Nats Blog Book Review: The Hardball Times Baseball Annual 2006

I mentioned a couple of weeks ago that I'd received a copy of The Hardball Times Baseball Annual 2006 and would review it for you. The book is terrific, and well worth the purchase.

It is organized into five sections: "The 2005 Season", "2005 Commentary", "History", "Analysis" and "Statistics". The first two categories recap the past year, with articles on each division and the postseason in the first chapter. These articles are written well but a bit mundane -- they are a good utilitarian review of what happened, but not much insight that those familiar with the races in each division won't have read before. The next chapter on 2005 Commentary is a nice collection of essays on the various big topics of the year, like the steroid scandal, World Baseball Classic, "The DePo Era" and other interesting pieces. I found these to be worthwhile reads.

The "History" section contains the two Bill James articles, one on what happens when a team uses young pitchers predominantly (not much good, according to him) and perhaps the definitive article (and, we can hope, the last) article on why Bert Blyleven should be in the Hall of Fame (James says he should, but without the hubris and condescension of some).

For me, the most interesting part of the book is the "Analysis" section, which contains statistical articles from a "macro" perspective rather than from a team or player perspective. This year many of the articles center around some research into "batted ball" data that THT has pored over. Instead of focusing on the traditional outcomes of baseball stats (singles, doubles, triples, etc.), they have reoriented the play by play data around types of batted balls (groundballs, outfield flies, infield flies, line drives). From this viewpoint you can start to more accurately asses "luck" for both pitchers and hitters by determining what typically happens on a line drive and compare it to what actually happens. For example, 71 % of line drives result in hits, making a line drive on average worth about 0.356 runs. In contrast, each 99.6% of infield flies result in outs, so each one of those is worth -0.243 runs, just above strikeouts.

Another fascinating thing they do is compare year-to-year correlation between types of batted balls for both pitchers and hitters to determine how much control each has over the different types. It turns out that pitchers appear to have some control over whether a ball is an outfield fly or not, but as to whether the fly ball will become a hit or not is not within their control, except for home runs. Interestingly, neither the batters nor the pitchers have much control over producing line drives, according to the data.

What is this data good for? It essentially takes Voros McCracken's DIPS theory and allows it to be explored for all types of events in a baseball game, not just strikeouts, home runs and walks. The data confirms McCracken's thesis for the most part, but gives us tools for really determining what parts of the game have the most luck in them. There are interesting and fun implications in all of this for ERV scoring that I am just starting to discover.

They also have batted ball data for ballparks too, which seems a very intuitively useful way to assess park impact. However, the article describing this data could be improved by including a chart with the batted ball outcomes for each park, much like the tables in the back of the book for each player and pitcher. The article gives only a summary of the various park data without the same level of detail found in the other articles.

The last section, "Statistics" provides a good, team-by-team collection of stats that is a handy reference. The charts could use some work, most notably color, as they have many different lines that are to be distinguished by different shades of grey that aren't that different.

One other important feature of the book that must be mentioned is the tone. The folks at THT aren't interested in telling you how smart they are and how dumb you are. They aren't part of any self-aggrandizing crusade against the establishment. They just want to write about baseball and look at it in new and interesting ways, and use statistical methods to help that effort. As A.E. Housman said, some like to use statistics the way a drunk uses a lamppost -- for support, not illumination. The Hardball Times Annual 2006 is full of illumination.