In the World Baseball Classic currently underway, Major League Baseball is testing out a rule change designed to resolve tied games more rapidly. When a game goes into the eleventh inning (the second extra inning), each team will begin their turn at bat with runners already on first and second bases. For readers not familiar with baseball, this will make scoring easier, and therefore ties should be broken more quickly. This rule has been used in international baseball for a few years (and amateur players may have encountered a version of it in local baseball and softball leagues as well).
Does this rule risk advantaging one team over the other? A post last month by Tyler Cowen on Marginal Revolution claimed this unambiguously was good news for the team that bats last, citing the conventional wisdom of the “last-ups” advantage in baseball. In fact, Cowen suggests that people who don’t subscribe to a belief in the last-ups advantage probably don’t “get” baseball. Yet, if something is so self-evidently true, it should be easy enough to demonstrate. A few years ago, while working on another project, it occurred to me I could use the same techniques to quantify how much of an advantage batting last was. The final product appeared in Journal of Quantitative Analysis in Sports, but, in short, no study had ever been able to provide evidence to support the last-ups advantage hypothesis.
As a behavioural economist, I’ve gotten interested not just in whether it is an advantage to bat last, but why such a universal belief came to be and persists, even in the absence of corroborating evidence. I’ll argue that there are three cognitive or behavioural biases at play here: framing, salience and certainty effect.
A framing effect occurs when one’s perception of a situation is affected by the choice of the words or labeled used to describe the situation. Implied but unwritten in the statement is the assumption that all the decision making in a game of baseball is done by the team currently batting. Yet, I expect few if any baseball fans would agree with that premise. Surely the defense is not a passive actor. It is true that the second-batting team knows the number of runs scored in the inning by the first-batting team; but it is equally true to say the team in defence second knows this information too, and can act accordingly.
Being the batting team – the team that has the opportunity to score run – is equated with having strategic control of the game. In baseball the batting team is always identified as the offense, while the team in the field is the defense. But have you ever wondered why this is so? In cricket, if you hear that a player is “coming into the attack,” the commentator is not talking about the batsman; he’s referring to the bowler (the analog to the pitcher in baseball). In general, cricket terminology is much more fluid in assigning offensive or defensive roles to the batsman and the bowler, depending on the conditions and the situation.
History plays some role here. Early baseball was developed in part by cricketers. One of the ways in which baseball was distinguished from cricket was that the design of baseball created an greater emphasis on the role of fielding. The job of the pitcher was simply to toss the ball so the batter could put it into play. The modern role of the pitcher, to try actively to get the batter out, evolved over time. So, probably, in early baseball, the batting team was truly more in control of the progress of the game.
It is also well-established that situations or incidents which stand out can play a disproportionate role in judgments and decisions. These salient situations are particularly memorable and easy to recall. Think about a game that is tied going into the bottom of the ninth inning. If the team batting in the bottom of the ninth scores, the game ends immediately. This is obviously a dramatic event, so much it has a name: a “walk-off” win. Such games are naturally memorable.
In contrast, think about a similar game, except it is tied going into the top of the ninth inning. The team batting first scores a run in their turn at bat, and holds the other team scoreless to win.
This is less dramatic, and therefore less memorable.
In the previous examples, the team scoring in the top of the ninth did not know for sure this would lead to victory, while the team scoring in the bottom did ensure victory for certain. While true, there is little, if any, difference in optimal strategy in a case where scoring wins “for sure” versus “almost for sure.” This is, in a nutshell, the implication drawn out in my article. However, psychologically “for sure” and “almost for sure” feel different; this is the certainty effect. The certainty effect causes us to place greater weight than we should on the events in the bottom half of the inning, simply because more of the uncertainty of the game has been resolved.
There are therefore a number of psychological reasons that might lead someone to conjecture a last-ups advantage. There’s a counterargument for each of them. Yet, belief in the last-ups advantage is essentially universal and taken as a given. To account for this near-universality, I propose correlation neglect. When I was growing up, I would hear from many sources – newspaper articles, commentators on television, friends, coaches – that it was an advantage to bat last. If every one of them had thought about it independently and came to that conclusion, then indeed this would be very strong evidence. But that is probably not how it works. For example, if my coach told me about it, and so did my teammate, it’s quite possible my teammate learned about it because he was told about it by my coach. Most of us do not account for this effect when we evaluate claims others make.
Returning to the question of if the last-ups advantage actually exists in baseball, in truth: I don’t know. All the evidence suggests there is no practical difference. People who believe in the last-ups advantage will be just as successful in baseball as those who don’t believe in it; there’s no practical cost to believing it. So, while parts of the baseball community might be uncomfortable with the introduction of the tiebreaker rule for other reasons, at a minimum, we probably do not need to worry about the rule giving one team an in-built advantage over the other.