For decades concepts like clutch hitting, team chemistry and desire to win were central to how baseball was understood and analyzed. They were default explanations that could be used to explain anything and which made it unnecessary for most analysts to think more seriously about the game. These explanations were one of the first targets of SABRmetrics which revolutionized how we understand baseball. Bill James and others created quantitative methods to explain baseball so that the conventional wisdom of chemistry, clutch hitting and the like could no longer be used as overly simple explanations of more or less anything that happens on the ball field.
A backlash has begun as thoughtful people, including James himself have started to consider that these very human factors which have an impact on most other human activities probably should not be entirely discounted in baseball either. The true problem with these types of explanations is not that they are wrong but they are the ugly offspring of the intellectual laziness and tautology which too frequently permeate conventional baseball analysis.
Team chemistry, for example, probably does exist, but when every team that wins is described as having team chemistry, the concept begins to lose its meaning. When teams that clearly lack this chemistry win, such as the Yankee teams of the late 70s, pundits often attribute the success to the team’s ability to respond to pressure or to the team being able to play better because of the tension. This is clearly nonsense too. An argument that asserts that teams with good chemistry win because they have good chemistry because they win is a tautology that cannot be tested and therefore has no meaning. It is this type of thinking to which serious quantitative analysts were reacting.
Clutch hitting is a similarly difficult issue. Some quantitatively oriented researchers have argued that clutch hitting simply does not exist. This is usually means that very few hitters demonstrate an ability to hit in the clutch over a period of several years. The data suggest that this is reasonably convincing, but is not quite the same as saying clutch hitting does not exist. This year, for example, Yankee catcher Francisco Cervelli has proven far better in the clutchthan in other situations. While it is very likely that this type of clutch hitting will either stop sometime soon or fail to carry over into next season, That does not mean it has not existed this year.
Every year there will be a few players like Cervelli who will hit particularly well in clutch situations, some will even continue to do this in the post-season. This is almost a mathematical inevitability due to the large number of players, expected deviations in performance and the role of luck in baseball. Unless it is repeated over time, however, it is difficult to meaningfully define a few clutch hits as being due to a player being a genuinely better hitter in the clutch.
Nonetheless, Cervelli’s ability this year in the clutch probably is due to little more than a few lucky bounces and breaks. The observation that Cervelli has done well in clutch situations in 2010 is obvious. Nothing in the evidence , however, suggests that Cervelli has some kind of special skill, unique ability to rise to a pressure situation, or is simply a winner with character, but that is the kind of narrative which still is very influential in analysis of baseball.
The idea that some players possess character, other intangible skills or are winners makes for fun journalism and good stories, but they do not stand up to more serious analysis. If, for example, having character was a real and variable skill, it would likely be randomly distributed across skill level at the major league level, but for some reason it is almost the great players who are described as having character and drive. The silliness and tautologies which permeate discussions of these characteristics, makes for an easy straw man for the more quantitatively inclined, but they cannot be entirely dismissed without some nagging possibility that team chemistry, being a winner, having character and other similar attributes exist and are important.
One of the next frontiers for baseball rigor is to bring the same rigor to understanding these ideas which has been brought to understanding offense, and increasingly, defense in baseball. The first step would be to clarify definitions so that it is understood what terms like winner, character and clutch mean. Currently, these terms are used very vaguely, broadly, and often as filler in lieu of more thoughtful analysis. It is easier to explain Derek Jeter’s success as due to him being a winner than it is to probe his numbers and find out that his consistently high on base percentage has really been what has made him so valuable.
In addition to clarifying terms, better methodologies need to be created. Instead of automatically ascribing great chemistry to whichever team wins the World Series or performs better than expected in a given year, it would make more sense to create measures of team chemistry before the season starts and to see if those measures have an impact on team or individual performance. Similarly, identifying ways to measure drive, being a winner, or being a good teammate in individual players and whether or not those characteristics effect the performance of a given team would also be a more serious methodological approach.
Making it more difficult to rely on often empty platitudes about clutch performance, chemistry or character as a way to explain outcomes in baseball has been one of the major impacts of the quantitative movement of recent years, but to some extent that has been the easy part. The harder work begins with revisiting some of these ideas, which are so influential in so many other areas of life, and seeking to gain a clearer, measurable and data based understanding of their import.