Review of The MVP Machine

Ben Lindbergh and Travis Sawchik's The MVP Machine: How Baseball's New Nonconformists Are Using Data to Build Better Players is an excellent, if slightly too long, baseball book. They argue that lots of people have noted the shift toward more granular ways of evaluating players, but few have looked into how data is used to development them, to actually change and shape them.

With the ever-annoying but driven Trevor Bauer as a major case study, they dive into the world of high speed cameras, pitching grips, throwing regimens, launch angles, arm slots, swing planes, and the like, all of which challenge conventional wisdom. As Bauer reasonably notes, he doesn't want any advice that is not based on data (unfortunately, he is a jerk to people who give him advice he believes it bad and in general seems unable to be nice to anyone).

In abundant detail, they make a convincing case for why players who use data improve, while those that don't will be at a disadvantage. Often, the people who can help interpret data aren't former players at all, which is transforming both dugouts and front offices. Scouts are becoming a thing of the past--the human eye just doesn't have the ability that technology does. People who can use and interpret data effectively also get MLB jobs--the website Fangraphs seems to always need writers because their best are continuously snapped up.

In the conclusion, they make two particularly important points. First, baseball is unlike other sports where data-driven improvements make the game more exciting (e.g. longer gold drives or three-point shots). Data decreases the action in a baseball game--fewer ground balls, less base running, more strikeouts. Second, getting data takes money, so kids with poor parents cannot benefit from it and therefore will likely fall behind. This is troubling in our era of year-long sports and club teams. Data has created real problems.

In a book full of examples of how players used data to improve, there is one glaring problem. All of those examples are success stories. In other words, the cases are selected on the dependent variable, thus proving it by definition. They note in passing players that only briefly benefited, but to flesh out this story they need a more detailed discussion of someone who did not benefit. My hunch is that it would bolster their case.