Understanding Statistics: Batting Average
Over the course of the next few months, MLB teams will spend more money than most people make in their entire life. Their moves will be scrutinized and examined under a microscope by journalists everywhere. The professionals who make the signings spend hours deciding which players to pursue. What criteria do they use and what criteria should we use to examine the signings?
Last week I introduced the idea of statistical regression analysis in this article. This week, let’s apply the same concept to determine which statistics are the best indicators of success. The first, and most important, concept to understand is that all statistics are flawed. All statistics have inherent errors that make them imperfect indicators of a player’s success. Still, statistics are the best piece of insight we have into a player’s success. MLB players live and die by their statistics. The difference between a .280 batting average and a .300 batting average can cost a player millions of dollars. But, should batting average really have such a strong sway?
Batting average is the standard barometer of a player’s success. Every baseball fan understands that a .300 average is very good, a .270 average is decent, and a .250 average is fairly poor. Because fans understand the concept of batting average so well, it is commonly used to measure a player’s skill level; however, the statistic of batting average is extremely flawed. Batting average takes the number of hits and divides it by the number of at bats. The statistics measures each and every hit as exactly equal. A homerun is worth the same amount as a single. Clearly, there is a big problem in this unjust measured equality.
We know batting average is flawed, but teams continue to use the metric as a measure of a player’s skill. When you look up on the scoreboard, you commonly see a player’s batting average. All of the hype surrounding batting average begs the question: how strongly does a team’s batting average determine the number of wins a team earns? The following graph represents the correlation between batting average and wins. The graph is followed by the correlation statistics I discussed last week.
r: 0.337001
r squared: .113569
As you can see from the r score, there is a moderately weak correlation between a team’s batting average and the number of wins they had in 2008. In fact, only about 11% of a teams wins can be attributed to their batting average. Why should fans of the MLB place so much weight in a player’s batting average, when there is little correlation between batting average and a team’s overall success? There must be other statistics we can use, right?
I’m glad you asked. Indeed, there are. Over the next couple of weeks, I’ll be analyzing a number of other statistics using the same methods in order to determine which statistic is the best indicator of a team’s overall success and of a player’s true skill, but keep one thing in mind: all statistics are inherently flawed.
Last week I introduced the idea of statistical regression analysis in this article. This week, let’s apply the same concept to determine which statistics are the best indicators of success. The first, and most important, concept to understand is that all statistics are flawed. All statistics have inherent errors that make them imperfect indicators of a player’s success. Still, statistics are the best piece of insight we have into a player’s success. MLB players live and die by their statistics. The difference between a .280 batting average and a .300 batting average can cost a player millions of dollars. But, should batting average really have such a strong sway?
Batting average is the standard barometer of a player’s success. Every baseball fan understands that a .300 average is very good, a .270 average is decent, and a .250 average is fairly poor. Because fans understand the concept of batting average so well, it is commonly used to measure a player’s skill level; however, the statistic of batting average is extremely flawed. Batting average takes the number of hits and divides it by the number of at bats. The statistics measures each and every hit as exactly equal. A homerun is worth the same amount as a single. Clearly, there is a big problem in this unjust measured equality.
We know batting average is flawed, but teams continue to use the metric as a measure of a player’s skill. When you look up on the scoreboard, you commonly see a player’s batting average. All of the hype surrounding batting average begs the question: how strongly does a team’s batting average determine the number of wins a team earns? The following graph represents the correlation between batting average and wins. The graph is followed by the correlation statistics I discussed last week.
r: 0.337001
r squared: .113569
As you can see from the r score, there is a moderately weak correlation between a team’s batting average and the number of wins they had in 2008. In fact, only about 11% of a teams wins can be attributed to their batting average. Why should fans of the MLB place so much weight in a player’s batting average, when there is little correlation between batting average and a team’s overall success? There must be other statistics we can use, right?
I’m glad you asked. Indeed, there are. Over the next couple of weeks, I’ll be analyzing a number of other statistics using the same methods in order to determine which statistic is the best indicator of a team’s overall success and of a player’s true skill, but keep one thing in mind: all statistics are inherently flawed.
