Abstracts From The Abstracts
Part Ten: 1986 Baseball Abstract
The Bill James Baseball Abstract in 1986 was the tenth in a series that lasted twelve years. It was probably nine more than what James anticipated when he self-published the first edition in 1977. Who would have thought back then that the Baseball Abstracts would end up selling hundreds of thousands of copies and find their way onto the New York Times bestseller list?
The price of the 1986 Abstract was raised one dollar for the second consecutive year to $8.95. I still have a Crown Books sticker for $6.71 on the front cover of my book, a 25% discount off the publisher's price. The value obviously rests in the record 340 pages of fresh content rather than in the design of the cover that features an amateurish home plate that only a pitcher could love.
James dedicates the book to John and Sue Dewan and thanks, among others, Susan McCarthy ("heckuva good wife"), Dan Okrent ("still appreciated for the role that he played several years ago in bringing this book to the attention of the nation"), Pete Palmer ("esteemed colleague and occasional competitor"), and Craig Wright ("treasured friend and compatriot").
I wanted to thank
In refuting the statement that "baseball is 75% pitching," James points out in the Introduction and Methods section that, "No pitcher allows home runs as often as
"This is true of every significant area of performance, including those things like walks and hit batsmen, which are usually considered to be controlled by the pitcher. And what does that mean? It means that in order to create a working model or simulation of a baseball game, you must allow the hitters to be the dominant, shaping force in the game. And if baseball were 75% pitching, one would not expect that to be true."
In Here We Go Again, James explains "the major principles and methods which are used and/or referred to repeatedly in sabermetrics." He reviews Runs Created, Value Approximation, the Defensive Spectrum, the Brock6 System, Range Factor, Minors to Majors Projection System, and The Favorite Toy.
With respect to runs created, James acknowledges that "the formulas are imperfect. . .there are many factors of a real-world baseball economy which we cannot measure. . .Base-running is one of those. The ability to move baserunners by making outs--that legendary little thing which doesn't show up in the box score--is another one. The ability to create extra runs by producing at key moments, if such an ability exists, is poorly measured, and not really accounted for in the current formulas."
Nonetheless, James believes "there are essentially stable relationships between batting average, home runs, walks, other offensive elements--and runs. The relationship is not random or arbitrary." He shares three conclusions from studies made by Palmer, Paul Johnson, Wright, and himself over the past few decades:
1) The old idea that a high-average hitter is the man who makes an offense go, and that low-average power hitters don't really do much for the team, is nonsense...Power is an extremely important element in the production of runs.
James goes on to explain that a team with a high on-base percentage and slugging percentage "will always do well in runs scored, no matter what else they don't do. They can be slow as the devil, they can be terrible bunters, bad clutch hitters, stupid baserunners and completely inept at hitting behind the runner. They will still score runs."
In explaining that baseball statistics are circular, James emphasizes that "the sum total of measured successes and measured failures" is always .500. As such, he believes it is "of paramount importance to try to understand the meaning of what the player has done in its own context. Missing this essential point, one would wind up with the conclusion that almost everybody who played in 1930 was a great hitter, while almost no one who played in the 1960s was of the same level."
To understand the value of any accomplishment in baseball, we must constantly relate the accomplishment to the context.
James amplifies his point with the following statement: "There are two major variables which define context: time and place. The level of run production has changed dramatically over time, and adjustments must always be made for this when comparing players in different eras. . .The levels of run production also change dramatically from place to place. . .Or, to put it another way, ballparks create gigantic illusions in player statistics."
It is not possible to make accurate evaluations of any player's accomplishments without adjusting for this bias. The effect of the park in which the man plays must constantly be kept in mind when evaluating his accomplishments.
In Baseball's Big Honor, James provides an updated version of his Hall of Fame Monitor, which he originally published in the 1980 Baseball Abstract and re-printed in the 1983 edition.
What we are trying to do here is not to decide who should go into the Hall of Fame. What we are looking at is who will, who is likely to and who is not likely to. . .I constructed this method by an after-the-fact analysis of voting patterns, combined with a limited amount of intuition to cover things that can't be entirely cleared up by the voting. . .The system is intended to help the baseball fan monitor the progress of a player toward the Hall of Fame. It is not intended to say who should or should not be selected.
James introduces Similarity Scores as a means to objectively compare the "degree of resemblance" between two players. "The similarity scores begin with the assumption that players who are all identical in all respects considered will have a similarity score of 1000." Points are subtracted based on the statistical and position differences between players.
James is quick to point out that "uniqueness is one of the fundamental tests of quality," something that I'm not sure most followers of similarity scores fully comprehend. For example,
James suggests several areas in which similarity scores might be useful, including Hall of Fame voting, making minors-to-majors adjustments and career projections, salary negotiations, and evaluating trade proposals. He also believes similarity scores have an even greater potential to "(1) define control groups which have the characteristics of the group under study in all areas except the one being investigated, and (2) construct theoretical models (or 'profiles') and identify real teams which are similar to the model."
Other things being equal, a catcher's career will be shorter than that of a player at any other position. Using this method, we'll be able to measure for the first time exactly how much catching shortens a player's career on the average.
In How is Project Scoresheet Doing?, James answers, "Very well, thank you." Owing to the "Herculean efforts of John and Sue Dewan," accounts of all major league games are now available to the public. He also mentions that STATS (of which James is "a minor shareholder") markets some of the information developed through Project Scoresheet to organized baseball.
In Section II (Team Comments), James devotes 30 pages to his hometown Kansas City Royals, split between "A History of Being a Kansas City Baseball Fan" and "A World Series Notebook." The latter recounts the 1985 World Series between the Royals and the St. Louis Cardinals.
James talks about Gene Mauch and his one-run strategies in the California Angels comments. "Strategies such as the sacrifice bunt and the stolen base are called one-run strategies because they tend to increase the number of times that a team will score one run in an inning, but tend to decrease the number of times that a team will score 3, 4, 5 or more runs in an inning."
The 1985 Chicago White Sox "were a terrific ballclub" in the clutch, begging the question by James if clutch ability exists. "Clutch performance certainly exists, but whether it is a function of ability or random chance is an issue on which there is no definitive or convincing evidence."
James cites a study by Don Zminda regarding the White Sox success in moving a runner from second to third with no outs. He proclaims, "And anybody who thinks you can win baseball games by making outs is probably one of those guys who tries to tell you that you can get rich by remembering to write your underwear off on your taxes." Is Buster Olney a CPA?
James lists the young talent on the Seattle Mariners--
In the New York Yankees segment, James introduces the term "secondary average," which he defines as "the sum of his extra bases on hits, walks and stolen bases, expressed on a per-at bat basis."
Unlike total average, runs produced, estimated runs produced, runs created, base/out percentage, linear weights and runs ad infinitum, secondary average does not attempt to sum up all of a player's offensive contributions; rather, it focuses on the major areas of offensive productivity which are not reflected in the player's batting average.
James describes secondary average as "a summation of the strength of the 'kickers' to the player's primary average." James makes two points worthy of our understanding. "Overall secondary averages are almost identical to overall primary batting averages" and "secondary average is...a better indicator of hitting ability than is batting average."
The relationship of secondary average to batting average didn't begin to approximate one another until the post-World War II years. Since James introduced the concept in 1986, the secondary average for the major leagues (including caught stealing as an offset against stolen bases) has been .257 vs. a batting average of .263. In 2004, the secondary average and batting average were .268 vs. .266, respectively. The spread between the first (
In reviewing the Cardinals, who had the second-most stolen bases (314) since 1912 while leading the National League in batting average and walks, James repeats "the one most universal truth about good offenses is that they get lots of people on base. If there is one thing that separates a good team from a bad team, it is the ability to get runners on base--as well as, defensively, the ability to keep runners off the bases." According to James, the fact that St. Louis had the best OBP in the league--and not base stealing (despite claims to the contrary)--is the reason why the Cardinals led the league in runs scored.
In "The Devil's Theory of Ballpark Effects" within the Chicago Cubs commentary, James speculates that "baseball teams tend to develop those characteristics which are least favored by the park in which they play."
Park illusions create unequal and misplaced pressures upon teams and players, which in the long run yield results which are precisely opposed to the characteristics of the park.
James theorizes further on ballpark effects when discussing the Philadelphia Phillies:
1) The way in which people think about baseball players is essentially formed by their statistics; 2) Those statistics are heavily influenced by the park in which the player performs; 3) The image of the park tends to become confused with the image of the player.
In the Houston Astros segment, James writes about "Late Season Success" and determines that teams that finish strongly have "an unmistakable advantage" the following year, particularly those with records of .500 or better. In "Hot Streaks," James commissioned a study performed by Steven Copley, who found that there was "a moderatley strong inverse relationship between a player's average in his last ten games and his likely performance 'today'...a tactical corollary to Bill's Plexiglass principle."
James wonders why inexperienced managers (such as Jim Davenport of the San Francisco Giants) aren't required to manage their teams "through a thousand or so games of table baseball" before taking the helm if, for no other reason, "just to get a feel for what works and what doesn't." He answers himself, "Because those games are for fans, that's why. We're professionals, you know; we don't have anything to learn from these fans."
In many other professions, simulations are much prized as educational tools; a major airline would never think of sending a pilot up with lives in his hands unless he had pulled a few dozen planes out of simulated crashes. And what is an APBA game, anyway? Why, it is a simulation of a manager's job, nothing more nor less.
In the Introduction to Player Ratings, James, in defending his decision to rank players by a poll of the scorers who participated in Project Scoresheet, writes what some may believe is contradictory to the development of Win Shares nearly 15 years later.
...I've always said that the best evaluation of players is subjective judgment; this is just the first time I have acted in a way that is consistent with what I have written. I've always railed against "great statistics," arguing that it is inappropriate to try to summarize everything a player can do in one number unless or until you can actually measure everything that he does. The problem with formal rating structures is that there are simply too many things that we don't know. To rate players by strictly objective methods, we have to construct a model of the baseball world. The real baseball world is inevitably going to be hundreds of times more complicated than the model that we construct, and therefore we are going to have to a) leave out many factors, factors which are very real and very important even though we can't measure them, and b) make assumptions about things that we don't really know.
Well, I checked and, sure enough, ol' Bill hit the nail squarely on the head once again. After hitting .270 and walking a career-high 113 times in 1985, Harrah's batting average fell to .218 in 1986 with no change in team or home ballpark. He was granted free agency that November and never hooked on with another team again.
"So what you have is: 1) the greatest defensive player ever; 2) at one of the two most important defensive positions; 3) who is also the best hitter in the league at his position; 4) having his best season offensively as well as possibly defensively; 5) holding together a team expected to collapse; 6) and leading them to the league championship. That is about as good a definition of an MVP as one can write--yet Ozzie finished eighteenth in the MVP voting! He was mentioned on only two ballots, placing eighth and ninth on those two.
"I didn't expect that, I don't understand it, I can't justify it, and I don't think it reflects very well on the award or the men who did the voting."
James goes on and says George Brett should have been named the MVP, citing "a much higher offensive winning percentage" than Mattingly or Henderson as well as a Gold Glove and, "with the pennant on the line in the last week of the season...as good a week as any player ever had under those conditions."
In Section IV (Essays, Etc.), James writes about "Age and Performance" and provides all-time all-star teams for each age from 18 to 42 along with single-season, career, and active player records along with the pace of the record holder at each particular age for the categories shown. He also names the Most Valuable Player for each age. I remember being mesmerized by this information when it was revealed as it was something that nobody (as far as I knew) had ever put together. Back then, I felt like I had gotten my money's worth on this feature alone.
Lastly, Bill's wife, Susan McCarthy, writes a two-pager "Looking Backward at Ten" in response to the tenth anniversary of the Baseball Abstract. It is an interesting and revealing history of the Abstract, dating to its primitive beginnings and covering the progress made over the ensuing decade.
Ten down, Susie, and two to go.
Next up: 1987 Baseball Abstract
[Additional reader comments and retorts at Baseball Primer.]
Abstracts From The Abstracts:
1977 Baseball Abstract