Love it or hate it, extra innings won’t look the same starting July 23.
When Major League Baseball begins its 2020 season, a new rule will place a runner on second base to start each half-inning once the game reaches the top of the 10th. For us traditionalists, this will be a moment to breathe deeply, count to 10, and embrace change.
The purpose of this article won’t be to support or oppose the new rule. If that’s what you crave, you’ll have no trouble finding hot takes elsewhere. But one aspect of the new rule is interesting from a strategic perspective: Does a manager improve his team’s chance to win by bunting his pre-positioned runner to third base? And as a result, might the rule mitigate or even reverse the recent sharp decline in the frequency of sacrifice bunts?
To start, let’s examine how rapidly the decline in sac bunts has accelerated in recent seasons. A graph follows, but these are the key takeaways:
- At its wartime peak in 1945, there was one sacrifice bunt for every 56 plate appearances.
- Over the next 13 seasons—as home runs more than doubled—the rate of sacrifice bunts declined by 37%.
- There followed a period of more than a half century during which sac bunts plateaued and then declined slowly, from one per 90 plate appearances in 1958 to one per 111 PAs in 2011.
- The rate of sacrifice bunts has dropped in each of the eight seasons since then. That is the longest such streak of year-over-year declines in recorded history, dating back to 1894, and it accounted for a 54% reduction during that time.
Another new rule all but guarantees the decline will not only continue but accelerate: the universal DH, with designated hitters finally to be used in all NL games. Last season, pitchers were credited with more than half of all sac bunts in the majors (56%). By contrast, leadoff batters in extra innings accounted for only 0.5% of all plate appearances. No strategy designed for such an occasional situation could even dent the collapse in bunts that will result from pitchers no longer batting.
As for whether managers will tend to bunt in those newly legislated situations, we have historical data that answers that question as well. The rule is new, but the situation is not. Managers have always had to decide whether to bunt in extra innings of a tie game if a runner reaches second base before the first out is made. For the past 30 years, they have overwhelmingly answered, “No.”
As to whether those decisions were sensible, this isn’t as simple as comparing the run-scoring rates from a none-out/runner-on-second situation and a one-out/runner-on-third state (1.17 and 1.28 runs, respectively, last season). Those two figures are similar enough that the specific game situation takes precedence over a theoretical one-size-fits-all approach. Who’s on second, who’s at bat, who’s on deck, and who’s available to pinch hit are all key elements in the equation. Is there a high-strikeout pitcher on the mound? Does his ball have late movement? How well does he field his position? Managers should be assimilating all that information and more when deciding whether to bunt.
A road manager also must decide whether to play for one run or a multiple-run inning. If a shutdown closer is warming in the pen, one run should be enough; otherwise, it might be prudent to play for a bigger inning. But the home manager has a much simpler equation to consider. Since one run wins the game, plating the runner on second base is all that matters.
So one would think that regardless of how often managers choose to bunt in these none-out, runner-on-second situations, they would do so more often at home than on the road. But counterintuitively, that hasn’t been the case. Over the last 30 seasons, road managers bunted on 29% of those plate appearances compared to 23% by home managers. The sample size in any one season isn’t significant. But across 30 years the odds that such a disparity would have occurred at random are 90-to-1.
The following graph highlights the difference between the percentage of home and road bunts. In this case, we didn’t limit the analysis to successful sacrifices. We included any plate appearance that ended in a bunt: hits, sacrifices, and other outs, including third strikes on attempted bunts. (For the record, we performed the same study using only sac bunts, and the results were virtually the same.) These are three-year moving averages to help smooth out the single-season percentages:
Informed fans can intelligently debate whether a bunt is the right move for the situation created by the new extra-innings rule for 2020. But is there a plausible explanation for why home managers have chosen to bunt less often than road managers when faced with that situation in the past?
It could be that the solution is as much about psychology as it is about baseball. Here’s an interesting take from my former colleague and statistical analyst, Dr. Gilbert Traub. Gil is a long-suffering Mets fan with a background in both applied mathematics and psychology. He suggests that the visiting manager is more inclined to bunt for fear of squandering his team’s one and only chance to score.
Extrapolating on Dr. Traub’s idea, consider the home manager who, unlike his counterpart in the opposing dugout, can’t lose without getting at least two more turns at-bat—one in the current inning, and another to follow if they fail to score. To the extent that there’s a stigma attached to bunting, might he be tempted to forgo the sacrifice and give his players three swings at driving the runner home? After all, the worst outcome would be another jump ball in the next inning. The home manager is playing with house money.
For the record, over the last 30 years, the winning percentage of teams that bunted in those situations was higher than that of teams that didn’t bunt, but not by a statistically significant margin: by 36 percentage points on the road (.699–.663) and 32 points at home (.845–.813). Note that the record for road teams will not be as high this season as in the past since the home team is guaranteed to start the bottom of the inning with a runner on second base.
The situational elements of bunting aren’t going anywhere, so it will never be as simple as settling for blind adherence to baseline percentages. But to ignore those percentages altogether might be just as foolish.
Research support provided by Jacob Jaffe and Tom Paquette.