I recently got a chance to read Rob Vollman’s Hockey Abstract (available for purchase in either hard copy or .PDF format here). What follows is my review of the work.
For those unfamiliar with Rob Vollman, his work at Hockey Prospectus, especially player usage charts has done a great deal to both further statistical analysis in hockey, but also to simplify it in order to make it easier to grasp. This is his first year publishing Hockey Abstract, a project intended to help do for hockey analysis what Bill James’ Baseball Abstract did for Sabremetrics in that sport. He has also worked for the last three years as a co-author of Hockey Prospectus, whose fourth edition is due out soon.
Hockey Abstract is 200+ pages with numerous chapters covering a lot of different angles, but really it is a two-part book. In the first half, Vollman dedicates significant time into the due diligence necessary to answer the kinds of questions that analysts would hope to be able to confidently answer: who is the best player in hockey? Which goalie is the best? Can we confidently predict future performance? How much luck is involved?
The second part of Hockey Abstract goes into great detail on the statistics involved in making those kinds of determinations, as well as the strengths, shortcomings, and current state of analysis in hockey. Each step of the way, Vollman meticulously catalogs the various sources used and resources available, a good reminder that this kind of analysis is a living ongoing community project.
It’s hard to say whether it was the right decision to set the book up as it was. The first half is definitely the more “fun” half, which does a good job of showing the usefulness of a topic which can be very dry. For somebody who isn’t familiar with a lot of the concepts of the analysis, a person just kind of has to take a few things at face value until he or she gets into the chapters which more thoroughly explain some of the metrics.
I’ll also cite Draglikepull’s review in his biggest criticism of the book on Pension Plan Puppets, which manifests in the second half:
While the first half of the book is a lot of fun, the second half is a more clinical look at the kinds of statistics that modern hockey analysis uses, and it’s in this second half of the book that my major criticisms manifest. Rob discusses a number of statistics that I personally think are of fairly limited value, and they all have one central problem: they’re over-complicated. In Thinking Fast and Slow Daniel Kahneman makes a very good case that making a formula more complicated rarely adds much value past a certain point.
I recommend you head over there and read because he goes into much more detail to properly flesh out that criticism.
As a Red Wings fan, there’s also a criticism I have which I fear is the kind of mistake which allows far too many people to jump to the defensive and cliched argument against stats-based analysis: the “you don’t watch the game!” argument. In the closing thoughts on the ‘Who is the Best Defensive Player?’ section finishes up on 16 pages worth of detailed analysis with a cringeworthy sentence about how Pavel Datsyuk “doesn’t kill penalties.”
It’s a technical mistake and one that doesn’t necessarily defeat the conclusion at hand, but it’s also an unnecessary absolute and, seeing as how the book is by its very nature technical, it’s the kind of mistake that can fester. Datsyuk hasn’t spent as much time per game killing penalties as Patrice Bergeron, but to say he doesn’t kill penalties is as avoidable as it is incorrect.
Despite the two-paragraph rant (and the fact that I intentionally took a few days off reading Hockey Abstract after that because I didn’t want to let it color the rest of the reading for me), I think the book is well put-together and well-argued. The author’s voice is both authoritative and respectful, which is a balance that hasn’t always been easy to achieve in the discussion about hockey analysis.
Overall, whether you’re brand new to the concepts discussed in the work (in which case you may want to read the book backwards) or somebody interested in both a deeper understanding of the state of hockey analytics and a fun look at their application for this year, I think this would be a worthwhile purchase. I’m afraid that those who consider themselves advanced students of the topic probably aren’t going to find much value in the work, although if you’ll excuse a platitude for a moment, I’ll also say that only a fool really ever says he’s done learning.
