs he put into play this year, 49% of them resulted in one. If we multiply Morton’s early-count fastball rate with Barnes’ ground ball rate on fastballs to calculate a shoddy-but-reasonable probability, we would get 35.5%. Not all ground balls with a runner on first result in double plays, but that’s significant. There’s some risk there.
But Barnes didn’t have to engage with Morton’s fastballs — he could taken some pitches and waited for the curveball instead. Bad news: Barnes isn’t particularly good against the curveball, either. For the past three years (2018 to 2020), he’s recorded a lackluster .191 wOBA against it. Zoom into this year, and he has a .000 wOBA. Charlie Morton is not an easy pitcher for anyone, but Barnes’ characteristics raise the difficulty a notch.
Plus, it turns out that Austin Barnes is not a bad bunter! He’s had 9 bunts across his career, and only two of them have resulted in the worst outcome: no runners advancing, unnecessary outs added.
In fact, one of his bunts this season against the Padres strongly resembles that against Morton, leading me to speculate that Barnes might have practiced for situations like these. Check out the identical execution:
However, all of this assumes a perfect success rate from Barnes. If he had failed the squeeze bunt, his attempt might have been forgotten, and I might not be writing about it. Success can distort our perceptions. So for a second, let’s imagine that in an alternate dimension, Austin Barnes screwed up his bunt. Inning over. Would the odds have shifted in Tampa Bay’s favor?
Not really. By the time Joc Pederson sent Cody Bellinger to third with a single, the Rays’ win expectancy, according to FanGraphs, sat at a lowly 14.7%. It’s highly unlikely that an inning-ending double play would raise that percentage by a significant amount. Plus, in the 5th inning, the Dodgers could simply attack with the heart of their order––Mookie Betts, Corey Seager, Justin Turner, and others who have a greater chance of stringing together a run than Austin Barnes alone. The Dodgers may not gain much from a success, but they also don’t lose much from a failure.
I considered other factors, such as if batter and pitcher handedness affected the direction, and thus success of a bunted ball, but there doesn’t seem to be a meaningful relationship. The Rays didn’t shift against Barnes, so that’s a non-factor. Ultimately, I think what activated the Dodgers’ green light was a combination of the matchup and a perceived lack of risk. Charlie Morton versus Austin Barnes has a good chance of resulting in a grounder or a strikeout. Barnes has successfully attempted multiple squeeze bunts in a career. And if he fails, well, then onto our NLCS MVP.
Like Devan said, there probably is a model that combines all the factors I’ve mentioned and more to generate one, neat number. I don’t work for the Los Angeles Dodgers, so all I can do is pick up pieces of the puzzle and combine them into a decent image. But that’s good enough for me.
All statistics from Baseball Savant and Fangraphs
