How much impact do managers really have on their teams?
Lou Piniella's recent and somewhat sudden retirement as Cubs manager last Sunday has made managerial maneuverings a hot topic, and appropriately so. With Piniella's departure, the Cubs became the sixth team to change managers this season, only two of which have yet committed to a new skipper beyond the end of this season. Meanwhile, seven other managers are in the final months of their contracts, including Bobby Cox of the first-place Braves and Cito Gaston of the Blue Jays, both of whom have announced they are stepping down at season's end. Rounding out the group are Ron Washington, skipper of the first-place Texas Rangers, and the big-name quartet of Joe Torre, Tony La Russa, Dusty Baker and Joe Girardi, who have a combined 13 pennants between them, and three of whom are still deeply involved in this year's pennant races. Add in the Mets' Jerry Manuel and the A's Bob Geren, who have options for 2011 that have yet to be picked up, and exactly half of the teams in the majors will have had to make decisions about who their 2011 manager will be between July 28 of this year, when the Orioles gave Buck Showalter a four-year deal to be their new skipper, and next Opening Day.
All of which begs the question: Just what kind of impact does a manager have on a team's performance anyway? In the essential 2006 book Baseball Between the Numbers, analyst James Click tried to tease some signs of managerial impact out of the statistical record but came up empty. After examining the measurable impact of in-game strategies (bunting, stolen bases, intentional walks), wins and losses relative to run differential, playing time distribution, in-game substitutions (pinch-hitters, relief pitchers, and defensive replacements), and direct impact on player performance (coaching), Click was unable to find evidence of a repeatable skill in any one of those five areas for any of the 456 managers he studied. That is to say that, much like clutch hitting, individual performances varied so much from season to season that the results appeared to be as much the result of chance as anything else.
The comparison to clutch hitting is apt here. The analytical stance on clutch-hitting is often misrepresented along the lines of "those eggheads don't believe clutch hitting exists." That's not exactly true. If a batter hits a three-run homer in the ninth inning with his team trailing by two runs, that is undeniably a clutch hit. However, the statistical record strongly suggest that clutch hitting for the most part is not a repeatable skill. In part because of the small sample sized involved, a player's performance in clutch situations will vary too much from season to season for that performance to be considered a meaningful indicator of future performance. Over the length of a career, as that sample size grows, those performances will trend toward the player's baseline level of production. Consider Derek Jeter, a.k.a. Captain Clutch, who early in his career was lauded for coming up big in the postseason. Now that he has had roughly a full season worth of postseason plate appearances (637 PA in 138 games) we can see that his performance in the postseason (.313/.383/.479) has been roughly equivalent to his career performance in the regular season (.315/.385/.455).
However, in a 2004 article in The Baseball Research Journal, Bill James famously, and controversially, described this method of determining the persistence of a phenomenon as "the Fog," as in "just because you can't see it doesn't mean it's not there." Indeed, in Baseball Between the Numbers, Nate Silver did manage to tease out some minutely measurable amount of repeatable clutch hitting ability. Though no similar breakthrough has occurred in the field of managerial evaluation, both James and BP, in their respective annual publications The Bill James Handbook and Baseball Prospectus, have begun to list managerial statistics as best as they can compile them.
The problem is that even those purely cumulative statistics contain a significant amount of gray area. In tallying the in-game strategies of the 2,430 regular season baseball games played every year, it is impossible to definitively determine when a sacrifice bunt or pitchout was called from the bench or instigated by the players on the field, when a runner in motion was stealing or part of a hit-and-run, whether a base stealer was explicitly told to go or simply had a "green light," when a sign for any of those tactics was missed or misinterpreted, when a day-to-day injury, illness or even absence rendered a top pinch-hitting or relief pitching option unavailable, or when an in-game injury prompted a substitution. Lacking the ability to accurately determine exactly what a manager is up to, it is no surprise that it has thus far proven impossible to determine exactly what impact he has had.
It's interesting to note, however, that Click's 2006 study did make one relatively firm conclusion regarding the impact of certain in-game decisions. "Only six times in thirty-three years has any manager used sacrifice attempts, stolen base attempts, and intentional walks to increase his team's win expectation over an entire season. Even the best managers cost their team more than a game per season by employing these tactics. At worst they can cost a team three games per season." Over multiple seasons, no manager employed those tactics for a positive effect.
That supports the belief that the best baseball manager is one with a strong roster who is smart enough to let his players play and stay out of the way. Think sabermetric hero and Hall of Fame Orioles manager Earl Weaver, who often said he managed for the three-run homer (which is to say, he let his hitters hit), or Charlie Manuel, the hitter-friendly manager of the two-time defending National League champion Phillies.
Rather, a manager's most important job is widely believed to be the distribution of playing time. It's intuitively true that a manager is only as good as the players he's given, but a good manager can get more out of those players than a bad manager by knowing when, how, and how much to deploy those players. One of the secrets to Casey Stengel's success with the Yankees, which he passed down directly to Billy Martin, was his knack for maximizing performance through the use of platoons, a tactic he himself picked up from having been platooned by John McGraw as a lefty-hitting outfielder for the Giants in the early 1920s (Stengel's best season came in 1922 when he made just 10 plate appearances in games started by left-handed pitchers).
Consider the impact Gaston had when he returned to the Blue Jays in June 2008. The Jays had gone 35-39 (.473) under John Gibbons, but surged to 51-37 (.580) under Gaston. Part of the reason for that change was Gaston benching David Eckstein, still playing shortstop under Gibbons, in favor of Marco Scutaro, calling up Adam Lind to take over in left field for the past-date trio of Shannon Stewart, Brad Wilkerson and Kevin Mench, and dropping a then-struggling Scott Rolen down to sixth in the batting order. Gibbons could have done the same, but he didn't.
Of course, as Showalter is already discovering with the Orioles, who have won just four of their last eleven games under their new skipper, such improvements can be fleeting. The Jays went 78-51 (.605) in Gaston's first 129 games back at the helm of the Jays, but just .48-73 (.397) over the next 121 through the end of the 2009 season, and 125 games into this season have effectively split the difference at 66-60 (.524).
Distributing playing time extends to the deployment of pinch-hitters and relief pitchers as well as to the batting order, which at its most basic level is a way of distributing plate appearances, with each successive spot receiving roughly 20 fewer plate appearances than the one above it over the course of a full season and the leadoff spot receiving about 150 more chances at the dish than the ninth-place in the order. Certainly the deployment of the right players can have a huge impact on team performance, but the degree to which we can credit even those decisions to the field manager remains difficult to discern.
Just as we can't be sure when a certain in-game tactic is being ordered by the manager or enacted by players acting on their own (or missing signs), we cannot be sure to what degree a manager is pulling the strings regarding playing time and other broader strategies and to what degree they are being pulled from above by the front office. The A's under general manager Billy Beane are notorious for handing down lineups and other specific instructions to their field managers, and the Red Sox since the arrival of Theo Epstein have a similar reputation. In a very different way, the late George Steinbrenner was known to order his managers, including Martin, to bench or play a given player based on anecdotal evidence at best.
Many mangers and general managers work together to maximize their 25-man roster, but not all managers get their way. It was Brian Cashman, not Joe Torre, who finally and mercifully took the Yankees center field job away from a badly aging Bernie Williams in early 2005 and installed Robinson Cano at second base and Chien-Ming Wang in the starting rotation, and when Torre wanted Williams back in 2007, it was Cashman who refused to offer the faded Yankee great more than a non-roster invite to spring training, an invitation the proud Williams declined, likely to Cashman's relief. Similarly, Billy Martin famously wanted Joe Rudi, not Reggie Jackson, from the first big free agent class in the winter of 1976, calling in to question just how much credit Martin deserves for the Yankees 1977 world championship.
Given all of that, it seems impossible to make any sweeping generalizations about the impact of managers as a group. It makes little sense try to credit Whitey Herzog and Art Howe, for example, with a similar level of influence on their teams' fortunes. Herzog, who was just inducted into the Hall of Fame last month as a manger, was both manager and GM of the Cardinals in 1981and Howe "was hired to implement the ideas of the front office, not his own," as then-A's GM Sandy Alderson told author Michael Lewis in Moneyball. Perhaps that's why even the best-paid managers in baseball -- Torre, La Russa, Piniella -- make roughly as much as a well-paid set-up reliever. For instance, Torre's current contract pays him an annual average of $4.3 million; the Dodgers are playing lefty reliever George Sherrill $4.5 million this year.
According to WXRL, Baseball Prospectus' cumulative win-expectancy based statistic for relief pitchers, the best set-up man in a given season (the White Sox' Matt Thornton last year, the Red Sox' Daniel Bard this season) is worth between four and five wins. According to the market, the best baseball manager isn't worth any more than that. We can't prove that that's true, but I'm willing to believe it.
