By The Numbers: The Baseball Time Machine
New York Yankees
Buy Yankees Tickets
Mets CentralShop for Mets Gear
Buy Mets Tickets
By Father Gabe Costa
» More Columns
With a nod to H. G. Wells, I suspect the only real way to compare a team from one era with a club from another time is to build a Time Machine.
Since we don’t seem to have the necessary parts (yet?) for time travel, we have to appeal to other approaches.
In the past, for example, countless baseball fans tried to settle questions like, “Would the Big Red Machine of 1975-76 beat the Gas House Gang of 1934?” by playing table games such as APBA Baseball, a topic addressed in one of our previous blogs.
Sometimes “All-Time tournaments” have been simulated, in one way or another, by the use of random numbers in conjunction with computers.
A third approach is more analytical in nature. The aim is to try to contextualize great teams of the past and present. Two authors, Bob Neyer and Eddie Epstein, came up with an interesting quantitative measure known as the Standard Deviation Score (SDS), which does precisely that.
In a book they published, Baseball Dynasties (W. W. Norton and Company, New York, NY (2000) – ISBN0-393-32008-1), these authors developed the concept of the SDS, defining it as follows:
Where the abbreviations denote the following quantities:
RS = Runs Scored by the Team
RSL = Average Number of Runs Scored by each team in the League
RA = Runs Allowed by the Team
RAL = Average Number of Runs Allowed by each team in the League
= Standard Deviation of Runs Scored in the League
= Standard Deviation of Runs Allowed in the League.
As the reader can see, there are two parts to the SDS: an offensive part and a defensive term.
Those fans that have had some background in statistics will recognize each of the two components as kind of a “Z-score” which “normalizes” the data. Because of this, the larger the SDS, the better the team as compared with its rivals for that particular year.
The authors assert that any SDS of +3.00 or greater is very good.
Let us apply the formula for the 1927 Yankees. That year Murderer’s Row and the American League ended up with the following numbers: RS = 975, RSL = 762, RA = 599, RAL = 762, = 115, and = 88.5. These values give a SDS score for the 1927 Yankees of +3.69, as we can verify below:
Some other great Yankee teams of the past have had the following SDS: 1939 Yankees (+3.52), 1947 Yankees (+3.05), 1961 Yankees (+2.97) and 1998 Yankees (+3.88).
The 1998 Yankee squad, incidentally, had the highest SDS for the 20th century.
Let me address two other points about the SDS. First, the scores can be added; therefore, some sense of a “dynasty” can be discussed over a two-, three- or four-year period. Up through the year 1999, the 1997-1998 Yankees had the highest two-year SDS with a measure of +6.70. For a three-year period, the 1969-1971 Baltimore Orioles had a cumulative SDS of +9.86. And for a for a four-year stretch, the 1936-1939 Yankees recorded the best score, coming in with a SDS of +12.83.
A second point the authors bring out is that SDS scores can be negative; thereby reflecting how poor a team was in the context of the league during the year(s) in question. The 1962 Mets had the lowest SDS ever, with a mark of – 5.91. The runner up team in this category was the 1916 Philadelphia Athletics; Connie Mack’s A’s SDS was – 4.39.
Many sabermetricians like a statistic such as the SDS; it is revealing and easy to compute, especially if the researcher can import data from a website into a spreadsheet. And such a measure, in conjunction with other instruments, can increase the plausibility of sabermetrical conclusions. For example, one could argue that the 1906 Chicago Cubs with a SDS of +3.73 was a better team than the 1942 St. Louis Cardinals, a team which had a SDS of +2.94.
But I still wish we had a Time Machine.
Got all that? Give us your thoughts on sabermetrics below…