How Caution is Costing Teams on the Basepaths

Baserunning: Where 100% success is a failure

If your phone is on 1%, or if you’re just eager to get to the point, here’s the punchline: MLB teams are not particularly adept at deciding when to wave runners around at third base and are leaving runs on the table by being far too conservative in determining when they send guys.

Now, let’s back up. First off, since we’re new here, a short introduction is in order.

We’re the Oyster Analytics team, Maxfield Lane & Owen Riley - we’re excited to be starting a new weekly column for Down on the Farm, covering an eclectic selection of baseball and prospect-related topics. Our main gig is running a prospect projection site that we talk about a lot on X (@oysteranalytics) where we evaluate the tools of MiLB players relative to their peers and provide projections of their likely career outcomes.

That’s who we are; now on to our baserunning analysis, starting from a bird’s eye view.

If forced to summarize the insights brought to sports in the 21st century by analytics in one pithy piece of advice, it would be difficult to do better than these four words: be less risk averse.

Variations on this theme can be found just about everywhere you look. In football: go for it more on 4th down, even on your own end of the field. Basketball: take more three-point shots, even though you’ll miss them at higher rates than midrange jumpers. Soccer: play out from the back, even at the risk of a potentially costly turnover.

In baseball, too, examples already abound. Don’t give away outs with bunts, swing away. Don’t be cautious and strike out averse, swing for the fences.

This piece adds another risk-embracing tenet to the gospel: send runners home from second on singles more.

From 2021 to 2024, teams sent runners from second to home on singles about 60% of the time. Understandably, they were far more aggressive with more outs, sending over 80% of runners with two outs, but less than half with no outs. Teams rarely made the mistake of sending someone when they shouldn’t have — or at least were rarely punished for doing so. Less than 2% of runners from second trying to score on singles have been thrown out at the plate over the past four years.

Second Base Runner Outcome on Outfield Singles, by Out

One explanation is that third base coaches and runners, in some combination, are simply really good at understanding when to take the extra 90 feet. At first glance, and on an instinctive level, this seems plausible. As the figure below illustrates, they do adjust their approach by runner speed, sending top quintile speedsters far more than base cloggers. Even though these fast players are sent more, they’re not thrown out more, with the fastest quintile being sent at the highest rate and thrown out at the lowest rate. It would appear, then, that when teams flip the light to red, they’re doing so for good reason.

Second Base Runner Outcome on Outfield Singles, by Speed Quintile

On further inspection, however, this explanation doesn’t seem to cut it. If send rates were determined by teams’ impeccable abilities to determine when to send players, we’d expect, controlling for the speed of runners, over large samples for either: a) all teams to send runners at the same, optimal rate and for runners to score at the same rate or b) teams that sent runners more frequently than others would see their runners thrown out more, since those on-the-fence opportunities must be more risky. In reality, we see neither of these patterns.

Looking first at all instances, regardless of player speed, there is virtually no correlation between how often teams send players and how often these players are successful in scoring.

Send Rate vs. Success Rate, Scoring from Second on Outfield Singles (All Runners)

Of course, one potential explanation for this lack of relationship would be that we’re not yet considering differences in speed between teams. Perhaps the Rays, with the highest send rate, having a similar success rate to the Mariners, with the lowest send rate, is due to the Rays’ possessing more speed on their roster in the 2021-2024 period. If that were the case, maybe the Rays could send runners more aggressively without being thrown out more.

Not so. The figure below replicates the plot above but includes only runners in the middle quintile of sprint speed.

Send Rate vs. Success Rate, Scoring from Second on Outfield Singles (Middle Quintile Runners)

Teams are sending runners of the same speed at vastly different rates and obtaining the same results. Sure, some of this is down to differences in the actual content of those plays for each team (deep vs. shallow, small differences in speed), but even limited to this middle quintile, we’re working with nearly 4,000 observations. It’s unlikely confounding factors worked together perfectly across this large sample to perfectly counteract any semblance of correlation. Park factors are also less relevant in second-to-home situations than they would be, for instance, on runners scoring from first on a double. In that case, the depth and shape of the outfield becomes crucially important.

While dimensions still can impact where fielders line up and field balls hit on singles, it certainly plays a more muted role in these situations. What’s more, the fact that a team like the Reds did not have a single player thrown out in 152 opportunities to send middle-quartile runners home suggests a pretty flagrant over-avoidance of risk. Six other teams similarly didn’t have a single middle-quartile player thrown out in the four-year sample.

Essentially, this indicates that teams are not great at making margin calls in these situations. The Mariners are sending average-speed players at a 13 percentage point lower clip than the Rays, but aren’t mitigating any of the risks of the situation in doing so. In other words, the Mariners aren’t identifying situations where runners are more likely to be thrown out and then throwing up the stop sign–they’re just indiscriminately holding runners more.

This stuff matters. Teams who send more are getting a far greater run expectancy out of these situations.

There’s a strong correlation between teams sending runners more and generating more expected runs: on average, every additional percentage point of send rate generates .012 more runs per opportunity. To put that in more concrete terms: the Rays are generating 0.2 more runs per send/hold decision than the Mariners. Over the course of a season of opportunities, that translates into about 14 additional runs.

Expected Run Value per Decision by Send Rate (All Runners)

Some readers may still (understandably) be thinking this might be all about speed. Speed certainly helps score runs, but it’s not the explanation for this phenomenon. The chart below, which includes only the middle quintile of runners, shows the existence of a slightly less tight but still strong relationship between send rates and run expectancy.

Expected Run Value per Decision by Send Rate (Middle Quintile Runners)

In both the case of all runners and strictly the middle quartile, this relationship is statistically significant, with p-values well below 0.01. For the sake of space, the other quintile-specific analyses are not included, but they produce similarly significant and notable positive correlations.

So, let’s get to the takeaways for teams. To maximize run expectancy, teams should encourage baserunners and third base coaches to be far more aggressive in almost all situations. The matrix below displays the “break-even” send percentages for a runner on second base given the situation (e.g., if it is bases loaded no outs and I am 70% sure I will make it home I should go, whereas if I’m 50% sure I will make it home I should stay at third):

In Bold: Breakeven Success Rate In Italics: Actual Success Rate

In cases where there is no one out and no runner on first, coaches and runners need to be incredibly sure they will score to justify a send attempt (when there is a runner on first, this becomes less paramount since it also enables, in most cases, that runner on first to advance to third). In all other cases, the likelihood of scoring doesn’t even have to break 75% for the right decision to be to send the runner. In fact, with two outs, a near 50/50 scoring rate still results in an increase in run expectancy!

Of course, runners and base coaches don’t have some internal barometer capable of infallibly calculating success probabilities down to the decimal places. That’s not necessary to take action here. The gap between current strategies and the optimal approaches is a wide one. It’s fair to say that the average runner or coach has some kind of intrinsic understanding of whether their likelihood of scoring is closer to a coin flip than it is to a certainty.

As it stands, teams are sending runners too conservatively in every single one of these situations. It’s understandable. When a runner gets thrown out at home, the cameras find the third base coach, and in some cases, he finds the door out of the organization in the offseason (see: Phil Nevin and the Yankees post 2022). Baserunners themselves get castigated by announcers and fans.

Nevertheless, the reality is that teams need to get more ambitious. Their risk avoidance when sending runners is not well-founded in scoring more runs or winning games; it’s based on a fear of making individual incorrect decisions that reflect poorly on the team and coaches, who are expected to play baseball in alignment with league norms. A coach who gets 40% of their runners thrown out at home with two outs may very well be making run-positive decision-making but would no doubt be subject to fan and media inquest. On the contrary, hitters are blamed far less when they fail to score held runners in subsequent at-bats (“that’s baseball!”). Shifting the perception of send/hold calls from a series of individualized decisions to a team philosophy is an integral part of implementing a more aggressive running mindset, as it would help alleviate and dissipate some of the stress that coaches would endure from a philosophical shift (and winning helps too).

Perhaps the Dodgers’ well-chronicled and successful approach to the Yankees in the World Series–forcing them to execute plays in the field–will provide a more welcoming environment for this philosophical shift. A lot has to go right for a player to be thrown out at the plate. As teams hopefully aim to get over their fears on the basepaths in these situations, they can take solace in the fact that even if they get the call “wrong,” the chance of a critical defensive imperfection is far from negligible. Or maybe it will be a very slow burn, as we’ve seen with teams fighting with themselves to gradually take more chances on fourth down in the NFL. In any case, there’s an advantage to be had for whichever teams are bold enough to seize it.