One of the things I missed when I had to skip out of SABR 45 Saturday was the committee meeting for SABR’s largest research committee, Statistical Analysis. Unlike many of the other committees, the Stat Analysis committee didn’t have a group project to work on, in part due to the individual nature of most members’ research. A couple ideas were bandied about the meeting during SABR 44, but it took until SABR 45 to get one of those ideas off the ground.
A few weeks back, Phil Birnbaum, the chair of the committee and editor of the By the Numbers newsletter, announced that group project. The idea is to create a crowd-sourced list of key resources for helping newcomers to sabermetrics learn what has been done and provide to him or her the foundation for additional contributions.
There are plenty of books and articles which I could cite, so I’m going to start with the broad resources that cover multiple topics. That means it does skew towards books. They are listed in the order they came out of my head.
Before I get into my long list, I want to invite you, dear reader, to contribute your recommendations to this project. If you do so in the comments, I’ll be sure to pass them on.
- The Numbers Game, by Alan Schwarz. This book came up recently when Graham Womack of Baseball Past & Present and I discussed the importance of this book and a few other titles that will make there way onto this list as for which one we’d recommend first. We both agreed that this title is where we’d tell others to start. A fantastic history of baseball’s numbers, and the understanding of how a particular stat like batting average or OBP came to be is key to understanding any analysis with those measures.
- The Hidden Game of Baseball, by John Thorn and Pete Palmer. It’s over 30 years old, and it might be the most important book in sabermetric history. There’s a reason I started my sabermetric research database project with this book: it was The Numbers Game before Schwarz wrote his book with its concise history of baseball statistics AND it introduced the linear weights model to the world, which is much more of the mathematical foundation of modern sabermetrics than anything put out by the most famous name in the field.
- The Bill James Abstracts, both the annuals printed from 1977-1988 and the Historical Abstract (first published in 1986, revised and updated in 2001). For the many who grew up before Al Gore’s invention came to the masses, these books were how they were introduced to sabermetrics. Bill isn’t a statistician in the academic sense, but his understanding of baseball endows his analyses with tremendous insight.
- Curve Ball, by Jim Albert and Jay Bennett. I have a rare relationship with this book. I read it before I ever read anything by Bill James. It steered me from being a pure mathematics major in college to a statistics major, which is one of the 5 best decisions I have made in my life. So yeah, I hold this title in high esteem for many personal reasons. That being said, it might be the best book for helping aspiring saberists to start understanding mathematical statistics, which is essential to advancing the field.
- The Book, by Tango, Lichtman, and Dolphin. For many saberists, this is the modern treatise on the subject. Grounded in an understanding of Palmer’s Linear Weights system, they introduce wOBA and use it to explore every facet of the game.
- For online reference guides, the FanGraphs Sabermetric Library is my preferred site, as I consider to be the most complete. Neil Weinberg is also authoring weekly posts to explain the ins and outs of various metrics, helping keep the reference guide current with new research.
- The Best of Baseball Prospectus: 1996-2011 is a 2 volume set that is a compilation of the most important articles from the first 15 years of that sites’ history. This is essentially my proxy for the excellent writing on that website, including Voros McCracken’s article on DIPS Theory and Keith Woolner’s “Baseball’s Hilbert Problems“.
- Baseball Hacks, by Joseph Adler. The ability to analyze data is great, but it is useless if you can’t get data to analyze. While the book is somewhat dated, it’s a great introduction to many of the coding skills required to do sabermetrics efficiently in the computing era, and one I still find worthwhile to have on my shelf.
- SABR101x, the massively open online course at edX administered by Boston University and designed by Andy Andres et al. If you prefer a class-based method for learning sabermetrics, this is as good as you’ll find. There are tracks on the history of sabermetrics, statistics, SQL/R skills needed, and a build up to understanding some key metrics used by saberists.
One thing I want to keep separate from this list is SABR’s own Guide to Sabermetric Research, which was put together by the aforementioned Phil Birnbaum. His involvement spearheading this SABR 101 project is why I leave it out for now. I have a sense that it will be that guide that is updated as a result of this group work.