When To Steal Home and Safety Squeeze

We’ve spent a lot of time this season discussing (and lamenting) the Twins’ offensive approach and its lack of balance. Through games played on Saturday, they continued to lead the league in three true outcome percentage with 39.7% of their plate appearances ending in a strikeout, walk, or homerun. They have 74 more three true outcomes than Seattle, the next closest team, but have only played one more game than the Mariners. Similarly, they rank 18th in ideal plate appearance percentage.

One of the complaints often levied against the three true outcomes style by which baseball is often played today (it’s not just Minnesota) is that it has eliminated the use of small ball strategies. Hit-and-runs, stolen bases, and sacrifice bunt attempts have become less frequent as the game has become more data-driven and teams play ever more by the numbers.

A counterpoint to that is that how baseball is played is far from static. It evolves over time based on the players involved and, perhaps most especially, the rules. When MLB changes the rules, such as lowering the mound in the late 1960s or limiting pickoff throws and increasing the size of the bases before this season, there is a response in style and strategy.

For example, stolen base attempts this year are up roughly 33% over just last season. At the league level, with about six weeks remaining in the regular season, there have already been more steals attempted (about 3,200) than there were in all of the 2021 season (about 2,900).

Not only are clubs running much more, but they are also doing so successfully at the highest rate in baseball history:

Data from FanGraphs

MLB incentivized stolen bases by making it easier to steal successfully and teams responded. Baseball is not a perfectly rational and efficient environment (far from it), but over time, the strategies and tactics employed by teams tend to adjust in response to risk and reward realities.

For all of baseball history until 2023, a rough rule of thumb was that a stolen base attempt was only worth the risk if the chance of stealing successfully was about 75%. That was especially true of attempting to steal second base. With the league average success rate currently pushing 80% this season, you can see why teams have been running more (and maybe should attempt to run even more still!)

Smarter and More Aggressive

The Twins have not leaned into the new baserunning environment with the same kind of volume as other teams. As a team, through Saturday, they had 67 stolen bases this season, which ranked 24th in the league. Utility player Willi Casto is the source of 29 of those, more than double Michael A. Taylor’s second-place 12.

While that team stolen base total might not stand out in the context of the league, it is already significantly more than last season’s 38 stolen bases and the 54 thefts from 2021. As a fraction of opportunities, something tracked by Baseball Reference, the Twins’ stolen base aggressiveness has roughly doubled over last season.

We heard reports in Spring Training when Paul Molitor was in camp that being smarter and more aggressive on the basepaths was a priority this year. This data seems to suggest that’s played out, at least on the aggressiveness side, albeit only modestly.

How about on the smarter side? Like everyone else, the Twins have been stealing bases more effectively this season. Not only did the 2022 Twins not run very much, but when they did, they were bad at doing so. They were caught stealing 17 times in their 55 attempts, a 69.1% success rate that ranked 26th.

This season, though, that script has flipped. They’ve been caught just 11 times in 78 attempts, which works out to an 85.9% success rate and a 2nd-best in the league ranking. They still don’t run very much, but when they have decided to go, they’ve had a pretty good chance of making it safely.

Like so many other things in baseball strategy, the situational context in which these decisions get made makes all the difference in the risk-reward analysis. That 75% chance of success number I mentioned above is a generality that can vary pretty significantly based on the score and the base-out situation. For example, in a situation where one run is disproportionately valuable (say, like it would win the game in the bottom of the 9th), the math tilts in favor of taking the risk to try to steal the base.

When to Steal Home

The baserunning play with perhaps the largest tilt toward taking a risk is attempting to steal home with two outs in an inning. Zach pointed out last offseason that the Twins had not stolen home since Clete Thomas did so successfully in 2013, as part of a kind of double-steal where another runner (Doug Bernier) got into a rundown between 1st and 2nd base.

Zach wondered when the Twins’ next steal of home would occur and in the process, managed to conjure three(!) of them into being just this season, all by Castro, and all as part of a double-steal action with a runner on 1st base.

Despite the Twins ranking near the bottom in overall stolen base attempts, they lead the league in steals of home. I’ll link to the videos so you can check them out:

May 24, 2023 vs. San Francisco

July 15, 2023 vs. Oakland

July 28, 2023 vs. Kansas City

Notable about each of these attempts is that they came with two outs in the inning. That’s important because the break-even math for attempting to steal home drops dramatically with two outs, as you can see here:

Credit: Josh Goldman | FanGraphs | Accessed from: LINK

You’re probably familiar with logic behind the old baseball rule of never making the 1st or 3rd outs in an inning at third base. That same explanation can be extended to a runner on third base and thinking about taking home as well.

With 0 and one out, the runner can likely score on a ball in play (even an out), so the threshold for taking a stolen base risk is higher — 87% with no outs and 70% with one out. With two outs in an inning and many of the previously available avenues to scoring the run closed, the required success rate drops to just 34%.

In Castro’s three attempts of swiping home, the Twins had the high-strikeout rates of Joey Gallo (42.8%) and Taylor (34.3%) and Ryan Jeffers and his 40% whiff rate against breaking balls against a breaking-ball-heavy right-hander (Anthony Desclafani, 50% breaking pitches) at the plate.

In two of the situations, the Twins were ahead on the scoreboard and looking at adding insurance runs. In the other (Kansas City), the score was tied, but only in the fourth inning.

Needing only a 1-in-3 chance to be safe and those game situations, it’s not too hard to see how Rocco Baldelli would conclude that trying to steal the run was worth the risk.

Another Way to Get A Run Home from 3rd

That same kind of conceptual logic can also be extended to the Twins’ recent propensity for using safety squeeze bunt plays to scratch across runs. Minnesota leads MLB with 9 bunt attempts with runners on 3rd base this season. Going back to last season, they’ve attempted 15 bunts with runners on 3rd base, also a league-high.

All but two of those 15 attempts have come with one out in an inning. The two that weren’t were in the 11th and 13th innings, respectively and in situations where the pitcher batter matchups were unfavorable (Gallo vs. a LHP with two outs in a tie game as the home team, Sandy Léon as a double play candidate with no outs in a 5-5 game).

Deploying a bunt with a runner on third and 0 or two outs doesn’t make much sense, with the exception of the bottom of the 9th or in extra innings as the home team. With no outs, you’re potentially forgoing the opportunity for scoring mulitple runs with a bunt play. With two outs, a force out at first ends the inning. So, one out is the most plausible situation to deploy a bunt play.

In 14 of the 15 attempts, the Twins were tied or ahead on the scoreboard, and generally in close games where the value of adding a single run could also reasonably be thought to be worth the cost of forgoing multiple runs.

With one out, in certain score situations and with certain hitters at the plate, a bunt is a great way to add an insurance run or break a tie. Matt Winkelman, whose work has appeared at Baseball Prospectus and other sites, wrote about safety squeezes in 2016.

He was writing then about the Phillies, who had the league’s worst offense by wOBA and wRC+, and had the fifth-highest strikeout rate that season. He argued that bunts with runners on 3rd base were a useful tool for adding runs to the ledger for teams with poor overall offensive production.

He estimated, with 2015 run expectancy data, that the break-even rate for bunting with one out and a runner on third was about 75% at the league level. That assumes a league average offense. For a team that produces below that mark, especially with a high rate of striking out (and therefore less chance to bring in a runner from 3rd base with a batted ball in play), the break-even point for using a bunt drops even lower.

Of the Twins’ 15 attempts the past two seasons, only three have resulted in an out without a run scoring. Of the remaining 12, ten have resulted in a run scoring and another two times the runners were safe at both 3rd and 1st. That’s a 66.7% rate of scoring a run and a 80% rate of avoiding an out when attempting to bunt with a runner on 3rd. Those both likely exceed the break-even thresholds required in those situations.

Here again, it seems the Twins and Baldelli are being smart with the deployment of small ball tactics when the situations, and the risk-reward math, support it. While there’s nothing wrong with the complaints about the lack of balance and resilience of approach in the Twins’ offense, at least they are trying to complement that with calculated risks where they make sense.

---

John is a writer for Twinkie Town and Pitcher List with an emphasis on analysis. He is a lifelong Twins fan and former college pitcher. You can follow him on Twitter @JohnFoley_21.