Posts tagged ‘Bill James’

Bill James on ... uh, about Everthing

Awesome article by Baseball guru Bill James about rule-breaking and the core of what makes America dynamic.

You Too Can Be Billy Beane

As a baseball fan, you may have heard something about Bill James, Billy Beane, and/or Sabremetrics, but were afraid all the math was too difficult.  Well, you too can use simple numbers to out-manage most major league skippers.  For today's introduction, you only need one simple table of numbers:

RE 99-02 0 1 2
Empty 0.555 0.297 0.117
1st 0.953 0.573 0.251
2nd 1.189 0.725 0.344
3rd 1.482 0.983 0.387
1st_2nd 1.573 0.971 0.466
1st_3rd 1.904 1.243 0.538
2nd_3rd 2.052 1.467 0.634
Loaded 2.417 1.65 0.815

These are the run expectancy numbers, compiled from data in the 1999-2002 baseball season.  Here is how to read the table: With a runner on 2nd (row three) and two outs (column three) a team on average can expect to score .344 runs the rest of that inning.

So, to test your understanding, how much does a leadoff double increase a team's chance of scoring?  Well, the base run expectancy at the beginning of an inning is .555 runs.  After a leadoff double, you are in the square for man on second, still no outs, which has a run expectancy of 1.189.   On average, then, a leadoff double increases the scoring expectations for the inning by 0.634 runs, which is a lot.  So here are a few simple sabremetric type conclusions you can reach just from this data:

  • Outs are extraordinarily valuable.  For example, man on first and third with two outs has a WORSE run expectancy than you have at the beginning of the inning, ie it is worse than nobody on and no outs.
  • Bunting almost never makes sense.  Assume a runner on first, no outs -- a typical bunting situation.  After a succesful bunt, you have runner on second and one out.  Notice that this has REDUCED the run expectancy from 0.953 to 0.725.  The reason I say "almost" never is that an even worse outcome is a strikeout, which would take you to man on first and one out for a RE of .573.  For batters highly likely to strike out or pop up in the infield (think: pitchers) bunting can make sense.
  • You can actually calculate what percentage chance of success you need to justify stealing second.  Lets again take man on first, no outs.  The RE is 0.953.  If he steals successfully, the RE goes to 1.189.  If he gets thrown out, the RE goes to 0.297 (bases empty, one out).  If X is the probability of stealing success, then 1.189X+0.297(1-X)>0.953.  X must be about 74% or greater.

Exercise: You have two hitters.  Assume they always lead off an inning.  One hits .300 with all singles.  The other hits .258 but a third of his hits are doubles, the rest singles.  Which is more valuable (assuming they walk and strikeout at the same rate)

NFL is Back, and the Cardinals Still Suck

I enjoy many professional sports casually, attending an event or two every year, but the NFL is by far my favorite.  In the pre-season, there was a lot of hype that maybe the long-time hapless Cardinals would be decent this year.  I knew better, even from the pre-season.   Heck, my 8-year-old daughter knew better.

We went to see the last pre-season game against Denver.  In that game, the Arizona starters played for quite a while against the Denver 2nd team, and got beaten up.  Specifically, they could not run the ball and in turn their defense could not stop the run.  So it was no surprise to see them get blasted in their first regular season game against the Giants. 

The problem with the Cards is this:  They have spent the last several years drafting high-profile position players, including spending a jillion 1st round picks on receivers.  Great teams got that way because they invested in their lines - both O and D, even when such picks might be less popular with the fans on draft day.  The Cards have instead focused on drafting "names" who might help sell season tickets in the new stadium.  This neglect is very apparent today.  It doesn't matter how good your position players are if there are no holes for the backs and the QB is getting plowed to the turf on every play.  This is a 5-11 team that is fortunately playing in the NFL's worst division, so they may eek out 7 wins.  You heard it here first.

By the way, if you are an avid football fan, I recommend two sites to you.  The first is Football Outsiders, who have taken a Bill-James-like approach to football stats, rethinking metrics to provide a better insight into what teams really are good.  Make sure to check out their DVOA rankings - basically they compare every teams performance on every play against other teams in the same situation (e.g. 3rd and 8 on their own 45).  The other site is Greg Easterbrook's always entertaining Tuesday Morning Quarterback column, providing large doses of football clear thinking and haiku.

The Power of Metrics and Expectations

This is my first and probably last baseball post - read this blog if you want more baseball.

I am fascinated with the psychology of the closer position.  Some background:  The best baseball pitchers start games, and on average get through about 6 innings of 9.  The baseball manager's job is to stitch together a number of less talented pitchers to cover the 7th, 8th and 9th innings.  One would expect that the manager would flexibly match pitcher skills against the lineup he is facing.  For example, if the most dangerous batters for the opposing team are scheduled up in the 8th inning, he might send in his best relief pitcher in that inning.  One would not expect to see any particular emphasis on one inning or another:  after all, a game lost in the 7th counts the same as a game lost in the 9th.

This, however, is not how most managers operate.  Most managers have one very highly paid and more talented relief pitcher they call the "closer" that they pitch solely in the 9th inning.  Why?  Why is the 9th more important and deserving of a valuable player than the 8th?

The answer is part baseball conventional wisdom, which is as strong as in any old-line industry.  However, the other part of the explanation must lie in metrics.  If a manager loses a game in the 7th, it is just a loss.  If a manager loses a game in the 9th, the game was "blown".  Newspapers and talk shows keep and publish stats on games blown in the 9th, but not games lost in the 7th and 8th.  Games lost in the 9th are in a sense portrayed as more of a management failure than games lost in the 7th, and this is made worse by the fact that a game lost in the 9th is somehow more psychologically devastating for fans and media.  Managers are not dumb - recognizing that they get dinged on their performance rating more for a game lost in the 9th than the 8th, they have invented the closer role.  General managers take a disproportionately large part of their salary budget for relief pitching and dedicate it to this closer role.

A guy named Theo Epstein a couple of years ago, as a general manager, challenged this conventional wisdom.  He observed that more games were lost in the 7th and the 8th than the 9th, so hypothesized that relief pitching emphasis and salary dollars should be spread more evenly across the three innings.  One of his consultants was the famous Bill James, who has challenged baseball conventional wisdom with facts for years.  Epstein was roundly criticized by media and local fans alike for his "Closer by Committee" approach.  Eventually he was forgiven, when in the following year he brought his town its first world championship in 86 years.

For more on this and similar baseball topics, the book Moneyball is fabulous, and tells this story of the clash of fact-based analysis and baseball conventional wisdom, in a way that might be familiar to change agents in any number of Fortune 500 companies.

Interview with Bill James

If you were to make a list of 10 people in the 20th Century who had the ability to rethink whole industries, you might come up with names like Sam Walton or Herb Kelleher.  One guy you might not think of, but who should make the list, is Bill James.  James has helped to single-handedly rethink the game of baseball, one of the great bastions of not-invented-here thinking.  Here is an interview of James that is pretty interesting.  Hat Tip to Cafe Hayek, who also has some thoughts on James the economist.

James sounds a lot like Hayek, and more recent authors like Virginia Postrel, when he says things like this:

If I were in politics and presented myself as a Republican, I would be
admired by Democrats by despised by my fellow Republicans. If I
presented myself as a Democrat, I would popular with Republicans but
jeered and hooted by the Democrats.
        I believe in a universe that is too complex for any of us to
really understand. Each of us has an organized way of thinking about
the world"”a paradigm, if you will"”and we need those, of course; you
can't get through the day unless you have some organized way of
thinking about the world. But the problem is that the real world is
vastly more complicated than the image of it that we carry around in
our heads. Many things are real and important that are not explained by
our theories"”no matter who we are, no matter how intelligent we are.
        As in politics we have left and right"”neither of which explains
the world or explains how to live successfully in the world"”in baseball
we have the analytical camp and the traditional camp, or the
sabermetricians against the scouts, however you want to characterize
it. I created a good part of the analytical paradigm that the
statistical analysts advocate, and certainly I believe in that paradigm
and I advocate it within the Red Sox front office. But at the same
time, the real world is too complicated to be explained by that
paradigm.

Or this, closer to the sports world:

Honestly, major league baseball"”and all sports"”would be far better off
if they would permit teams to do more to make one park distinctive from
another"”even so far as making the bases 85 feet apart in one park and
95 in another. Standardization is an evil idea. Let's pound everybody
flat, so that nobody has any unfair advantage. Diversity enriches us,
almost without exception. Who would want to live in a world in which
all women looked the same, or all restaurants were the same, or all TV
shows used the same format?
        People forget that into the 1960s, NBA basketball courts were
not all the same size--and the NBA would be a far better game today if
they had never standardized the courts. What has happened to the NBA
is, the players have gotten too large for the court. If they hadn't
standardized the courts, they would have eventually noticed that a
larger court makes a better game"”a more open, active game. And the same
in baseball. We would have a better game, ultimately, if the teams were
more free to experiment with different options.
        The only reason baseball didn't standardize its park
dimensions, honestly, is that at the time that standardization was a
dominant idea, they just couldn't. Because of Fenway and a few other
parks, baseball couldn't standardize its field dimensions in the
1960s"”and thus dodged a mistake that they would otherwise quite
certainly have made.
         Standardization destroys the ability to adapt. Take the high
mounds of the 1960s. We "standardized" that by enforcing the rules, and
I'm in favor of enforcing the rules, but suppose that the rules allowed
some reasonable variation in the height of the pitching mound? What
would have happened then would have been that, in the mid-1990s, when
the hitting numbers began to explode, teams would have begun to push
their pitching mounds up higher in order to offset the hitting
explosion. The game would have adapted naturally to prevent the home
run hitters from entirely having their own way. Standardization leads
to rigidity, and rigidity causes things to break.

I love it.  Maybe those guys who want to use baseball as a paradigm for life had something after all.

Rethinking Football Metrics

I find that most experienced managers have become experts at identifying and gaming flaws in measurement systems. The in and outs of measurement systems have always interested me, both in business and in sports (how about that segue-way?)

Those of you who are baseball fans may be familiar with Bill James. Bill James came to the conclusion that baseball stats really didn't say very much about what went on in a game, and were misleading in evaluating individual performance. He and people like him have asked questions like "is RBI production really a fair measure of individual performance (since it depends on teammates getting on base)" and "why are walks left out of traditional hitting stats". My post is really on football, but if these baseball questions interest you, check out the book Moneyball.

Much like these baseball stat pioneers, there are a number of people trying to rethink football statistics. For example, is total yardage given up a good measure of defensive productivity? Won't a mediocre defense on a team with a great offense that grinds out 8 minute drives sometimes look better on this stat than a good defense on a team with an offense that is always 3 and out? A site called Football Outsiders is one example of the search for better football understanding. If you are numerically inclined, and are tired of the "its all about execution, about taking it one game at a time" football analysis, check these guys out.