Swarthmore Statistician Uses Data to Analyze MLB Managers

Alisa Giardinelli

Swarthmore Statistician Uses
Data to Analyze MLB Managers

Uses Stats to Describe Managers as They Do Players

by Alisa Giardinelli
02/25/2008

Steve Wang

Associate Professor of Statistics Steve Wang

 

Baseball players have long been judged by their stats. Just in time for opening day, a statistician at Swarthmore has devised a new way to put managers under similar scrutiny.

"With players, you can use their stats as a language for their strengths and weaknesses," says Steve Wang, an associate professor of statistics. "With managers, we don't have that language. A lot of what we talk about is subjective and anecdotal. I'm looking for a systematic way to describe the skills and preferences of different managers."

Wang analyzed six year's worth of raw data for current managers on 23 criteria - from how often they go to their bullpen to the number of lineups they use in a season. Wang says the information has been available for years in the annual Bill James Handbooks, but is unknown to most fans. He presented his findings at the AAAS annual meeting in Boston earlier this month.

Wang determined that many of a manager's preferences are consistent over time and can therefore be measured. For example, in the American League (AL), the Angels' Mike Scioscia has consistently used near the fewest number of relief pitchers, while Buck Showalter in Texas was always close to the league lead. How is that useful information?

"Of course, a good manager should adapt to his personnel," Wang says. "But managers have styles as well, and certain styles might be more effective with certain kinds of teams. A manager who prefers to stay with his starters might be best suited for a team with veteran starting pitching, whereas a team with fragile young arms might do best with a manager who uses his bullpen aggressively."

Another telling difference, especially in the National League (NL), is the number of lineups used. At the top of the list in 2007 is the Cardinals' Tony LaRussa, who used 150 different starting lineups in 162 games. In contrast, the Phillies' Charlie Manuel used the fewest variations, 87.

"If you have good players who stay healthy, the tendency is to leave them where they are," Wang says. "But it's also a manager's preference and shows a clear difference in style."

Perhaps not surprisingly, Wang says the connection between a manager's preferences and his win-loss record is tenuous since it depends so much on the players. But he is careful to note that last year's three AL division-winning managers - Terry Francona, Mike Scioscia, and Eric Wedge - "are similar to each other in that they are moderate in most pitching-related categories."

Wang's next steps include studying how managers can be grouped into types, much like players can be described as sluggers or leadoff men. "As a statistician, you try to find patterns that are not obvious on the surface," Wang says. "I'm using a complicated data set about which not much is known. It's a fun challenge."

Wang, a Yankees fan, came to statistics from his interest in baseball, not the other way around. But although he wrote his master's thesis on Markov chain models of baseball lineups and has taught a class on the Baseball Hall of Fame, he primarily works on statistical problems in paleontology.

Additional examples from significant manager stats:

  • letting a closer stay in for more than one inning. "With the Yankees, Joe Torre did this 10 times a year and was near the top in this category in the AL because he had Mariano Rivera," Wang says. "Eric Wedge in Cleveland never allowed this in 2004 and has done so only one to three times a year since."
  • use of pinch hitters. "At over 100 a year, Minnesota's Ron Gardenhire used about twice as many pinch hitters as Buck Showalter did when he was in Texas. A typical NL team, on the other hand, uses 200-300 or so in a year, since the pitchers bat."
  • number of long starts and quick hooks: "Bruce Bochy and Ozzie Guillen were more conservative in staying with their starting pitchers the longest. At the other extreme are Bobby Cox and Manny Acta. Joe Torre also became more aggressive about going to the bull pen over time (2002-07), particularly because of injuries to his front-line starters in 2007."