Read Understanding Sabermetrics for Free Online

Earl Weaver once remarked that saving a run is as important as scoring a run, a sentiment that displayed his appreciation of Gold Glove players who might not always hit for power or average (Mark Belanger or Paul Blair). Rod Carew is in the Hall of Fame for his hitting (3053 career hits), yet his 33 errors at second base in 1974 were the most ever by a second baseman since 1946. His fielding percentage was .015 below the league average, but his batting average was .101 above the league mean.
    Hitting, pitching, and fielding are individual talents. Can the peak-value analysis be applied to teams? Teams win games by scoring more runs than their opponents. Maybe Weaver’s remark is noteworthy in this regard. At the end of a 154-game or 162-game season, did the team who scored the highest number of runs win the pennant? Or, did the team with the most victories prevail and capture the World Series trophy? Can we use the considerations of Figure 4.1 and apply them to teams? (For more on the relationship of runs scored and runs allowed to wins and losses, see Inning 4: The Linear Weights School — Offense).
    The first part of the process is to determine which teams to study. Suppose we look at some teams that have been labeled “dynasties” — those teams that went to the World Series at least three years in a row. Maybe they didn’t win the series all three years, but they played on the big stage for the opportunity. The New York Yankees immediately come to mind (1921-1923, 1936-1939, 1949-1953, 1960-1964, and 1998-2001). Wow! Also, the Chicago Cubs (1906-1908), the New York Giants (1921-1924), the Philadelphia Athletics (1929-1931), the St. Louis Cardinals (1942-1944), the Baltimore Orioles (1969-1971), and the Oakland Athletics (1972-1974 and 1988-1990). Our analysis could determine if these teams exhibited a peak value during their stretch of appearances in the Fall Classic. One drawback of such a study, however, is the homogeneity of the data. In how many cases were the teams fielding the same roster (with the same output) year after year? Were free agents brought into the mix? Were trades a major part of the championship? Perhaps we need to analyze individual teams and compare them to various eras.
    In 2004, the Boston Red Sox fielded a powerful offensive team en route to the World Championship. The next year, the Sox brought back much of the same team, yet they failed to win the American League pennant. Should sabermetricians consider 2004 a peak season for Boston, and how does it compare to the peaks of other Boston teams? What measures should we use? Should we limit the study to just the number of victories? The number of runs scored? Runs allowed?
    Let’s gather some data (initially look at simple statistics — team batting average and earned run average). We should also take relativity into account, as we compare the Boston Red Sox of 2004 to the 1986 Red Sox, the 1975 Red Sox, the 1946 Red Sox, and the 1918 Red Sox. (Incidentally, the Boston Americans / Red Sox won five World Series titles from 1901 through 1918, the last three with a young left-handed pitching sensation named Ruth). In Table 4.1, we show the BA, ERA, Average Runs Scored (RS) and Average Runs Allowed (RA) for these five Boston teams, as well as the league batting average and ERA (the runs scored and runs allowed averages for the league are the same).

     
    Table 4.1 Boston Red Sox team statistics
     

    Which was the best Boston Red Sox team? Which team had the best peak value? Only the 1918 team had a lower-than-the-league batting average. Obviously, we would need to study more than just team batting average, ERA and runs scored or allowed. Keep in mind that in 1918 and 1946, pitchers’ batting statistics were incorporated into the team average in the American League. Look at the difference between runs scored and runs allowed for the Ted Williams 1946 team. In addition, non-sabermetrical factors might be incorporated. Finally, despite all of the