Charting The Starters

Colby Cosh used an off day recently to create an interesting graph representing all of the pitchers who put in enough innings to qualify for the ERA title last season. The X-axis represents number of walks (control), and the Y-axis represents strikeouts (power). The third variable Cosh introduces is the number of home runs allowed, which he portrays by the size of the dot that represents each pitcher on the diagram (the bigger the dot, the more dingers allowed).

Click here to see the graph (if you can't read the graph initially, just leave your pointer over the graph, and a box allowing you to view it at normal size will appear). It's a neat little exercise, one in which Colby has combined the three variables that Sabermetric analysis says are the only data points for which pitchers are completely responsible (note he uses ERA, which Sabermetric analysis has found wanting, as a method to determine his sample, not make any judgements on it).

Which leads Colby to note:

If you're a baseball fan you probably grasped all this in ten or 15 seconds. It's a picture of the "universe" of starting pitchers (it would be really cool if you could animate it), and I wonder if I am the first to make such a thing. It frustrates me that sports pages and magazines so rarely challenge you with data representations even as marginally sophisticated and complex as this. They're talking to a market of people who can tell you what Terry Sawchuk's goals-against average was in 1964-65 without checking, but they shy away from putting anything on the page more complicated than a grade-school bar graph.

What Colby is touching on here is the difference between being able to create statistics, and being able to understand and appreciate them. As a student like me who, to put it kindly, was "Math-challenged" from the earliest moments of my academic career, I understand this pretty implicitly (and this is even after taking college calculus and probability and statistics as a high school senior).

Then again, this really shouldn't be an excuse for being able to look at the data, parse it for real information, and then draw conclusions from it. For me, this wasn't a skill easily won, and it took getting slapped in the face by the business world to get there (especially my experience as a college newspaper editor, when I learned I was responsible for the bottom line financially as well as editorially).

I only point this out for one reason: to quote Barbie, "Math is hard". Which makes the work that literally thousands of amateur Sabermetricians have done all the more amazing. I wasn't always a devotee of Sabermetrics for precisely this reason.

But with Colby's work as a starting point, here's another question I'd like to see answered. Instead of charting one season with all the pitchers in the majors, why not chart one pitcher's entire career? Further, why not chart the career of every "starting" pitcher in the history of the game, and then average the results to discover at what age a pitcher's performance begins to erode? If you like, you could parse your sample size based on power (strikeouts) vs. control (walks).

Eric 12:40 PM

Trackback Pings

TrackBack URL for this entry:
http://www.ericmcerlain.com/mt/mt-tb.cgi/2996

Listed below are links to weblogs that reference Charting The Starters:

» Graphing Pitchers from Baseball Musings
Eric McErlain points me to this post by my favorite Canandian blogger, Colby Cosh. In the post, Colby creates a 2-D projection of the K/Inn, BB/Inn and HR/Inn of each ERA qualifying pitcher from 2003. Basically, if you are in... [Read More]

Tracked on February 27, 2004 02:25 PM

Comments

Posted by: at March 2, 2004 12:55 AM

Post a comment