The Fielding Bible

Editor’s Note (2/15/12): This review was written for the First Fielding Bible. The newest version, Volume 3 of the Fielding Bible, is now available at ACTA Sports.

The Fielding Bible is the first attempt to bring cutting edge fielding analysis to the mainstream. The book is the work of Baseball Info Solutions president John Dewan, with a helping hand from Bill James. On the cover of The Fielding Bible, the book is described as a “breakthrough analysis of Major League Baseball defense.” I wouldn’t go that far (and I’ll get back to that later), but The Fielding Bible is certainly one of the most interesting books of the year.

The whole book spans a little over 240 pages; only about 10% of that is reserved for essays, most of which are written by James. Perhaps one of the more interesting essays regarding baseball fielding that I have ever read is James’ comparison of Derek Jeter and Adam Everett. He goes in-depth looking at the differences between the two players, and systematically proves that Jeter is in fact “a below average defensive shortstop” even if we give him “every possible break on the unknowns.”

This essay embodies much of what is right with this book. Dewan is open-minded, and freely presents loads of data to the reader. There are charts of where hits fell against each team both at home and on the road so that we can see where each club’s defensive strengths and weaknesses were. Each player’s defensive “plus/minus” is broken down by home/away numbers and performances behind lefties and righties.

Dewan even devotes separate sections to analyzing how fielders perform on bunts (Mike Lowell is the best; Alex Rodriguez the worst), to analyzing performance in double play situations (Jose Castillo is spectacular; Todd Walker has hands made of solid granite), and another to rating outfielders’ throwing arms.

The amount of information included in The Fielding Bible is awesome, and will lend itself to further in-depth fielding analysis. At the heart of the book is Dewan’s plus/minus system, which is kind of like David Pinto’s Probabilistic Model of Range.

For each year, Dewan finds the probability of fielding a ball based on where it went (distance, direction), the batted ball type (fly ball, line drive, etc.), and whether the ball was hit hard, medium, or soft. So if on a hard ground ball to Vector 17, the shortstop only has a 10% chance of making a play, and he does, then he is credit with 1 -.1 = .9 plays above average. If he does not make the play, he is credited with 0 – .1 = -.1 plays above average. Dewan runs this analysis on every ball put into play in each of the past three seasons, and adds up the results for every player in the major leagues. The resulting rating is the player’s plus/minus.

Such a method was first employed in constructing Ultimate Zone Ratings, made famous by Mitchel Lichtman (also known as “MGL”). The problem is, The Fielding Bible never acknowledges this. On one hand, I understand that Dewan is the first person to present such data to a wide audience. While UZR has been mentioned in the New York Times and on, Lichtman has never tried to disseminate his results to very many people. Dewan deserves credit for that. But in the sabermetric community, it seems that some have decided to do away with giving credit where credit is due, and that irks me. Not only does it lead to re-inventing the wheel in many cases, but it leads to people wanting to conceal their ideas in fear of being ripped off. Dewan’s not at fault for that, but it still would have been nice if he mentioned those who pioneered play-by-play fielding analysis somewhere in the book. Nice, but not necessary.

On the other hand, such unawareness (or disregard) of fielding analysis leads to some errors in the book. For example, both Dewan and James state that every “extra” play made is worth around half-a-run. That’s just not true: In fact, an extra play is worth around three-quarters of a run. (See an explanation for why at the end of this review.) Now this may not be too big a deal, but it will cause casual baseball fans who are seeing a play-by-play fielding system for the first time to underestimate the true impact of defense.

In the book’s longest essay, James reveals his new non-play-by-play system, Relative Range Factor. It is similar to the many other non-PBP systems; in my mind, too similar. It’s almost exactly the same thing as Clay Davenport’s system on Baseball Prospectus. Maybe James isn’t aware of Davenport’s fielding system—I don’t know. But this goes back to my earlier point of re-inventing the wheel. The system, frankly, adds little or nothing to the state of fielding analysis. If BP was more open about its system, James probably wouldn’t have come up with Relative Range Factor. He could have worked on something more interesting. Instead, this essay didn’t really do anything for me.

Minor criticisms aside, The Fielding Bible was a great book, and highly recommended. While the fielding system Dewan uses could have been improved with more detailed information, and be converted to runs, the strength of the book is not the final results. It’s that Dewan tells us how he got there. It’s that it allows the reader to analyze every player’s defensive abilities in detail. It’s that we see the why to the ratings, rather than just getting a number.

The player comments are spectacular, and the analysis contained within them is awesome. I know I’m going to pick up The Fielding Bible many times over the next season, whether it’s because I want to know how a player acquisition might affect a team, or because I’m arguing about a player’s strengths, or if I just want to look at some cool charts.

References & Resources
This explanation is regularly given by baseball analyst Tangotiger:

Suppose a team with Ozzie at shortstop gives up on average 12 non-home run hits, and 2.6 walks every game (which of course is 27 outs) . Applying .50 runs per non-home run hit (I know it should be closer to .55, but I just want to keep it basic), and .30 runs per walk, and -.10 runs per out, and we get 4.08 runs scored per game. And per game, we see that Ozzie’s team faces 41.6 batters (again, let’s not worry about double plays, etc).

Now, let’s say Ozzie was traded for Spike, and let’s say for every 41.6 batters faced, there is one ball that Ozzie gets to that Spike doesn’t. So, for those 41.6 batters, Spike’s team records 13 non-home run hits (1 more than Oz), 2.6 walks, and 26 outs (1 less than Oz). However, there’s still one more out to go! Since Spike’s team gives up 13 non-home run hits / 26 outs, we can estimate that this team will give up 13.5 non-home run hits, 2.7 walks, and 27 outs per game ( a total of 43.2 batters, a remarkable 1.6 MORE batters than Oz). Anyway, applying our linear weights constants, and we see that Spike’s team gives up 4.86 runs per game.

This number is .78 runs MORE than Ozzie. This is the result of Ozzie getting to one more hit than Spike. .50 runs for the hit, and about .30 runs for the out gives you the .80 runs.

Using Recurrent Neural Networks to Predict Player Performance
Technology is rapidly advancing possibilities in decision-making.

Print This Post

Comments are closed.