Developing a Method

How do you measure greatness?  The Baseball Hall of Fame is an institution built to try answering that question as it pertains to the game of baseball.  There have been lots of attempts at coming of with objective methods for comparing the top players, by such analysts as Bill James, Jay Jaffe, and many others.  One of the challenges for an objective measure for something like Hall of Fame worthiness, however, is that there’s by nature a subjective element to it.  Unlike a straight value metric, where the goal is to strip as much subjectivity out of the results as possible, an evaluation of a Hall of Fame case requires personal judgments on things like career value versus peak performance, how to treat career interruptions, how to weigh career highlights and postseason performance, and so on.  My goal in this series is to develop a framework to help weigh these factors.

The first step in this process will be to look at some of the issues that need to be addressed in developing this system.  How much weight should be put on brief excellence versus sustained effectiveness?  How should durability be factored in as compared to dominance?  Should offense and defense be addressed as two separate parts, or should the focus be on the total package?  There’s subjectivity involved in answering those questions, and one of my goals is to be able to show a range of outputs using the same data, in order to demonstrate how the core methodology can be tweaked to answer these questions from different angles.

One final note here is that this is not an attempt at building a superior value metric.  It’s not a value metric at all, but rather it’s an attempt at interpreting existing metrics for different purposes.  I will be using the WAR metrics on Baseball-Reference as a foundation, for three main reasons:

1. The data is easily accessible
2. Historical players are treated the same as modern players
3. The offensive and defensive breakdown of WAR is presented neatly

With that in mind, any flaws or limitations present in the baseline WAR will show up in my numbers as well.  With some tweaking to adjust for different scales or baselines, the framework I’m building can be applied just as easily to Win Shares, WARP, other variants of WAR, or any other value metric.  I’ll try to keep my discussions neutral of any metric, in the interest of modularity.  There’ll be some light number crunching here, but I’m more concerned with the theory than the specifics at the moment.