Let's start with the basics: my name is Chris Jaffe and I just wrote a book I'm immensely proud of (to be fair, as author I'm a bit biased, now aren't I?) titled: "Evaluating Baseball's Managers, 1876-2008" which you can order if you want.
It's because I wrote the book that I decided to start this blog. While I'd love to say the main motivation for this site is for the betterment of humanity or to foster greater understanding among the post-prehensile tail set, my concerns are a bit more self-centered.
After writing my book, I learned that it was important for me to try to promote the sucker actively. (DAMMIT!) Otherwise, it could easily come and go, unnoticed and unread. From what I know, having a blog is apparently one of the best ways to promote a book and build awareness for it. In theory anyway.
That sounds mighty self-interested, and to some extent it is, but only to some extent. I wouldn't bother going through the motions of starting up a site like this unless I truly did feel this was something well worth reading. To paraphrase the late Samuel Goldwyn, I don't care if my book makes a nickel - I just want every man, woman, and child in the country to read it. For what it's worth, based on the feedback I've gotten from people who read the manuscript, I'm right to think it's worth half a fart.
About the book
The book has a basic starting point: managers are primarily managers of men and not of the game. Thus much commentary of the profession is off-the-mark because it focuses on in-game strategy instead. This is especially true of virtually all sabermetric commentary on the profession, as it is far easier to quantify things like bunts that ability to get the most out of players.
Among other goals, this book tries to fill that gap. The idea for this book first popped in my head when I saw Phil Birnbaum give a presentation at the annual SABR convention in 2006. He created a database to determine how much teams under/overachieved in a given season. He termed the disparity luck. However, while luck is certainly one factor that explains why a team would do better or worse than one might expect, it's not the only reason. For example, looking at the results it was amazing how "lucky" teams managed by Earl Weaver always were, or how "unlucky" Don Baylor's squads were.
Two elements of Birnbaum's database really caught my attention. He created two algorithms - one for hitters and the other for pitchers - to determine if a player did better or worse in a given season based on how he performed in the two previous and two succeeding seasons. These algorithms (based on Runs Created and Component ERA respectively) appeared to be a real breakthrough to me. They promised to be the first things produced that could really measure the impact a manager could have on his players.
However, I didn't want to claim these things measured coaching just because I thought so. I sought some further evidence. To that end, I divided all teams in MLB history into one of four categories: those helmed by managers who lasted at least 2,000 games, 1,000-1999 game managers, 500-999 game managers, and 499 or fewer.
My rationale was simple - assuming there is any correlation between length of managerial career and skill (and that's an assumption I feel incredibly safe in making), the longer a group of managers lasted, the better they should do. If it were luck, that would not be the case - the longer a group lasted the more luck should even out. It's awful tough to be the luckiest group when each individual has a sample size of 13+ seasons on the job.
Furthermore, I theorized that if the algorithms represented skill, the gap between the two middle groups (the 1,000 to 1,999 guys and the 500 to 999 guys) would be the smallest overall. They were the two most mediocre groups while the 2000+ and 499- groups should have an undue proportion of the best and worst managers pulling them away from an average score. However, if Birnbaum's work (henceforth referred to as the Birnbaum Database) measured only luck, the gap between the 1000s and 2000s would be smallest, because of the large sample sizes involved.
Much to my delight, the test confirmed my belief that the Birnbaum Database indicated what impact a manager had on his team. Please note this falls short of 100% indisputable evidence, but the evidence is strong enough for me.
I should note the Birnbaum Database actually covered five different areas. Aside from the two algorithms, he also looked at : 1) deviation from expected win-loss records based on pythagenpat projections, 2) actual team runs scored versus expected run scoring based on overall offensive production (again using Runs Created), and 3) team runs allowed versus expected runs allowed (based on Component ERA).
I wasn't sure what to expect from these final three components, but the results from the 2000s/1000s/500s/499s test also indicated some managerial impact. The results aren't as strong as they were for the two algorithms, but they were far more in line with what one would expect from managerial skill than from pure luck. I theorized that the runs allowed/scored components focus on in-game tactical ability of managers (just because a stolen base is worth X-runs on the whole doesn't mean each stolen base is worth that), and pythagenpat deviation might say something about how he structures the roster. I'm willing to believe that all five components indicate managerial ability, but the two algorithms are the two I'm unquestionably sold on.
While the above inspired me to investigate managers, it wasn't until I created a database of my own that I felt I could write a book on managers. I call it the Tendencies Database. It's inspired by a brief segment in the Bill James's Guide to Managers in which James notes how often different managers led the league in certain statistics, ranging from bunts to homers. I took it a step further. With Excel, I incorporated how well all a manager's teams ranked in a given category rather than just the times they led the league.
The notion was simple: figure out where all a manager's teams ranked in a given category, adjust for league size (which is actually a bit more difficult than it sounds), and compare the managers across era. In the results, 1.000 is always an average score. The lower score usually means the manager fared better in that stat. (It might sound odd that lower means better, but it makes sense because the Tendencies Database is based on how teams rank in a given category. Ranking first is a good thing usually as well as the lowest possible rank).
I ran several dozen queries with the Tendencies Database, including obvious ones like which manager liked to sacrifice bunt or use the most relievers. I also ran trickier to research ones, such as who relied the most on their starting lineup and who used their bench the most often. To answer that question, I determined what percentage of a team's plate appearance went to their starting eight position players (or nine in a DH league) for every team in history. I've never seen it studied, but it certainly is important.
In fact, even with the simple queries such as who sacrifice bunts the most I adjusted to find a more accurate result. The sacrifice comes only when there is someone on base, so rather than look at simply SH, I take SH divided by opportunity, to see which managers bunted the most when they had the chance.
I limit the Tendencies Database to skippers who lasted at least ten seasons as a team's primary manager. As an added bonus, since an average score is always 1, you can combine results to determine items such as who was the biggest small ball manager of all time (add together the sacrifice hit and stolen base results) or any similar idea. The Tendencies Database is explained in greater depth on pages 110-113 of The Hardball Times 2008 Annual
Neither the Birnbaum Database nor the Tendencies Database are perfect analytical tools for evaluating managers, but then again imperfection is not a synonym for useless. For lack of a better way of putting it, I kept my BS Detector on when looking at the data to figure out when they illuminate our knowledge of managers and when it's distort it. By and large, they provide a lot of helpful information on the peculiarities and distinctions of baseball managers.
There is a third avenue of exploration used to evaluate managers, but this one works only with managers from the mid-1960s or earlier. It used to be that a manager would decide who his starting pitcher would be based on the identity of the opposing squad rather than have them pitch in the modern ABCDE rotating order. I call this practice pitcher leveraging, because the goal was to leverage your best arms for the most important games in order to make those contest count as much as they could in the team's favor.
I created a few stats to evaluate the degree of pitcher leveraging. First is AOWP+, (the acroynym stands for Average Opponent Winning Percentage) which I debuted in a column over at The Hardball Times. AOWP+ determines if an individual pitcher was leveraged in a season. A score of 100 means he wasn't at all, a higher score means he faced quality teams a disproportionate number of times, and a lower score means the opposite.
Based on AOWP+, I created a stats called LPA (Leverage Points Added) and LPA+ (which adjusts LPA for era, as pitcher leveraging ebbed and flowed over time). Both LPA and LPA+ apply to managers. The more extreme the AOWP+s were under a manager, the higher his LPA and LPA+ wil be. (Like AOWP+, an early version of LPA and LPA+ were used in a Hardball Times column).
I only look use AOWP+ and LPA+ for managers from 1876-1965 because leveraging clearly came to an end in the mid-1960s. And when I say it clearly came to an end, I looked up over 90% of all starts from 1876-1969 and a chunk of starts from 1970-onward, and there is no evidence of leveraging existing as a meaningful strategy. Ultimately, LPA+ and its ilk are a side note in my book, but it's still an important sidenote, especially given that as far as I know, no one has studied how managers leveraged their pitchers in any meaningful way.
The heart of the book, however, isn't on math, but on managers. The first part of the book consists of three chapters explaining the above areas in much more detail (and providing key results as to the best/worst/most extreme managers of all time), but that merely sets up the second part of the book. In the six chapters in part two, I apply the above to 89 of the most important managers in baseball history. Those 89 include the 77 managers who lasted at least 10 seasons as a team's primary manager from 1876-2008, and a dozen of the most important others. (Two of those dozen others - Mike Scioscia and Terry Francona - managed their tenth season in 2009).
I can honestly say that I'm proud of my entries on each of the 89 managers. They're not all of equal value - sorry all you Jimmy McAleer fanboys, but your guy isn't the highlight of the book - but I found something to say about all of them.
About the blog
Anyhow, since I'm proud of my book I'm willing to promote it, and rumor has it that starting a blog is a good way to promote a book. There must be something to the notion of blogging for promotion -- Joe Posnanski's first blog was The Soul of Baseball Blog, named after his Buck O'Neil book. Howard Megdal created a site for his book "Baseball Talmud". Perhaps most successfully of all, Tangotiger and MGL's The Book Blog has become a must-read site for those into advanced sabermetric thinking.
I'm not nearly as ambitious with this blog as some of the writers above are with their sites. My main writing focus will still be my weekly gig at The Hardball Times, where my columns regularly appear every Monday. This primarily exists for me to add in whatever thoughts on baseball I have that aren't in and of themselves enough to fit into a full-length column. That's not intended as a slam on what I'm doing here, either -- my typical column there is a bit over 2,000 words). A lot of what I'll write might be simple/basic bits such as my Random Facts of the Day pieces I've done at THT Live.
Obviously, manager-related information will be prioritized (this is, after all a blog spawned by a book on baseball managers), but that won't be the only item. Heck, I don't even feel the need to write only about baseball. I have a general plan of what I want to do, but I don't see any need to get too fixed in my plans in advance: after all being a bit free form is half the fun of having a blog. I just want to see how it plays out and see what the future holds.
If I like blogging, I'll keep blogging after it's stopped making sense for me to do so for purely book-related reasons. If it's too much a chore, I'll stop. That's another advantage of blogging. Either way, I doubt this place will get updated either day, but if you're interested I've established an rss feed (scroll to the bottom of the site) and it's possible it even works.
If nothing else, could ya show some support for someone with enough chutzpah to start a baseball blog in November? (And by "chutzpah" I mean "stupidity" - but you already figured that one out on your own, I bet).
Have any more questions? Send me an e-mail.
0 comments:
Post a Comment