February 8, 2007

Schrodinger's Bat

The Burgess Shale and Other Weighty Matters

by Dan Fox

the archives are now free. All Baseball Prospectus Premium and Fantasy articles more than a year old are now free as a thank you to the entire Internet for making our work possible. Not a subscriber? Get exclusive content like this delivered hot to your inbox every weekday. Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get instant access to the best baseball content on the web.

Subscribe for $4.95 per month Recurring subscription - cancel anytime. Subscribe for $39.95 for one year a 33% savings over the monthly price! Purchase a $39.95 gift subscription a 33% savings over the monthly price! Already a subscriber? Click here and use the blue login bar to log in.

"Inasmuch as all base ball managers are imitative all clubs now carry at least one pinch-hitter. For this may the Lord forgive [John] McGraw, as there is no doubt that the system has added to the expense of the clubs, the length of the box scores, and the vexation of the scorers; while there is much doubt as to the practical value of the system."

--Sporting Life, August 30, 1913

Before my brief hiatus last week the topic of this column turned back to triples as we explored whether it was the case that the ever-increasing bulk of players could be responsible for the steady decline in three-baggers since around 1930. It turned out that even when normalizing for the distribution of Body Mass Index (BMI) that prevailed in the 1920s when leaner players were more prevalent, the triples rate still declined, albeit slightly less dramatically. Given the other probable causes for the decline in triples, we concluded that in the aggregate the increasing scarcity of triples over time is a residue of the ever-increasing level and standardization of play.

As interesting as all that was, it was the following paragraph that engendered the largest volume of comments from readers.

 

As an aside, I'd be interested to hear reader theories as to why average height and especially weight dipped in the period from 1910 through the 1920s and then picked up where it left off in the 1930s. No systemic cause or data problem jumps out at me, and it doesn't seem likely that the slashing style of play popularized by the Baltimore Orioles of the 1890s would have such a latent effect.

I asked the question because one of the graphs I built indicated that height, and more dramatically weight, decreased for players who debuted in the 1910s and 1920s before resuming a steady upward trend in the 1930s. A more detailed graph below with weight graphed on the Y-axis on the left and height on the Y-axis on the right shows the same basic picture.

 

When looked at in more detail, it's clear that weight decline was more acute than height and in fact average height increased in each of the first three decades of the twentieth century, while weight decreased in the period 1910-1919 before rising slightly from 1920-1929 to just under its level from 1900-1909.

What I found most interesting, perhaps, was the variety of responses articulating theories to explain the trend, the most frequent of which were:

 

 

 

 

• Immigration. Some readers speculated that a spike in eastern and southern European immigration in the early years of the century might be responsible as more players with those backgrounds entered the league. This would be the case, so the theory goes, since the general health and nutrition (and hence height and weight) might be lower for those immigrants than it had been for players born in America.

  Country of birth is available in the database, but calculating the percentage of players born in the US versus those outside by debut decade revealed that by the 1910s 97% of players were born in the United States. That percentage did not begin to shrink until the 1950s (92%), and has since accelerated with the larger influx of Latin and Asian players in the 1980s (87%) and 1990s (78%) to the point we're at now, where in the 2000s the percentage of US-born players was 72%. It is perhaps the case that the data for country of birth is incorrect for players from that period, or that there is a residual generational effect in play, but absent more data there is no hard evidence that country of birth or ethnicity is the cause of the trend.

• The Flu. It occurred to me, as it did to several readers, that the inappropriately named (since it started in the U.S.) Spanish Flu pandemic of 1918 might in some way play a role here since it was paradoxically hardest on otherwise healthy young adults, and killed between 500,000 and 675,000 people in the United States. That large a disruption might affect the population from which baseball players are pulled and lead to a decline in the average weight and height.

  It turns out, however, that while there were certainly flu outbreaks in eastern cities as early as 1916, the trends we see (as revealed in the more granular graph above) began as early as 1910. That's not to say that the flu might not have reinforced the trend in some way, since it obviously had a devastating effect on the entire population, only that it couldn't have been the initial or major cause.

• The Great War. As with the flu, the effect of World War I on the pool from which players are selected was too late for the trend we see. Once again, however, the war clearly had an effect later in the decade of the 1910s and probably into the 1920s and may have helped perpetuate the trend. This is suggested by the fact that in World War II we see decline in height and weight from 1942 through 1945 before both resume their upward march in 1946.
• A Competing League. Although not mentioned by many readers, I thought perhaps the rival Federal League might have had an effect whereby the introduction of eight additional teams might have depressed the size of players available to the American and National Leagues. Once again the timing is not quite right, since the Federal League did not come online until 1914 and the trend had clearly begun before then.

The Young and the Restless

After cogitating on these theories for several days and satisfying myself that none were the underlying cause, I went back and took a second look at the data. Thinking that perhaps there was something wrong with the data itself, I began to look at the players who debuted during the upturn in the trend. What struck me was that the number of players who debuted in a given season rose from 99 in 1907 to 221 in 1912 and 201 in 1913. Since the number of teams remained constant during this period (at 16), this more than doubling the number of rookies in the game aroused my suspicion that this trend and the trend for decreasing height and weight were related.

But first, to confirm that teams made historically liberal use of players during this time, I created the following graph.

 

The average number of players employed per team rose sharply between 1904 (when it was 22.6) to 1912, when it topped out at 36.9. For the next 78 years, only in 1946--in the wake of World War II, as players returned from the service--was there a total number of players used (38.6) that eclipsed that 1912 high. Of course, in recent years teams have increasingly exercised their 40-man rosters to where the average per team has topped 41 since 2003 (excluding 1994 when the strike wiped out September call-ups).

And now we come to the link between the two. In looking closer you find that the average age of players who debuted also fell coincident with the 1908-1919 period as shown in the graph below.

 

In other words, as teams employed more players during this period, they were also selecting players who were slightly younger than players who had debuted in the previous decade, and who would debut in the 1920s and 1930s. Assuming that height and weight were recorded early in a player's career (remember we have only one data point for each player) it would also be the case that they would likely trend lower and especially so for weight, which is precisely what we see. Here, then, we likely have the primary cause. Players who debuted were slightly smaller because they were slightly younger.

Two additional points can be added. First, the impact of the war may also be evident here, although having the opposite effect of World War II, where the average age of a player at his debut increased to more than 25 years from 1943 through 1946 while height and weight decreased. Additionally, the average age dropped precipitously through the 1950s, and bottomed out in 1965, when 117 players debuted who averaged 22.6 years of age. Since then, the average age has increased and stabilized at just under 25 years old since the mid-90s.

The question then becomes, "Just why did so many more players break-in during this time?" I posed the question to Steve Treder (he was after all, a finalist in the trivia competition at last year's SABR convention), who reminded me that Bill James had written on the subject in last year's Hardball Times.

In his article titled "Young Pitchers," James shows that the percentage of innings pitched by young pitchers (age 25 or younger) in the decade 1910-1919 was 42%, the highest of any decade, with the 1960s coming in second at 37%. The percentage then drops sharply to 24% in the 1920s, and 23% in the 1930s. As to why this occurred James wrote that,

 

The pitching patterns of the 1910-1919 era were abnormally simple. At that time the pitcher was allowed to abuse the baseball – scratch it, rub it in the dirt, spit licorice on it – and then pitch with it…You just rear back and throw hard, and the baseball itself will take care of the rest…Pitchers in that era threw fewer breaking pitches and fewer off-speed pitches than ever before or since, and this obviously worked to the advantage of young pitchers. After 1920 the salaries of players grew rapidly, which probably accelerated the trend away from younger pitchers.

But an increased reliance on young pitchers isn't the entire story. As James himself recounts in The New Bill James Historical Baseball Abstract, the 1910-1919 period was also one where managers began to use players for more diverse reasons as well.

For example, as told by Peter Morris in A Game of Inches: The Stories Behind the Innovations that Shaped Baseball Volume 1, The Game on the Field, pinch hitting began in 1905 when Giants manager John McGraw purchased the contract of veteran Sammy Strange who would be called upon in both utility infield and pinch-hitting (in the sense of hitting "in a pinch") roles. The trend caught on, and by 1909 the New York Telegram observed that "Almost all teams in the National and American leagues carry some player these days who is supposed to be able to take his place at bat in an emergency…" As documented in the quote I led off with today, this was the case, and Ham Hyatt for the Pirates and Dode Criss for the Browns and later Moose McCormick with McGraw's Giants became successful pinch-hitters. By 1914 Hyatt had 58 at-bats as a pinch-hitter, whereas James mentions that in 1904 several AL players tied for the league lead in pinch hits with a paltry two.

In part, this trend was driven by the increased demand on pitchers to increasingly specialize at their craft, and as a result become less effective offensive players. The graph of Pitcher vs. Position Player OPS over time presented in a previous column bears this out. At the same time pinch-runners became fashionable, and once again it was McGraw who was the innovator and took it to the limit 72 years before Charlie Finley later would with Herb Washington, when he used Sandy Piez in that role full-time in 1914.

Entangled with the development of pinch hitting, removing starting pitchers and thereby necessarily developing a demand for the first true relief pitchers (such as Sad Sam Jones in 1915 and Dave Danforth in 1917) was an innovation characteristic of the era. As early as 1902 White Sox manager Clark Griffith was reported to be toying with the idea of removing his starter when he showed signs of fatigue after the sixth inning. In January of 1913, The Sporting News reported that, "The pinch hitter is becoming a factor in the big league races, and it may be only a short time until teams will have to carry great one and two-inning pitchers--men that hurl shut-out ball for a couple of rounds." Although it would take almost another 60 years for this concept to be cemented in the game in the form of the closer, the trend can be seen in the following graph that tracks the percentage of appearances that were made in relief by year.

 

The big jump takes place between 1904, when 11.3% of the appearances were in relief, to 1916, when 41.1% were.

And finally, as is widely known, the adoption of extensive platooning following its successful employment by George Stallings as manager of the 1914 Miracle Braves served to increase the number of players teams employed. Platooning reached its heyday by around 1919, and enjoyed a resurgence following World War II in no small part because of the return of Casey Stengel to major league managerial ranks. As a player, Stengel had been platooned by McGraw, as documented by Retrosheet event files from 1922 (incomplete though they were when I downloaded them, but accounting for 89% of Stengel's plate appearances).

 

Casey Stengel, 1922
     PA  AB   H  2B  3B  HR  SO  BB IBB HBP  SH  SF   AVG  OBP  SLG
R   246 216  82   7   9   6  15  19   1   7   1   2  .380 .439 .579
L     6   6   3   0   0   0   0   0   0   0   0   0  .500 .500 .500

James also notes that roster sizes must have become standardized around 1917 in order to control expenses and to limit the number of players who could be controlled by a single team in light of the developing minor leagues. Unfortunately, I can't seem to find any documentation to back that up, and the data on players used per team would tend to support a slightly earlier date around 1914 or 1915.

The Cone Inverted

The flowering of strategies and the increase in the usage of players brings to mind an analogy from the world of paleontology. In 1909 Charles Doolittle Walcott discovered a treasure trove of wonderfully unique fossils preserved in a layer of shale near the town of Field in British Columbia, specimens that would become known simply as the Burgess Shale. While Walcott placed his specimens in familiar phyla that were known to exist during the period (Middle Cambrian, 505 million years ago), it was a reinvestigation by Harry Blackmore Whittington, Derek Briggs, and Simon Conway Morris of the University of Cambridge in the 1980s that upended that traditional interpretation of the fossils' place in the evolution of life. By inverting the familiar iconography of the cone of increasing diversity in life forms, Whittington, Briggs, and Morris reinterpreted the Burgess Shale as replete with creatures in phyla that are now extinct. In other words, rather than life becoming increasingly more diverse in terms of its basic body plans over successive geologic periods, the Burgess Shale records an initial flowering of experimentation in structures just after the dawn of life before a later decimation or winnowing into the few surviving phyla we see today. Stephen Jay Gould devoted an entire book to this theme as an illustration akin to his theory of punctuated equilibrium in his 1989 book Wonderful Life: The Burgess Shale and the Nature of History.

This pattern from the history of life is akin to what we see in the history of baseball as ideas (usually after being initially successful) spread throughout the game and are either adapted into its fabric, standardized, and carried forward or are discarded. The period of 1910-1919 was one particularly active period of experimentation. I've discussed an example on this general theme with relation to batting styles from an 1888 article published in Birmingham Alabama.

The analogy fails, however, in one crucial respect. In the history of life, once a phyla becomes extinct it cannot be recovered since major body plans, as expressed through genes, lose their flexibility for major reorganization. That's fortunately not the case in baseball, where although some experiments are discarded forever (courtesy runners, "plugging" baserunners, legal doctoring of baseballs), styles of play that were once innovations can come back in vogue. Platooning in the 1950s and base stealing in the 1970s and 1980s are two examples, and some of us are hopeful for the return of the four-man rotation as another. Even so, the overall trend is towards standardization, as more effective strategies and techniques are adopted, resulting in an ever-increasing level of play. In that sense we can look at something as innocuous as a blip in the average weight of rookies as recording the continual evolution of the game.