Who is a better player — Ryan Howard or Chase Utley? Not who slugs the most home runs (Howard), or steals the most bases (Utley), or drives in the most runs (Howard), or plays the more difficult position (Utley), or scores the most runs (Utley). At the end of the day, taking everything into account, how many more games have the Phillies won because they've had Utley as their second baseman rather than an ordinary second baseman, compared to the number of games they've won because they've had Howard as their first baseman, rather than an ordinary first baseman?
Over the past thirty years or so, a lot of folks have developed answers to this question by converting traditional baseball statistics into estimates of a player's overall net value, in 'runs created' above a baseline level, taking into account all of the elements of the player's offense, the amount of time he plays, and the relative difficulty of the position he plays. These runs created estimates can also be translated into 'wins' created.
But we've left something out, haven't we? How many runs did the player 'save' or 'allow' in the field, relative to other fielders at his position? Yes, we have long had pretty good estimates of how much 'extra credit' a player should get because he happens to play a more difficult position (such as second base) rather than an easier one (first base). But that doesn't tell you how well he plays the position. If you don't know how to answer that question, in terms of a reasonably reliable estimate of runs saved or allowed relative to the baseline fielder at that position, then in any remotely close case of overall player evaluation you don't really know which player was better.
The problem of estimating the value of a player's fielding over the course of a season and career using traditional defensive statistics has been the holy grail baseball statistics problem at least since Branch Rickey declared it unsolvable in a famous Life magazine article in 1954. Precisely because it was still considered unsolvable even two generations later, despite some insightful attempts we'll be discussing throughout this book, private companies were formed to collect new kinds of statistics to answer this question for baseball teams. And those answers are probably very good. But they're based on proprietary data, and are only available for recent players.
A different approach must be taken to evaluate the all-time greats. It's clear that Barry Bonds has been a more valuable player than Ken Griffey, Jr. But if you had asked before the turn of the millennium which player had been more valuable over the first ten years of his career, it would have been a very close question. Going a little further back, who was the greatest player before the offensive explosion in the 1990s — Joe Morgan or Mike Schmidt? Schmidt had bigger numbers, but Morgan played what was then a much tougher position than Schmidt. From the moment Mickey Mantle retired in 1968, everyone knew that Willie Mays had had the greater career. But some have argued very persuasively that at his peak or in his prime Mantle was not just a greater player, but a much greater player than Mays.
Ted Williams was a proud man, and didn't hesitate to say that he was a better hitter than Joe DiMaggio, but always granted that Joe was the better all-around player. Was Joltin' Joe, in fact, a more valuable player, overall, than The Thumper? Ty Cobb won a lot more batting titles than Tris Speaker, but at the time they played, some thought Speaker a better all-around player, and Speaker oft en earned more money than Cobb. Back in the 1890s, Ed Delahanty was the greatest slugger in the game. But was he the best player?
None of the questions posed above can be answered unless we have a method of estimating fielding value, in runs, that is as transparent, logical, and at least close-to-reliable as the methods developed to assess the other components of overall value.
If we could find such a method, it would not only change our rankings of the players throughout baseball history we already 'knew' were among the best; it would bring to our attention for the first time many players who've been forgotten because of their modest batting numbers. And resolve one of the big unknowns of major league history: whether the rare players who have garnered fame primarily through their fielding — Rabbit Maranville, Bill Mazeroski, Brooks Robinson, Omar Vizquel, perhaps a few more — actually contributed all that much to the success of their teams.
To answer these questions, this book presents the first method for quantifying fielding value throughout major league history to which professionals who make their living working with mathematical and statistical models might give their approval. And the results produced by that method, along with a new approach introduced for ranking players who played in different eras, will, I believe, eventually result in the biggest changes seen in thirty years in the assessment of the historical statistical record of our nation's most history- and numbers-obsessed pastime.
As the saying goes, extraordinary claims require extraordinary evidence. So we are going to present a lot of evidence. We'll introduce each concept with relatively simple examples, and then refine things step by step. In addition, sufficiently detailed information is presented in the on-line appendices for anyone interested in replicating, testing, or improving upon the methods introduced here.
The level of detail will, nevertheless, be overwhelming in more than a few places. I can readily imagine how many of you will feel when confronted with some of the formulas and data (and analysis of such formulas and data), because I 'chose' to become an English major in college after a very undistinguished three-semester career in calculus. Although I've since had a reasonably positive mid-life encounter with some fairly advanced mathematics, and have managed to organize all the data and devise all the equations presented in this book, I know exactly what it feels like to confront books packed with math, data, and intricate arguments that seem at first glance, and even after many re-readings, incomprehensible and lacking any clear objective.
I hope I've at least made the objective clear: measuring the season-by-season impact, in runs, of fielders throughout major league history, and finding a way to compare them more fairly across time. When things get a little dense or difficult, do what every mathematician or scientist does when they first come to a part of a technical paper or book that they don't understand: skip ahead to something that recaptures your interest. Browse through the player essays in Part II. Go back and re-read something you felt you more or less understood, but would like to understand better. If and when you feel like it, come back to the parts that seemed a bit hard to follow. I hope and believe they'll be clearer the second or third time around. If not, the fault will be mine.