Baseball Prospectus Needs Your Help! Check out our call for contributors!

On June 23rd, 2011, Jim Riggleman resigned from his post as the Washington Nationals manager, convinced he wasn’t in the Nationals’ long-term plans. The Nationals didn’t look far for his eventual replacement, spending the three games bench coach John McLaren ran the team working out the details of front office consultant Davey Johnson’s return to the dugout after a hiatus of nearly 11 years. Johnson will manage the Nationals for the rest of the season and has an option to return for 2012, according to MASN. At whatever point he is done with managing, he will stay in the Nationals front office to help hire his successor.

How good of a manager did the Nationals pick? The career .563 winning percentage he carried from stints with the Mets, Reds, Orioles, and Dodgers, which ranks eighth all time among managers who skippered at least a thousand games after 1900, suggests that he either had something on the ball or else stumbled into some awfully promising rosters.

As unlikely as the latter sounds, one is hard-pressed to refute it without resulting to anecdote instead of data; managerial skills are not readily defined statistically. The majority of the impact that a manager has on a team is arguably behind the scenes. The impact of on-field, “inside baseball” strategic decisions such as when to bunt or steal is minimal, and even relief pitching decisions are thought out well ahead of time given the modern insistence on reliever “roles,” and most of the time the manager doesn’t have to think hard about when to bring in his best reliever, because he’s slavishly devoted to the saves rule.

Even when a manager does push a button for a bunt or call for a particular reliever, just because it worked out doesn’t mean he was right, and if it a given move doesn’t pay off it doesn’t mean he was wrong. “After therefore because” is a logical fallacy. The best the manager (and we) can hope for is to have a good rationale for his decisions. The actual execution is up to the player.

Where the manager has the most direct impact is in deciding who to play. Baseball is distinct among the major sports in that a team’s probability of winning changes daily, simply because of the fact that the availability of the players on the roster changes. Injury or suspension aside, a basketball, hockey, or football team will generally field the same players for every game (hockey’s one exception is the goalie, who needs periodic rest). A starting pitcher on a baseball team cannot pitch every game. In addition, a manager may decide to rest his best players on any given day. More significantly, when confronted with two players for one position, a manager may errantly choose the less-talented option, making his team worse even though the players on the roster are exactly the same.

Given this cavalcade of caveats, we’ll measure “manager skill” by looking at the difference between a manager’s actual win record and his Pythagenpat or expected win record. Decades ago, Bill James demonstrated that a team’s won-lost record is reflective of its run differential. Teams whose records differ dramatically from their expected record can be inferred to have been lucky, unlucky, or perhaps, well- or mis-managed. James called the simple calculation for a team’s expected record the “Pythagorean theorem.” Over time it has been revised until it reached its current form.  

With the Pythagenpat method, we can get a basic look at how many games a team “should have” won based on their run differential and then look at how many games they actually did win. If a manager’s teams that consistently out-perform their run differential, this may be due to the manager’s skill level, although it should be emphasized that so many different factors can serve to make a team’s final record diverge from its projection that this is no more than supposition. We could be giving the manager too much or too little credit. For example, it is theoretically possible that a manager can actually influence the number of runs a team scores or allows. Hypothetically, if a manager consistently makes the correct call over the course of a season, a team might score (say) 10 extra runs. Under the Pythagenpat method, this would be lumped into the team talent level and not the manager skill level.

In this analysis, Davey Johnson lands 10th out of the 30 current managers, with an average win differential of 0.93. This means that of the teams he has managed, they have averaged almost one win more than they “should have” won.  The current active leader in this category is Kirk Gibson, since his Arizona Diamondbacks have won almost 2.5 more games than they should have this year. Among managers with a little more experience, Fredi Gonzalez (572 eligible games) and Ron Gardenhire (1543) are at the top of the list, as their teams have averaged more than two wins than their Pythagenpat win record indicates. Mike Scioscia and Ozzie Guillen round out the top five.

Another way to see how a manager makes an impact on the game is to look at how many intentional walks were issued, stolen bases were attempted and sacrifice hits were converted for his teams. Baseball Prospectus’s R.J. Anderson recently hypothesized that these would be the most likely statistical categories a manager would directly affect. I compared the manager’s tendencies with the league average for that year. National League teams average more of each of these categories than their American League counterparts, so managers are compared only to their colleagues in their respective league.

I took the total number of each of these statistics for the team, divided them by the league average and multiplied by 100. This creates a stat similar to ERA+, where a value of 100 is exactly league average, and numbers below or above 100 indicate how far from average a manager is in that particular instance. For example, in 1985 the New York Mets—under the direction of Davey Johnson—issued 36 intentional walks. The National League average in 1985 was 65 intentional walks. This gives Johnson an IBB+ value of 36/65*100, or 55.

As it turns out, relative to the league, Johnson’s teams rarely give out intentional walks. The average IBB+ value for his career is 64. The highest value for any of his teams is 80, with the Baltimore Orioles in 1996. Under Ozzie Guillen, the White Sox have the highest average IBB+ value at 132. The Brewers this year under Ron Roenicke have the lowest average IBB+ value at 42.

Doing the same analysis for stolen base attempts and sacrifice hits produces varying results. The SBA+ metric is much more reliant on the players on the team than the others may be. A team full of Rickey Hendersons will average more stolen base attempts than a team full of Bengie Molinas, regardless of the manager’s tendencies, and while we can credit a Whitey Herzog with intentionally stocking his teams with this sort of player, the predilections of other managers, and their involvement in player procurement, is not quite so clear. Johnson’s teams have averaged a SBA+ of 102, just slightly more than league average. However, this covers a range from 56 with the 1997 Baltimore Orioles to 159 with the 1995 Cincinnati Reds. It doesn’t appear that Johnson has any set-in-stone belief for aggressiveness on the base paths.  Joe Maddon’s teams have the highest average SBA+ value, at 146. Mike Quade’s 2011 Cubs rarely run, with an SBA+ value of 50, exactly half of the league average.

Johnson seems to have the same belief regarding sacrifice hits, but in the opposite direction. The average SH+ for his career is 97, slightly below league average. The range here is not quite as large, going only from 75 with the 1989 Mets to 119 with the 1985 Mets. Ron Washington loves the bunt, as his teams average a SH+ of 127. Buck Showalter believes the least in the sacrifice hit. His teams average a SH+ of 61.

Johnson cut his eyeteeth on championship Orioles teams under Hank Bauer and Earl Weaver, and over time he has shown some preferences that show him to be an active thinker of the old-school kind, one who creates opportunities by platooning and creative defensive alignments (famously starting iron-gloved third baseman Howard Johnson at shortstop when extreme fly-ball pitcher Sid Fernandez pitched). If his long time on the sidelines hasn’t changed him, Johnson will find ways to squeeze something extra out of his long-suffering charges while spending little time worrying about the flashier but less meaningful aspects of managing.

A version of this story originally appeared on ESPN Insider Insider.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
I appreciate how hard this is. I expect that managers change tendencies over time, so that some of their record probably doesn't really predict what can be expected now. I also expect that some alter their tendencies to fit their rosters, so they may handle the current team differently. I also think that ignoring the manager's effect on pitching altogether, and their attempts to maximize the focus of the players, can't really be right. But I have no better ideas. I do think that we ought to take even our best attempts with big grains of salt.
Agreed. A good manager should be able to change their style based on their current team. What's good for the goose is not necessarily good for the gander. Separating manager skill from team skill is a difficult exercise.
Welcome to BP, Chris - what can we expect from you going forward?
Thanks! I'll try to focus mainly on research-type things, challenging conventional knowledge and looking at what the numbers really tell us. Also, I'll try to do things like this: - interactive visualizations of certain events.