What really sets the Hardball Times Annual apart, and makes it arguably the best of the bunch, is that it focuses on the latter component--insightful articles--far more so than its competitors. There are absolutely stats, which, in fact, are sophisticated, well-organized, and innovative. In addition to classic team and player stats, The Hardball Times makes use of a number of more advanced statistics to analyze both teams and players, including Win Shares, runs created and prevented, defense-independent pitching statistics, and detailed batted-ball statistics. It will be an excellent resource in the coming season, along with the wonderful Hardball Times website.
But the heart of the book lies in the 160 pages of insightful baseball writing by a huge cast of contributors. There is an emphasis in places on the previous season, but much of it is timeless, and will still be worth reading several years down the road. The book also spans a huge range of material, from baseball humor and blunders, to baseball history, to hardcore statistical analysis. In this way, it is more similar to a periodic editorial volume than a classic baseball year-in-review.
The book began with a series of articles recapping the 2006 season. After an introductory article by the always-interesting Dave Studenmund (10 things I learned last year), there was a separate review of the happenings within each division in baseball, as well as a recap of the playoffs. Generally, in each divisional review, each team was recapped for a paragraph or so, with some additional writing to describe how they interacted over the course of the season. The playoff review highlighted each game's major happenings, and did a nice job of summarizing the events as they transpired...though to be honest, I found this whole first section of the book to be the least interesting.
Fortunately, the next section was much better, with eleven short articles of commentary on the 2006 season. Highlights included a dissection how how Walt Jocketty assembled the 2006 St. Louis Cardinals by Brian Tsao, an investigation of the effect of the new drug testing policy on MLB performance by Steve Treder (basically, there wasn't much of a measureable effect), an interesting investigation of competitive balance by Vince Gennaro (he argues that it's only an illusion), and an analysis of the hugely successful Braves General Manager John Schuerholz's approach to his business by Mac Thomason.
From this point on, the articles broadened in scope, focusing less on the 2006 season and more on baseball in general, both in present day and in the past. Here are a few highlights:
- John Brattain describes the initial origins and ultimate downfall of the Federal League. I had only passing familiarity with this league, but it is a remarkable experiment and one wonders what would happen if baseball's anti-trust exemption was dropped and the same thing was tried today.
- There is also a wonderful article by David Gassko on Tommy John Surgery, in which he describes not only the origins and nature of this bizarre surgical procedure, but also its effect on current major league baseball. This "miracle" surgery has not only prolonged the careers of innumerable star-caliber pitchers, but it sometimes might even have improved their abilities.
- David Gassko (he is a major contributer to this volume, along with Studes, and for good reason) attempts to evaluate the most valuable (in terms of total career contributions) pitchers of all time. In doing so, he does something that I hadn't seen anyone do prior to this article: he attempts to correct for the competitiveness of baseball. Gassko correctly points out (as others have before him--it's just this is the first I'd seen it) that baseball today almost certainly must be far more competitive now than it was in "the good old days," because a) training and medicine are far better now than they used to be, and, perhaps more importantly, b) the population from which baseball players can be chosen is far larger now than it was before--especially post-integration. Somehow, I never had realized how important integration was to the quality of competition in today's game before reading this article, and the implications of this for interpreting past performances. Gassko has continued this work recently with hitters at THT.com.
- John Walsh has a fabulous look at outfield arms, going back to the late 1950's, using his outfield throwing metrics. Some notable Reds that pop up as having excellent arms: Cesar Geronimo, Pete Rose, Paul O'Neill, Ken Griffey Jr., Adam Dunn (earlier in his career, apparently--last season he was below average), and Reggie Sanders.
- Dave Studenmund conducted a very interesting study on what he calls Net Win Shares Value. He used a three-step process. First, he identified the typical payroll cost of a marginal win (a win over the 50-win threshold expected of a team with nothing but replacement players). He then looked at each player's salary, estimated the expected win shares above bench of the player given his salary, and then compared these expected win shares to actual win shares above bench. Finally, he converted this difference to a dollar value to determine the players who provided the most bang for the buck teams spent on them.
- Among the Reds, the most valuable players in terms of net win share dollars, were Bronson Arroyo, Aaron Harang, and Dave Ross, while the players who underperformed their salaries the most were Ken Griffey Jr. (by a LOT, due to his high contract and his only modest production), Jason LaRue, and, in third place, Eric Milton.
- Greg Rybarczyk, owner of the fascinating HitTracker website, has a fabulous look at some of the insights that can be gained with more detailed physical data on batted balls. In particular, he looks at how the weather and park effects affected production by both pitchers and batters, and revealed some interesting results. For example, he finds that Arroyo-victim Glendon Rusch was really the victim of insane bad luck last year, with 20 of 21 home runs allowed being hit with a tail wind, pushing them an average of 27 extra feet, and all were at ballparks with high elevations. As a result, he had a disastrous season...but may have actually pitched with skill that would result in a reasonable season: Rybarczyk calculates that Rusch would have allowed only 9 home runs had he pitched under average conditions. Even more striking is his look at how Craig Biggio's home run totals in 2006 were substantially improved by hitting into Minute Maid Park's Crawford boxes in left field.
- Dave Studenmund and David Gassko team up in a pair of articles to take a close look at batted balls: their run values, how splits between GB/FB/LD's vary among different sorts of hitters, and the extent to which players have control over their batted ball rates (on the latter point, pitchers have less control than hitters, though pitchers still do have some control, especially of ground ball rates). It's a fascinating pair of studies, and makes me think that much of our future insights into player performance will come from further investigation of batted ball tendencies.
Before I close, I will make one point of modest criticism. There is a tendency among some of the studies in this volume to go on what are sometimes (at least in biological research, which is what I do in my day job) called "fishing expeditions": studies in which you run a huge number of statistical tests on a huge number of variables in the same dataset and then go through and try to interpret those that are significant. This is a problem that I often see in baseball research, so I don't mean to single anyone out here. But it can lead to--and, I suspect in some cases, has led to--false positives.
The issue is that significance tests report how often you should find the result that is the observed amount different from the expected result (usually zero difference) just by chance. We typically define 0.05% as a cutoff probability for significance. The problem is that if you run 20 of these tests, all on the same dataset, you should expect to get at least one "significant" result just by chance alone (1/20 = 0.05), even when no difference exists. There are various ways of compensating for this issue, either by using different sorts of statistical methodologies, or by doing an adjustment to your significance cutoff (e.g. the Bonferroni method), but as far as I can tell, these were not employed.
Overall, though, this years' THT Annual offers outstanding value--insightful, innovative, and readable analysis, as well as an excellent selection of detailed and sophisticated statistics, all for a very affordable $20. This was the second year in a row I have purchased this publication, and I was again very impressed with it. There is no annual offseason baseball publication that I will be looking forward to more than next years' THT Annual.
You can purchase The Hardball Times 2007 Annual via amazon.com, or in many local bookstores. I bought mine at Borders. :)