Abstracts From The Abstracts
Part Eleven: 1987 Baseball Abstract
Ten years after self-publishing the first Baseball Abstract, Bill James was still producing "100% new material" with each and every new version of the baseball annual. In fact, the 1987 Bill James Baseball Abstract, in its Oakland A's green and yellow cover, is one of the best books in the 12-year series. It is replete with more essays than any other Abstract.
Despite the fact that the Baseball Abstract was still in its glory, the first visible sign of James reducing the quantity -- as opposed to the quality -- took place in this year's edition. After increasing the number of pages fivefold from 1977 to 1986, the 1987 Abstract saw a 40-page decline, followed by a 65-page drop in its final year in 1988.
Year Pages 1977 68 1978 115 1979 120 1980 200 1981 206 1982 213 1983 238 1984 273 1985 308 1986 340 1987 300 1988 235
James told me over breakfast at the Winter Meetings earlier this month that the reason for the reduction in pages was a direct result of the publisher pushing up the deadline to December 15th in order to beat the growing competition to market. As a result, Bill simply didn't have the time to put together as much information as he once had or would have liked. If he had to do it all over again, James told me that he would produce the book on his schedule rather than theirs.
Speaking of publishers, Ballantine Books also published "The Great American Baseball Stat Book" by Bill James, John Dewan, and Project Scoresheet in April 1987. The editors included Gary Gillette, Craig Wright, and Don Zminda.
There is some overlap in the information of the two books, but there are several distinctions. Let me clarify:
James reproduces a sample page --
In Meaningful and Meaningless Statistics -- a wonderful primer, a must-read for any aspiring sabermetrician -- James suggests that there are four basic criteria by which statistics can be judged:
1. Importance (Significance) - "Does it correlate with winning? . . .Winning games in baseball consists of three parts -- scoring runs, preventing runs, and putting the two together in a desirable combination."
2. Reliability (Integrity) - "What outside influences are there on this accomplishment?
3. Ease of comprehension ( Intelligibility) - "Can the average baseball fan make sense of this information?"
4. Construction (Putting the Elements Together) - The "stat value" score of which James calls a starting pitcher's ERA "the best basic stat in baseball" and on-base and slugging percentage the "two statistics that stand above the others" for hitters.
James admits to getting "letters every week telling (him) why won-lost records of pitchers are meaningless, or why saves or game-winning RBI are meaningless." He explains, "The problem with this is that, when you declare one category to be meaningless, you are left to rely that much more heavily on the ones remaining. And they ain't perfect either."
Now, if you say that won-lost records are meaningless because they depend on who the player plays for and how many runs he scores. . .then you're left judging the pitcher essentially by his ERA -- which is, in fact, also subject to outside influences. When people say that one statistic is meaningless, what they are really saying is that they have learned to see the distortions in that statistic -- but haven't yet learned to see the distortions in the alternatives.
In Beyond the Basics, James elaborates on "new" statistics (dividing them into two classes -- ratings and records), including baserunner errors, quality starts ("Vin Scully's favorite baseball statistic"), total average ("the product of the fertile mind of Thomas Boswell, possibly America's finest baseball writer"), runs produced ("invented by Spiro Agnew. . .Spiro was never too complex"), linear weights ("devised by my friend Peter Palmer"), and on base + slugging ("comes from the work of Peter and Dick Cramer").
I'm not big on on base plus slugging. Its accuracy is not quite that of linear weights, but it is high -- higher than total average. . But the sum of the two isn't quite as easy to make sense of as are linear weights. . .And the statistic is put together wrong. . .They shouldn't be added together, the should be multiplied. A team with a .400 on-base percentage and a .400 slugging percentage would score more runs than a team with a .350 and .450 although both would add up to .800.
In My Own Menagerie, James discusses stats that he has developed himself -- runs created ("my way of evaluating a hitter"), offensive winning percentage (an attempt to place runs created "in the context of the number of outs made"), defensive winning percentage, approximate value, power/speed number, secondary average ("the impact of a player's secondary offensive skills. . .not reflected in his batting average"), defensive efficiency record, isolated power ("the difference between batting average and slugging percentage"), established performance levels, The Favorite Toy ("not a statistic but a process. . .to estimate a player's chance of accomplishing some particular, very difficult goal"), and The Brock6 System (an attempt to "project the normal expectation for the rest of a player's career").
In chapters one through three, James spends 39 pages on "The Greatest Rookie Crops of All Time," "Outstanding Performances by Rookies as Rookies," and "Evaluating a Rookie," followed by a three-page Appendix with definitions and a list of rookies with comparable records at different ages.
The key points are as follows:
Furthermore, as a player ages. . .
As a result, "the further along in this progression the player is (regardless of age), the closer he is to the end of his career (or, conversely, the earlier he is in this progression, the longer he can be expected to play)."
Many players, perhaps most players, are driven out of the major leagues indirectly because they lose their speed. If you can create seven runs a game if doesn't matter how fast you are; you can play first base or DH. But as a player loses speed as he ages, he loses the ability to play the positions (center field, shortstop, second base) at which offensive ability is scarce, and thus loses the ability to stay in the majors without creating seven runs a game.
Of all the various conclusions from James' study, I was surprised to read that K/BB ratio for hitters is "not an indicator of potential growth or development for a rookie."
James also addresses defense in the section on rookies, proclaiming that "there are four basic kinds of defensive value.
* The ability to play a key defensive position, like shortstop, second base, center fielder or catcher, at which talent is always in short supply and where consequently the aggregate offensive performance is less.
* Range, which is the ability to maximize opportunities at the position assigned, and which is measured by range factor.
* Reliability, which is the ability to make the plays a player at the position is expected to make, and which is measured by fielding percentage or errors.
* Specialty skills, such as the ability to turn the double play or a strong throwing arm."
The above words of wisdom obviously apply to all players, rather than just rookies. I, for one, think center fielders unfairly get grouped in with corner outfielders -- especially when voting for All-Stars, MVPs, and Hall of Fame -- and would like to see more of a distinction made among the three OF positions.
In between each of the team comments, James provides a one- or two-page essay on a random topic. He first tackles runs created in which he identifies six factors "which can cause a discrepancy between expected runs. . .and actual runs scored:
1) Exceptionally poor or exceptionally good hitting with runners on base and in scoring position.
2) Baserunning errors or effective baserunning.
3) Exceptionally poor or exceptionally good lineup design.
4) An unusual number of opponent's errors.
5) Doing a particularly good or a particularly poor job of advancing runners with outs.
6) Luck, which would usually be expressed as unusual performance in one of the previous five areas."
James expects to better account for the differences in the future owing to Project Scoresheet but points out that "all of these factors, taken together, do not create huge discrepancies from the current estimates. . .and to the extent that teams do exceed (or fall short of) runs created estimates in one year, they tend to relapse (or improve) in the next year."
On the other hand, James is skeptical of the Cardinals going from 101 wins in 1985 to 79 in 1986 and back to 101 in 1987. (The Redbirds didn't make it all the way back, but they won 95 games, finished in first place in the East, and won the N.L. pennant before losing in the World Series to the Minnesota Twins in seven.) "Maybe, but if a man is thrown from a horse in a half-second, does that mean that he must be able to get back on in another? If you wrap your car around a tree, can you put it back together as quickly? It is a rule of nature that the processes of destruction, such as fire and violence, act more quickly than the processes of growth and development. In the course of a decade there are more teams that improve from season to season than there are that decline, which means that the declines are larger than the improvements."
In "The Fastest Player in Baseball," James attributes a truism that speed is the only thing you can use both on offense and defense to Whitey Herzog. James hypothesizes that "it is probably the only characteristic of a player that you can evaluate by looking at so many different areas of play."
In "Rushing 'Em," James concludes from a study -- despite claims to the contrary -- that the average number of minor-league games played by major-league regulars has been "remarkably consistent" since 1940. If anything, James found that there were actually fewer players rising through the minor leagues quickly then there were a generation ago.
James addresses what is now known as Defense Independent Pitching Stats in "Indicated ERA" [(HRA x TBB x 100)/Innings Pitched²]. The major difference between the two is that James doesn't account for strikeouts in his formula.
There are two elements of a pitcher's record that are independent of the team. Those are his walks and his home runs allowed. Those are the two elements on which, as the announcer says, the defense can't help you; if you don't throw strikes or the ball leaves the park there is nothing
Using James' version, in 2004 the American League had an Indicated ERA of 4.75 vs. an actual ERA of 4.67 and the National League had a 4.65 vs. 4.31.
In "Why Cleveland, Texas and San Francisco Shouldn't Be Expected to Contend in '87," James points out teams which improve by 20 or more games from one year to the next "very rarely" win even more games the following season. In fact, his study shows such teams will decline "at least 80% of the time. He contends the plexiglass principle is most compelling for teams that win a pennant or come very close because management "will be less inclined to identify and address its remaining weaknesses." However, "the tendency to relapse is somewhat weaker if the improvement was based on young players."
Three teams -- Detroit Tigers, St. Louis Cardinals, and San Diego Padres -- improved by 20 or more games in 2004. Based on James' work, the Cardinals are the most likely ballclub to recede in 2005. By the same token, don't expect the Tigers or the Padres to show much, if any, improvement this coming season.
James writes a pointed short essay about baseball and business, entitled "You Don't Say." His conclusion? "Anybody who tells you that baseball is basically a business is either badly confused or a jackass. And you can tell her I said so." Is "her" the operative word in that statement? If so, it's not all that surprising to me. I mean, whoever said Bill was politically correct?
The satirist in James spills out in "Score," in which Bill creates an aptitude test for prospective official scorers "in the American League Championship Series or possibly even the World Series." He asks four questions and gives humorous answers, yet requests that completed examinations be returned to "Urbane Pickering's School of Official Scoring, Valentine Design, and Toilet Traning, P.O. Box E-5, Cotton Balls, Iowa."
In "MVP," James discusses whether a pitcher (
A baseball roster consists of 24 players, usually 9 pitchers and 15 position players. If success in baseball is 37.5% due to pitching, then the average pitcher is exactly as valuable as the average everyday player. For Mattingly to be correct that a pitcher can't be as valuable as an everyday player, one must conclude that baseball is much less than 37.5% pitching. I don't think too many people are going to argue for that.
James runs the numbers for the 1986 season and concludes that "Clemens deserved the award." He also takes a second look at the 1978 A.L. MVP race and switches allegiance from
Ken Phelpses are just available; if you want one, all you have to do is ask. They are players whose real limitations are exaggerated by baseball insiders, players who get stuck with a label -- the lable of their limits, the label of the things they can't do -- while those that they can do are overlooked.
The players are once again rated by approximately 140 scorers from Project Scoresheet. "Polling people who know the players is the surest, best way to incorporate a broad base of knowledge about the players in to the rankings, and thus the most accrurate way that I know of to rate players."
Here is a sampling of James' more biting comments:
I also thought James' comments on
Next up: 1988 Baseball Abstract (the last in a 12-part series)
Abstracts From The Abstracts:
1977 Baseball Abstract
[Reader comments and retorts at Baseball Primer.]