What about wins? We know that with pitchers, wins are not an effective way to judge performance. Too many factors outside of the pitcher's control - no matter how many home runs Carlos Zambrano hits - effect the outcome of games: offensive performance, relievers, umpires, and rioting fans are the most obvious.
Managers, however, are not afforded the same luxury. The numbers in the wins and losses columns constitute the only stats that matter when evaluating managerial performance. The manager's job requires that he pull together and assert some degree of control over all those outside factors that let pitchers off the hook for their win/loss totals. Like a director who has to oversee and bring together the myriad elements in order to make a film, managers are responsible for the final product, the team you see on the field for 162 games and their W-L record through those 162 games.
Simple enough, but it all started to make me wonder whether or not stats we toss around so easily in evaluating players can be totally discounted in evaluating managerial performance. And just what exactly would those numbers reveal about managerial performance?
As you can see, this is a project that has the "nebulous" label stamped allover it. The factors that can go into this are wild and variant and subject to things that stats just can't evaluate. In the new economy of MLB, rosters vary greatly from season to season, one big free agent signing can represent the difference between a losing season and angry fans versus a winning season with star crossed eyes in every seat. What this little indulgence represents is a starting point, the first step in comparing managerial performances.
What I wanted to first is look at the basics. How did the hitters hit and the pitchers pitch from one manager to the next? To do that, I decided to start by looking at midseason managerial changes. The main reason was that the team's make-up, the guys on the roster, was far less likely to change dramatically during a season than from season to season where owners and/or GMs - who themselves carry lots of responsibility in the performance of a team because of their role in determining team composition - can significantly reshape a team. It's also helpful because the manager's supporting cast of characters - pitching and hitting coaches - are also less likely to change. A third reason for starting with the midseason manager switch is that we see players in the midst of a single season, rather than between seasons where they can undertake various individual programs to alter the way they play the game (a physical conditioning program for example).
Last season, miraculously, there were no midseason manager changes, but the season before that, 2005, saw four teams sack a skipper before playoffs started. Those teams were Baltimore, Cincinnati, Kansas City (who actually had three managers that year), and Pittsburgh. I needed a starting point in this process, and for that I decided to start by comparing the two different managers who managed close to the same number of games for the respective teams that fired and hired them. In Pittsburgh, Lloyd McClendon got axed after managing 136 games. His replacement, Pete Mackanin managed just 26 games in September with a roster of call ups. It was the opposite in Kansas City, where Tony Pena managed 33 games, interim solution Bob Schaefer stood in for 17, and Buddy Bell took over for the last 112. With the Orioles, things looked like they might be a little easier, with Sam Perlozzo replacing Lee Mazzilli after 107.
Analysis: First Step on a Long, Long Road
Cincinnati, however, had exactly what I was looking for as the ax fell on Dave Miley on June 20, 2005, after 70 games, and Jerry Narron got the call for the 93. It's not a dead-even split, but it's close enough not to be skewed by minor league replacement players. The tumultuous 2005 season endured by the Reds became my starting point.
I started with the most basic stats, the wins and losses.
As you can see, Narron fared much better than Miley after taking the helm. So let's look at the next most basic stat, runs scored and runs against.
Scoring runs was never much of a problem for the Reds that season. Throughout the tenures of each manager that year, they ranked towards the top of the NL in runs scored.
Compare the overall batting lines for the Reds under the two managers.
|Manager||AVG||OBP||SLG||BB rate||K rate||Contact rate|
As you can see, although the Reds still scored runs in grosses, their overall hitting line got better. They walked slightly more, they struck out less, and they hit better. Managers can make a difference by emphasizing plate discipline and the other things that can impact a player's approach, but tweaking the lineup can make a more immediate impact. That's exactly what Narron did, although his changes didn't come fast. Miley used Dunn in the fifth spot primarily, but by the end of the season Narron had used the slugger in the third and fourth spots for the most part (he continued to shift Dunn's spot into 2006).
Before we delve deeper into that, let's look at the pitching splits between the two managers. Now here, I started with the starting pitcher splits because they're readily available via Pinto's day by day database. Unfortunately, there's not much out there (in my admittedly limited knowledge that allows you to break down day by day for relievers or even teams stats on the whole). Starting pitching problems were synonymous with the Reds through the recent past until things started to level out a bit last year. Dissonance between the stats for Reds' starters under the two managers still allow us to identify key differences as a starting point. From there - and this is my next step - is to look at results under the two managers by looking at the results shown by the individual players.
The starting pitching got better under Narron. Starters walked fewer batters and gave up fewer home runs, at the cost of decreased strike outs. What this tells us instantly is that pitchers under Narron altered their approach slightly, opting more for control to keep runs off the board and limit damage from the long ball in a miniature ballparks known to favor hitters. It's not much, and it's circumstantial at this point, but this emphasis is a well-known short-term fix for a mediocre pitching staff that bleeds runs in park known for the long ball.
However, we must acknowledge the presence of other factors that could impact this. Most glaring, is the stability in the rotation's fifth spot. Paul Wilson went down after 9 starts, in which he posted a lucky 7.77 ERA. Four other pitchers made starts that season filling in for injuries, with Luke Hudson getting 16 starts for the fifth most on the team that season.
Quick note, here's another great limitation (and again this could be a limitation in my knowledge of statistical tools available) in that batted ball data would be useful here, but there is nothing out there that can break down batted ball data on a day by day basis. Again, next time I'll look at individuals and try and break it down by G/F ratio pre- and post- All Star. It's not a perfect split by any means, but it can give a decent approximation for our purposes.
Next time, (Oh God, there's more?) I'll use qualitative evidence to back up what we see in the stats. Did Narron's lineup changes result in more production from certain players? Did Narron emphasize control over Ks from his pitchers? At the moment these findings are circumstantial, but that kind of evidence will help us build a case.
Splits from individual players between the Narron and Miley days in Cincy can help confirm the thesis as well. I suspect, while the case is far from a solid one, that the stats will confirm a different approach from manager to manager.