Understanding Sabermetrics (13 page)

2. In 1983, Baltimore Orioles shortstop Cal Ripken won the American League Most Valuable Player Award over his teammate, first baseman Eddie Murray, in a close contest. Based on the original linear weights batting run production, who should have won the award? Here is the data:

3. Using the original linear-weights model, calculate the batting runs for the starting line-up of the 1960 Pittsburgh Pirates: C Smoky Burgess, 1B Dick Stuart, 2B Bill Mazeroski, SS Dick Groat, 3B Don Hoak, and Outfielders Roberto Clemente, Bob Skinner, and Bill Virdon. Here is the data:

1. Calculate the batting runs for the top five home run hitters of all time, for the season in which each won his first Most Valuable Player Award: Hank Aaron in 1957, Barry Bonds in 1991, Babe Ruth in 1923, Willie Mays in 1954, and Sammy Sosa in 1998. Who contributed the most wins to his own team? Here is the data:

2. Consider the following Hall of Fame catchers, all of whom played before 1930: Roger Bresnahan, Buck Ewing, and Ray Schalk. Who contributed the most runs to his own team (use the condensed form for Linear Weights Batting Runs)? Here is the data:

3. Only six players have hit 60 or more doubles in a season (notice in which era they all occurred). They are:

The linear-weights pitching-runs model assumes that pitchers are responsible only for earned runs. Thorn and Palmer argue that the earned run average statistic is an indication of a pitcher’s rate of efficiency, not an indication of his actual benefit to a team’s overall performance. As an example, if a team has two pitchers with identical ERAs who are compared against a higher league ERA, and pitcher A pitched in twice as many innings as pitcher B, then pitcher A must be worth twice as much as pitcher B to his team. The linear-weights pitching-runs statistic seeks to measure the number of runs, beyond the average, that a pitcher has prevented from scoring. The formula for earned run average is:

What about the average number of runs for a pitcher? This will equate to a pitching runs number of zero. The average number of runs is:

For example, in 2002, Roger Clemens had an ERA of 4.35 in 180 innings pitched for the New York Yankees. The American League average ERA was 4.38 runs (almost identical). Had Clemens’ ERA also been 4.38, he would have held opposing batters in check at the league average, regardless of the number of innings he pitched. In 2005, by contrast, Clemens, then with the Houston Astros, authored an amazing ERA of 1.87 in over 211 innings pitched. This was the lowest ERA for a season in Clemens’ 23-year career. The National League average was 4.14 runs. Surely, Clemens saved a significant number of runs that an average pitcher, pitching in his place, might have allowed. How to find that number? Thorn and Palmer’s linear-weights pitching-runs formula is straightforward and simple: