I.
I have a couple of research questions for you, anybody knows how to do these.
First, has anybody studied the willingness of managers to break the “rule” against intentionally putting the potential winning run on base? It seems to me. . . .just my impression. . .it seems to me that managers now are much, more willing to issue an intentional walk putting the potential winning run on base than they have ever been before. Since intentional walks are pretty easy to spot in the accounts of the game, it seems to me that this should be something we could study.
Among the things I would like to know are:
1) How frequent is this occurrence now, as opposed to 40 years ago? (I will guess that it is 200% more common now than it was in 1970.)
2) Do we have records of how many times each manager has done this?
3) How often does it backfire, “backfiring” defined as “runner who was put on base scores the go-ahead run.”
II.
The All-Important “Loss” Column
This is the time of year at which teams which are not in first place start hearing about the all-important loss column. The Red Sox are 6½ games back, we hear, but 7 back in the all-important.
Well. . .what makes the Loss Column more important than the Win Column? Wouldn’t it make as much sense to talk about the all-important Win Column?
There is a difference, which is this. Teams which are in a pennant race win more games than they lose. Let us suppose that, with a week to go, the standings are these:
Team
|
W-L
|
Pct.
|
GB
|
Indianapolis
|
93-65
|
.589
|
…
|
Portland
|
90-67
|
.573
|
2 ½
|
New Orleans
|
91-68
|
.572
|
2 ½
|
OK, in that situation, I grant you that the team which is 90-67 is in much better shape than the team which is 91-68. In this scenario, in fact, if we assume
a) that each team has a 60% chance to win each remaining game,
b) that they don’t play one another, and
c) that in the event of a tie, all the teams are equal
Then Indianapolis’ chance of winning this race is 92.4%, Portland’s is 5.5%, and New Orleans’ is 2.1%. Portland has a very meaningful advantage from having two more games left on the schedule.
Suppose, however, that the standings on August 1 are these:
Team
|
W-L
|
Pct.
|
GB
|
Indianapolis
|
63-39
|
.618
|
…
|
Portland
|
60-41
|
.594
|
2 ½
|
New Orleans
|
62-43
|
.586
|
2 ½
|
New Orleans is four games behind in the loss column; Portland, only two. But in that situation, wouldn’t you rather be New Orleans than Portland? Portland has to play four more games, with the same number of days left on the schedule. That means that New Orleans has four more days off in August and September. Given that you’re in roughly the same position, wouldn’t you rather have the extra days off than the possibility of winning an extra game because you haven’t played it yet?
Is there any research on this? Has anybody ever studied the effect of days off during a pennant race? How exactly would you do that?
Your research is appreciated.
Part III—Ozzie’s Latest Rant
Our beloved half-crazy manager, Ozzie Guillen, has burst forth with some strongly-worded thoughts about the disparate treatment of Asian and Latin baseball players, and, as I write this, MLB is busy defending itself from charges of unequal treatment of these groups.
Ozzie’s only problem is that he tells the truth. It’s a bad habit, get you in a lot of trouble, but. . .he can’t seem to break it.
Look, we all want to believe that we are doing our best to accommodate all of our players, and to deal fairly with all of them. At the same time, just because we think we are trying to do that doesn’t mean that we are succeeding.
When you see something that you think is wrong, you think is an injustice, what should you do about that? Should you keep your mouth shut, or should you speak out against it?
I’m not saying that Ozzie is right, and here’s one element of the problem that I haven’t heard comment on. The difference between English and Spanish is not at all like the difference between English and Japanese. A person who speaks Spanish as a native language can learn to speak English with a reasonable effort, because many of the words are nearly the same and the structure of the language is similar. This is not true of English and Japanese.
But Ozzie saw something that he thought was wrong, and he spoke out against it. Good for him. We shouldn’t criticize him for that; we should listen to him, and hear what he has to say.
Part IV—Age Deviation as an Indicator of Quality
The sum total of successes and failures, as measured in the statistics, is the same in all leagues. 99.9% of sabermetrics is based on circular measurements, assuming that whatever is a success for one team is a failure for the other. Sabermetrics—except in the recent era, when we have inter-league play with which to calculate adjustments—sabermetrics implicitly assumes that all leagues are created equal, that a player from 1957 is facing the same level of competition as a player from 2007. The only problem is, it isn’t true. It’s what we might call a necessary assumption, and it may be a workable assumption the majority of the time, but we know that it’s untrue on some level—we know, indeed, that it is always untrue on some level, since the quality of competition in two different leagues can never be precisely the same.
It bothers me to operate on what I know to be a false assumption, and so at some point I began worrying about this in the back of my head. In 2002, while I was coaching third base for a team of eight-year-old boys with the one requisite eight-year-old girl who plays on all such teams, I had a conceptual breakthrough on this issue. “It may be true that the sum total of successes and failures is the same in the statistics of every league,” I realized, “but that nonetheless the statistics of high-quality leagues could be systematically different than the statistics of lower-quality leagues. Thus, it may well be true that the statistics of a league do indicate or even measure the quality of the competition within the league.”
Once I had made that breakthrough, I began to ask the question “In what ways are the statistics of high-quality leagues different than the statistics of lower-quality leagues?” In what ways, in other words, do the statistics of a league indicate the quality of the league? And once I had begun to ask that question, I found more than 40 ways that the statistical summary of a good league is systematically different from the statistical summary of a weaker league.
It was a big project for me; I worked on it for several weeks on end. Just as I was about ready to write up what I had found, however, the Boston Red Sox called. I put what I had found on the back burner, and went to work on Red Sox stuff, and I never did get around to writing up what I had learned.
OK, here’s one way in which the statistics of good leagues are systematically different from the statistics of weaker leagues.
Age.
The more players in the league who are ages 25 to 30, generally speaking, the higher the quality of the competition within the league. The further away from age 27 the players in the league are, the lower the quality of competition.
Think about it. During World War II, the quality of competition obviously declined. What happened to the age spectrum? Baseball in World War II was often described as Old Men, Teenagers, and the physically unfit.
If you compare college ball to high school baseball, which has players who are nearer in age to 27? Which has better quality competition?
If you compare college baseball to Triple-A baseball, which has more players who are nearer in age to 27?
If you compare major league baseball to minor league baseball, which has more?
If you compare high school baseball to little league baseball, which has more?
If you compare a first-place team to a last-place team, which is more likely to have a 19-year-old in center field and a 38-year-old at first base?
In all of these cases, the team which competes at a higher level is likely to have more players who are nearer in age to 27.
We can use this fact, then, as one element among many to make a measurement of the level of competition within a league. The first thing we need to do is to establish a method. Here’s the method I came up with.
1) If a player is less than 27 years old, subtract his age from 27.
2) Square that.
3) Multiply that by his plate appearances. The result is called the “Age Discrepancy Contribution”, or the ADC for short.
4) If a player is more than 27 years old, subtract 27 from his age.
5) Divide by two.
6) Square that.
7) Multiply by his plate appearances to get the ADC.
8) Find the team or league total of the ADC.
9) Divide by the team or league total plate appearances.
10) Take the square root of that.
11) Repeat steps 1-10 for pitchers, but replacing “plate appearances” with “innings pitched”.
12) Multiply the “batters result” by two and the “pitchers result” by one, and divide the total by three.
In other words, age 26 is equivalent to 29, 25 is equivalent to 31, 24 is equivalent to 33, 20 is equivalent to 41, 18 is equivalent to 45, and 15 would be equivalent to 51.
The result is called the “Age Deviation Score”, and the Age Deviation Score is an inverse indicator of the quality of play within a league.
Let’s walk through the data for the 1976 Cincinnati Reds, a rather good team, and the 2003 Detroit Tigers, who lost 119 games. These are the Plate Appearances and Ages for the members of each team:
Player
|
PA
|
AGE
|
|
|
Player
|
PA
|
AGE
|
1976 Cincinnati
|
Reds
|
|
|
|
2003 Detroit Tigers
|
|
|
Pete Rose
|
759
|
35
|
|
|
Dmitri Young
|
635
|
29
|
Dave Concepcion
|
636
|
28
|
|
|
Bobby Higginson
|
538
|
32
|
Ken Griffey Sr.
|
628
|
26
|
|
|
Carlos Pena
|
516
|
25
|
George Foster
|
627
|
27
|
|
|
Ramon Santiago
|
507
|
23
|
Joe Morgan
|
599
|
32
|
|
|
Craig Monroe
|
458
|
26
|
Tony Perez
|
586
|
34
|
|
|
Alex Sanchez
|
423
|
26
|
Cesar Geronimo
|
555
|
28
|
|
|
Shane Halter
|
393
|
33
|
Johnny Bench
|
552
|
28
|
|
|
Warren Morris
|
377
|
29
|
Dan Driessen
|
268
|
24
|
|
|
Brandon Inge
|
366
|
26
|
Doug Flynn
|
235
|
25
|
|
|
Eric Munson
|
357
|
25
|
Bill Plummer
|
168
|
29
|
|
|
Kevin Witt
|
289
|
27
|
Mike Lum
|
164
|
30
|
|
|
Omar Infante
|
244
|
21
|
Bob Bailey
|
141
|
33
|
|
|
Andres Torres
|
185
|
25
|
Ed Armbrister
|
90
|
27
|
|
|
Matt Walbeck
|
144
|
33
|
Gary Nolan
|
88
|
28
|
|
|
Gene Kingsale
|
140
|
26
|
Pat Zachry
|
77
|
24
|
|
|
Ben Petrick
|
129
|
26
|
Jack Billingham
|
70
|
33
|
|
|
Dean Palmer
|
98
|
34
|
Joel Youngblood
|
60
|
24
|
|
|
A.J. Hinch
|
82
|
29
|
Fred Norman
|
59
|
33
|
|
|
Danny Klassen
|
78
|
27
|
Don Gullett
|
51
|
25
|
|
|
Craig Paquette
|
33
|
34
|
Santo Alcala
|
50
|
23
|
|
|
Cody Ross
|
22
|
22
|
Pedro Borbon
|
20
|
29
|
|
|
Hiram Bocachica
|
22
|
27
|
Rawly Eastwick
|
19
|
25
|
|
|
Ernie Young
|
15
|
33
|
Pat Darcy
|
14
|
26
|
|
|
Nate Cornejo
|
5
|
23
|
Will McEnaney
|
9
|
24
|
|
|
Mike Maroth
|
4
|
25
|
Manny Sarmiento
|
7
|
20
|
|
|
Adam Bernero
|
4
|
26
|
Don Werner
|
5
|
23
|
|
|
Matt Roney
|
2
|
23
|
Rich Hinton
|
1
|
29
|
|
|
Jeremy Bonderman
|
2
|
20
|
|
|
|
|
|
Steve Avery
|
1
|
33
|
|
|
|
|
|
Gary Knotts
|
1
|
26
|
If you follow the process above, it works out to an Age Deviation Score of 2.26 for Cincinnati, 2.10 for Detroit.
Fooled you, didn’t I? But this is the data for the pitchers on the two teams:
Pitcher
|
IP
|
AGE
|
|
|
|
Pitcher
|
IP
|
AGE
|
Fred Norman
|
180.3
|
33
|
|
|
|
Steve Sparks
|
89.7
|
37
|
Jack Billingham
|
177.0
|
33
|
|
|
|
Steve Avery
|
16.0
|
33
|
Pedro Borbon
|
121.0
|
29
|
|
|
|
Danny Patterson
|
17.7
|
32
|
Rich Hinton
|
18.0
|
29
|
|
|
|
Jamie Walker
|
65.0
|
31
|
Joe Henderson
|
11.0
|
29
|
|
|
|
Brian Schmack
|
13.0
|
29
|
Gary Nolan
|
239.0
|
28
|
|
|
|
Adam Bernero
|
100.7
|
26
|
Pat Darcy
|
39.0
|
26
|
|
|
|
Gary Knotts
|
95.3
|
26
|
Don Gullett
|
126.0
|
25
|
|
|
|
Chris Spurling
|
77.0
|
26
|
Rawly Eastwick
|
107.7
|
25
|
|
|
|
Fernando Rodney
|
29.7
|
26
|
Pat Zachry
|
204.0
|
24
|
|
|
|
Matt Anderson
|
23.3
|
26
|
Will McEnaney
|
72.0
|
24
|
|
|
|
Eric Eckenstahler
|
15.7
|
26
|
Santo Alcala
|
132.0
|
23
|
|
|
|
Mike Maroth
|
193.3
|
25
|
Manny Sarmiento
|
43.7
|
20
|
|
|
|
Nate Robertson
|
44.7
|
25
|
|
|
|
|
|
|
Chris Mears
|
41.3
|
25
|
|
|
|
|
|
|
Nate Cornejo
|
194.7
|
23
|
|
|
|
|
|
|
Matt Roney
|
100.7
|
23
|
|
|
|
|
|
|
Franklyn German
|
44.7
|
23
|
|
|
|
|
|
|
Shane Loux
|
30.3
|
23
|
|
|
|
|
|
|
Wil Ledezma
|
84.0
|
22
|
|
|
|
|
|
|
Jeremy Bonderman
|
162.0
|
20
|
When you do the same for the pitchers, you get 2.75 for the Reds, and 3.75 for the Tigers. Combining the two scores in a 2-to-1 ratio, then, we get an overall Age Deviation Score of 2.42 for the Big Red Machine, and 2.65 for the 2003 Tigers.
The score for a good team will usually be lower than the score for a bad team; not always, usually. Sometimes things which are usually true can be added together to reach conclusions which are clearly true, although sometimes they cannot. Let’s look now at the scores over time.
When major league baseball “began” in 1876 (it had begun earlier than that, and it did not truly become major league baseball until later than that, but we have to start somewhere). . .when major league baseball began in 1876, the Age Deviation Scores for MLB were in the range of 3.5 and higher:
1876
|
1877
|
1878
|
1879
|
3.57
|
3.42
|
4.31
|
3.85
|
You may wonder at the sudden surge upward in 1878. “Major League Baseball” in 1878 consisted of only six teams, therefore the data is volatile. A handful of personnel decisions impact the data for the league. Monte Ward pitched 334 innings that year; he was 18. The Only Nolan pitched 342 innings; he was 20. These things move the league average upward in that environment.
Anyway, over the next few years, as the league stabilized, we would expect the Age Deviation Score to drop, suggesting an improvement in the quality of play, and it did:
1880
|
4.12
|
1881
|
3.43
|
1882
|
3.31
|
1883
|
3.13
|
It dropped sharply. The Age Deviation Score dropped from 4.12 in 1880 to 3.13 just three years later. The 4.31 figure for the National League in 1878 remains the highest ADS on record for a league represented in the history books as major league.
In 1884, however, baseball expanded very, very rapidly, causing the Age Deviation Score to spike upward:
1880
|
4.12
|
1881
|
3.43
|
1882
|
3.31
|
1883
|
3.13
|
1884
|
3.79
|
In 1884 there were three leagues. The Age Deviation Score for the National League was 3.57, for the American Association 3.70, and for the Union Association 4.16—consistent with the general perception that the National League was the strongest of these leagues, and the Union Association the weakest.
The UA folded after the 1884 season, and from 1885 to 1889 the Age Deviation Scores drifted lower, consistent with a fairly rapid improvement in the quality of play:
1884
|
3.79
|
1885
|
3.46
|
1886
|
3.39
|
1887
|
3.48
|
1888
|
3.45
|
1889
|
3.17
|
Throughout all of this era, the Age Deviation Scores were lower in the National League than in the American Association, as we would expect that they would be.
In 1890 there was a baseball war. The players formed a union and started their own league, the Players’ League. Most of the best players went into the Players’ League, but the other leagues signed replacements and carried on. As we would expect, this sent the Age Deviation Score somewhat higher:
1884
|
3.79
|
1885
|
3.46
|
1886
|
3.39
|
1887
|
3.48
|
1888
|
3.45
|
1889
|
3.17
|
1890
|
3.43
|
The scores for the three leagues were 3.62 for the American Association, 3.79 for the National League, and 3.10 for the Player’s League, suggesting that, for that one season, the National League may actually have been the weakest of the three leagues. Again, this is not a surprise to people who are knowledgeable about that season.
After the 1890 season the Player’s League folded, and after the 1891 season the other two leagues consolidated. Thus, baseball dropped from 24 teams to 12 teams in two years. As we would expect, this led to a sharp decrease in the Age Deviation Score:
1890
|
3.43
|
1891
|
3.27
|
1892
|
3.03
|
The 3.03 figure for 1892 was a new low at that time, but the figure continued to drop lower throughout the decade, indicating that the game was continuing to mature, continuing to weed out the teen-agers and the old men who had been common in the game through the 1880s.
1890
|
3.43
|
1891
|
3.27
|
1892
|
3.03
|
1893
|
2.99
|
1894
|
2.84
|
1895
|
2.87
|
1896
|
2.87
|
1897
|
2.87
|
1898
|
2.76
|
1899
|
2.72
|
In 1900 the National League eliminated four teams, dropping major league baseball to one highly competitive 8-team league. The Age Deviation Score dropped to 2.62.
In 1901, however, the American League started up, disrupting the established order again, and driving the average higher:
1899
|
2.72
|
1900
|
2.62
|
1901
|
2.79
|
To this point the data has consistently behaved as we would have expected it to behave, with the score increasing whenever new leagues are added, and decreasing in almost all other years.
After 1901, however, this is not always true. Baseball in the first decade of the 20th century was much more successful financially than it had been in the 1890s. The players made more money; the game continued to get organized and continued to thrive. We might expect, in these circumstances, that the Age Deviation Score would continue to go down, but in fact it did not. After dropping to a record-low 2.56 in 1903, the deviation score began then to go up:
1900
|
2.62
|
1901
|
2.79
|
1902
|
2.64
|
1903
|
2.56
|
1904
|
2.70
|
1905
|
2.69
|
1906
|
2.87
|
1907
|
2.94
|
1908
|
2.90
|
1909
|
3.05
|
1910
|
2.96
|
The deviation score reached to around 3.00 in 1909, and flattened out at that level for several years after that. This is one of three respects in which the score in that era does not behave as we might expect. The other two are:
1) That although the American League totally dominated the World Series of the 1910-1918 era, and the American League clearly had more stars, the deviation scores in the American League were consistently higher than in the National League,
2) That one would expect the addition of the Federal League to drive the scores up, but this did not happen. Instead, the scores remained level or dropped slightly, and the scores for the Federal League itself were actually lower than the scores for the established leagues:
1910
|
2.96
|
1911
|
2.89
|
1912
|
3.02
|
1913
|
3.03
|
1914
|
2.94
|
1915
|
2.87
|
1916
|
2.88
|
And these are the scores of the three leagues in those years:
|
American
|
|
|
National
|
|
|
Federal
|
1910
|
3.08
|
|
1910
|
2.85
|
|
|
|
1911
|
3.09
|
|
1911
|
2.68
|
|
|
|
1912
|
3.24
|
|
1912
|
2.82
|
|
|
|
1913
|
3.29
|
|
1913
|
2.79
|
|
|
|
1914
|
3.23
|
|
1914
|
2.91
|
|
1914
|
2.68
|
1915
|
3.02
|
|
1915
|
2.88
|
|
1915
|
2.73
|
1916
|
2.94
|
|
1916
|
2.82
|
|
|
|
It was not until 1918 that the American League scores caught up to the National League.
I am not arguing that the Federal League was equal in quality to the National League or the American League. I think it’s pretty clear that it was not. However, I will alert you that, if we get the chance to look at other internal indicators of league quality, we will see this again; the Federal League will again, at other times, appear to be on an equal footing with the other two leagues.
Anyway, after 1916—after the Federal League folded—the age deviation scores contracted rapidly:
1916
|
2.88
|
1917
|
2.62
|
1918
|
2.66
|
1919
|
2.47
|
1920
|
2.49
|
1921
|
2.51
|
And then swung back upward in the Babe Ruth era:
1920
|
2.49
|
1921
|
2.51
|
1922
|
2.67
|
1923
|
2.70
|
1924
|
2.73
|
1925
|
2.74
|
1926
|
2.72
|
1927
|
2.89
|
1928
|
2.98
|
1929
|
2.89
|
1930
|
2.78
|
This is not entirely un-expected. Baseball salaries increased rapidly in the 1920s. When salaries increase, one of the things that happens is that older players stay in the game longer. A player has been making $3,000 a year to play baseball, and the average salary goes to $7,000, that player is strongly motivated to hang on for another year. This causes the age deviation score to increase, which is generally—but not universally—indicative of a decline in the quality of play.
In any case these numbers worked their way back downward through the 1930s:
1930
|
2.78
|
1931
|
2.78
|
1932
|
2.79
|
1933
|
2.69
|
1934
|
2.77
|
1935
|
2.71
|
1936
|
2.67
|
1937
|
2.66
|
1938
|
2.73
|
Through this era, as you would probably expect, the numbers in the American League tended to be a little bit lower than the numbers in the National League. The American League was probably the stronger league. The numbers went up briefly over the next few years:
1939
|
2.86
|
1940
|
2.99
|
1941
|
2.93
|
1942
|
2.89
|
1943
|
2.71
|
We come then to World War II. Since we all know that the quality of play went backward during World War II, we would expect the Age Deviation Score to have increased substantially—and in fact it did:
1940
|
2.99
|
1941
|
2.93
|
1942
|
2.89
|
1943
|
2.71
|
1944
|
2.94
|
1945
|
3.15
|
The 3.15 figure in 1945 was the highest of any season since 1891, indicating a substantial move backward. However, it should also be noted that the 1943-1944 figures are not remarkable; it’s really only 1945 that has an out-of-line Age Deviation Score. The 1943-1944 figures are consistent with the rest of baseball history in that era. In other words, there really were no more teenagers and old men in baseball in 1943 than in 1933 or 1923. There were more in 1945, yes, but only in 1944 and 1945.
Anyway, when the “real” players returned in 1946 the Age Deviation Score dropped sharply:
1945
|
3.15
|
1946
|
2.65
|
1947
|
2.68
|
But then went back up in 1948-1949, when the young players who entered baseball after the War began to reach the majors:
1945
|
3.15
|
1946
|
2.65
|
1947
|
2.68
|
1948
|
2.82
|
1949
|
2.80
|
1950
|
2.76
|
The numbers were stable throughout the 1950s:
1950
|
2.76
|
1951
|
2.69
|
1952
|
2.85
|
1953
|
2.74
|
1954
|
2.71
|
1955
|
2.75
|
1956
|
2.81
|
1957
|
2.79
|
1958
|
2.84
|
1959
|
2.80
|
1960
|
2.84
|
We come, then, to the expansions of the 1960s. For any method designed to measure the quality of play in the major leagues, this is a critical juncture, as people disagree strongly about the effects of expansion. Some people—like me—believe that the effects of expansion on the quality of play were transitory; other people believe they were more or less permanent. This is not a system to evaluate the overall quality of play, but it could be an element of such a system. We thus need to look carefully at what is happening.
The American League expanded from 8 teams to 10 in 1961; the National League did the same in 1962. As it happens, neither league saw an immediate movement in the Age Deviation Score. The American League ADS went from 2.73 in 1960 to 2.80 in 1961; the NL went from 2.99 in 1961 to 2.93 in 1962.
However, there was an expansion effect. It was just delayed. What happens in an expansion is this. The immediate effect of the expansion is to let into the major leagues a number of career minor leaguers, players who have been waiting for a chance but who aren’t quite good enough to force their way in, and these players tend to be closely bunched around 27 years of age. What happens a year or two later, though, is that these first expansion players fail, for the most part, and many of them are replaced by very young players or by older players who have been released by other teams. The Houston Colt .45s in 1962, their first season, had no teenagers on their roster. In 1963, however, they had eight teen-agers on their roster at one point or another—one of whom (Rusty Staub) led the team in games played.
In 1965, then, the Age Deviation Score reached up to 2.98—the highest figure since 1945, and one of the highest figures of the 20th century. This suggests a substantial “backstep” in the quality of play due to expansion:
1960
|
2.84
|
1961
|
2.87
|
1962
|
2.80
|
1963
|
2.94
|
1964
|
2.90
|
1965
|
2.98
|
After 1965 these figures went down rapidly—until the next expansion in 1969. After the second expansion, they again worked their way upward:
1965
|
2.98
|
1966
|
2.85
|
1967
|
2.70
|
1968
|
2.59
|
1969
|
2.70
|
1970
|
2.70
|
1971
|
2.80
|
The 2.59 figure in 1968 was the lowest since 1921. Another observation here. It is generally believed that the National League in this era was stronger than the American League—but this indicator does not reflect that. In fact, the American League numbers in his era were consistently lower than the National League numbers—lower every year from 1957 through 1965. It wasn’t until 1966 that the National League scores caught up to the American League—and even then they were almost the same.
In the 1970s the Age Deviation Scores declined slowly:
1971
|
2.80
|
1972
|
2.79
|
1973
|
2.77
|
1974
|
2.77
|
1975
|
2.84
|
1976
|
2.71
|
1977
|
2.75
|
1978
|
2.74
|
1979
|
2.69
|
1980
|
2.63
|
1981
|
2.65
|
In the late 1970s, after the beginning of the free agent era, salaries escalated very rapidly. This kept older players in the game, and this caused the Age Deviation Score to go up again. By the mid-1990s, however, the indicator had worked its way down to the all-time low:
1980
|
2.63
|
1981
|
2.65
|
1982
|
2.77
|
1983
|
2.81
|
1984
|
2.84
|
1985
|
2.83
|
1986
|
2.92
|
1987
|
2.81
|
1988
|
2.63
|
1989
|
2.62
|
1990
|
2.60
|
1991
|
2.68
|
1992
|
2.61
|
1993
|
2.53
|
1994
|
2.47
|
2.47 remains the lowest figure ever, in 1919 and in 1994. I’ll discuss this a little later.
After 1994, of course, baseball was in the heart of the steroid era. Steroids are a youth drug; they enabled players to do things at ages 35 and above that had never been done before at any age. This causes the Age Deviation Score to move consistently higher for more than ten years:
1994
|
2.47
|
1995
|
2.48
|
1996
|
2.54
|
1997
|
2.62
|
1998
|
2.73
|
1999
|
2.74
|
2000
|
2.68
|
2001
|
2.74
|
2002
|
2.71
|
2003
|
2.75
|
2004
|
2.76
|
2005
|
2.81
|
2006
|
2.91
|
It moved higher, yes, but not all that much higher. The ADS moved back to the levels of the 1960s.
Since 2006, with the gradual elimination of steroids and their lingering after-effects, the number has reduced significantly:
2006
|
2.91
|
2007
|
2.88
|
2008
|
2.78
|
2009
|
2.67
|
Providing further evidence, if more is needed, that we are moving beyond the steroid era.
The Age Deviation for the National League has been higher than the score for the American League in every season since 1999, often much higher—consistent with many other types of evidence showing that the quality of play in the American League has moved ahead of the quality of play in the National.
It is my belief that the Age Deviation Score is in general an indicator of the quality of play within the league, but we don’t want to be overly confident of this. The rising scores in the steroid era clearly do not indicate a decline in the quality of play, however much we might prefer to believe that that was true. The steroids enabled players to increase their athletic abilities at ages when these would normally have been in decline. This may have been morally wrong, but it is not evidence of a decline in the quality of play in the game, and it should not be interpreted that way.
It is my belief that the quality of play in the major leagues has improved steadily over time—but the Age Deviation Score does not exactly support this belief; other evidence does, but the Age Deviation Score does not. The Age Deviation Score shows a very rapid improvement in the quality of play from 1876 to about 1900, but relatively little improvement since 1900. The lowest score ever was posted in 1919, and matched in 1994.
But that argument, too, should not be overstated. The score for 1919 was the lowest ever, true, but the average score for the years 1910 to 1919 was 2.86. The average score for the years 1990 to 1999 was 2.60. The data for 1919-1921 is just kind of a fluke, an anomaly. Not too much can be read into it.