01 March 2010

Beyond Batting Average and interview with author Lee Panas

As many of us who discuss sabermetrics and statistics in general and how they relate to baseball; it is a subject that carries a broad range of awareness and acceptance in conversations about the game. The most recent statistical discussions have occurred over the past decade and taken place on the internet. As conducive as the internet is to share information and quickly acclimate to new truths and create or discover new applications, it has some problems as it is a very ephemeral library. Little is written down and much is carried on in a digitalized form similar to Homer. Web sites close shop, people get hired and their work gets scrubbed, and many baseball critics do not follow rules of citation very well. It makes it difficult to get caught up to speed on the conversation even if you are interested in how to use metrics in baseball.

A problem I have often seen when a book is written about sabermetrics is that they are often grooved for those in the mid- to hi- level of competency in the field. Introductory level works are hard to find and are a major reason why many web sites are beginning to post running SABR 101 FAQ's on their sites. It has also led to SABR series on Yahoo! and other online publications. Thankfully, I think I have found an actual book that is somehow both a handy introduction to what statistics mean and their utility in baseball as well as a resource for more advanced users of metrics (which is quite needed as metrics are often used incorrectly). Lee Panas' Beyond Batting Average is this book.

Panas does well to present the historical time line of statistics from Chadwick's original descriptive statistics to Branch Rickey and his usage of ISO to present day concepts like linear and nonlinear metrics. One amusing side note indicates how folks complained about new metrics in the late 1800s just as people do now with stats like WAR or xFIP. Panas does well to give you the formulas for advanced stats if you wish to play around with them, but moreso they act as an illustration letting you know what each statistic accounts for and how they are weighted. In the text, he simply and straight-forwardly explains what the formulas mean and how that effects statistical outputs. He informs the reader what each stat does well and their limitations.

As you have noticed I rarely review a book, but this is one that I think merited that. It is a solid addition to anyone's library. The introductory baseball statistics reader who wants to learn more and be able to start engaging in conversations about metrics would find this most useful. More advanced readers will see this as being useful at times for the quick to grab metrics and have a solid citation when making a claim about repeatability for a certain statistic and how that might relate to a skill.

Lee was kind enough to take part in an interview with Camden Depot. It is shown in full after the jump.

Camden Depot: I often find it important when reading a book to understand who the author is and from where personal experience is he/she writing. I was hoping you could give us your background, information about the blog you write, and other items you think potential readers of your book would be interested in.

Lee Panas: I have been interested in sabermetrics since the early 1980s. Bill James was writing his Baseball Abstracts at the time and he was a big influence on me. In addition, I was studying mathematics and statistics in college and some of my classes required me to apply statistical methods to real data. So, I used baseball statistics in a few of my projects. This gave me the opportunity to prepare for my career as a research analyst while learning more about baseball at the same time.
I have been writing at DetroitTigerTales.com since 2005. I discuss a lot of topics besides statistics there - Tigers history, prospects, transactions, etc - but sabermetrics has always been the overriding theme. I have written a lot about baserunning and fielding statistics in particular. In addition to my blog , I am currently writing about the Tigers for John Burnson's Heater Magazine. I have also contributed to books such as Tigers Corner, Graphical Player and How Bill James Changed Our View of Baseball.

CD: From experience, I have found that taking on a large endeavor (i.e. a book) often comes from a moment of transcending excitement and then several months of trying to pay dues to that initial moment of inspiration. Was there a singular moment that convinced you that you had to write this book? If not, where do you think this urge originated?

LP: There was not really a single moment where I decided to write a book. It's something that has been building for a long time. I have avidly been discussing baseball on message boards and blogs for a couple of decades. Much of my time has involved explaining sabermetrics to people. In the early years, there wasn't a great deal of interest in the subject. However, I have noticed a lot more people getting curious about it in recent years. Blogging has been a good way to educate more people about the field but I can only give small does in that format. Thus, it became increasingly clear that I wanted to write a book where I could tie all the information together into a complete story.

CD: The sabermetric field has seemed to me that it is moving toward a goal that is to evaluate the true talent of a player. That is, a progression from simple descriptive statistics to predictive/evaluative ones. From what I have read in your book as well as the outline, it seems to me that you very much support this perspective on moving toward predictive statistics. Do you think anything (i.e. blind luck, MVPs) is lost or being lost (or devalued) in terms of appreciation by doing this and is that bad?

LP: For the most part, I think trying to evaluate a player with predictive statistics is a good thing. As you know, some of the more traditional statistics are based on things that are largely beyond a player's influence. It makes sense to eliminate as many outside factors as possible and to evaluate a player on things he can control most easily. I think that we are getting increasingly better at finding statistics which more accurately define a player's true talent.
There is, however, the potential for skills to get lost in this approach if we are not careful. I'll use ERA as an example. Statistics such as a strikeout rate, walk rate, ground ball percentage and FIP are more predictive than ERA for most pitchers. However, they don't take into account the ability of a pitcher to pump up his fastball or to induce a double play ball with runners on base. It has been shown that this has more to do with overall pitcher quality than clutch pitching ability. However, I think more work needs to be done to identify pitchers who may have more ability to pitch in high impact situations than others. For these pitchers, ERA might be telling us something that the so called true talent stats are missing.

CD: One of the more fearsome aspects of writing a book on baseball statistics is that the community is internet based and the level of information progresses so quickly. I think you have done a great job is writing something that is rather resilient in light of that. How did your recognition of the blinding pace of statistical innovation direct you in writing this book?

LP: First, I think the fast pace of sabermetrics has left some people frustrated because they can't keep up. That is one of the reasons I wrote the book. By organizing all the information in one place, I think it helps people catch up. I also hope through reading the book, fans will understand more about the reasoning behind the statistics. This should help readers make sense of future measures more quickly.

One thing I have done throughout the book is to include a bit of history showing the evolution from simple traditional measures to the more advanced measures of today. I think this better prepares readers for the future advances in statistics than if I just dove into the present and talked about the statistics developed in the last two years.

CD: You have mentioned before that you largely relied on peer review for this book. Can you tell us some of the people you consulted in writing this book?

LP: I think peer review is essential in this kind of effort. It's a complex subject and I wanted make sure I got everything right. There are a couple of important sections of the book that might have been left out if it were not for the suggestions of others. There was also a matter of educating myself in certain areas. I had a good handle on most of the statistics coming in but there were a couple of instances where I was not interpreting statistics as accurately as I could. So, I was glad to have the developers of the statistics correct me. Finally, I think peer review helped me write the book in a way that would make it accessible to as many people as possible. For example, there were some sections in early drafts which reviewers felt were a little too mathematical for my target audience. I either eliminated those sections or wrote them more simply.

The reviewers ranged anywhere from very talented sabermetricians who know the field better than I do to intelligent baseball fans who are relatively new to sabermetrics. Some of the key contributors in alphabetical order are John Dewan, Brandon Heipp (aka U.S. Patriot), Chuck Hildebrandt, Justin Inaz, Mitchel Lichtman, Kurt Mensching, Pete Palmer, Samara Pearlstein, Tom Tango and Geoff Young. There were many others but I think this gets the point across that I got input from a lot of different kinds of people.

CD: Thanks again Lee. Very well done.

LP: Thanks very much for the interview.

Lee Panas' blog is Tiger Tales and has recently written, Beyond Batting Average.

1 comment:

Steve L. said...

Thanks for this Crawdaddy. These interviews just get more and more interesting. This definitely looks like a book to check out. When people use all these acronyms I feel just like the citizens who show up on base and don't know the lingo.