MLB : How much does a win "cost"?

Manny Acta, coach of the Washington Nationals, was fired early today. The Nationals are currently last in the National League East and have the dubious honor of sporting the worst record in baseball: nearly 10 games back of the next worst team and on pace for a pathetic 50 wins or less. Not good, to say the least.

From the article,

"We feel that the team has underachieved," [General Manager Mike] Rizzo said. "We feel we have a better ballclub than we've shown on the field."

Easy to say in the heat of the moment, when you're looking for a scapegoat or a reason for having dumped Acta. But, is it true? Has the team underachieved?

Washington has many problems: the worst ERA in the National Leage (5.21), the most errors committed in baseball (82), a rotation full of rookies, they rank at the bottom of most offensive categories, etc. But, in my opinion, most of these problems are simply the result of having one of the lowest team salaries in the majors: around $59 million, 27th out of the 30 teams. And this begs the question: what is the relationship (if any) between team salary and wins?

We can find team salaries these days readily available on the internet and plot that against the current number of wins for each team. The result looks like this:


Several assumptions need to be listed and justified. First, the Yankees and Marlins are notorious outliers but for competing reasons: the Yankees are willing to spend nearly twice as much as its nearest competitor (always have to stay ahead of the Red Sox) and the Marlins have the ability to do more with less. After removing these two, we're left with 15 NL teams and 13 AL teams. If we're going to compare leagues, it would help to remove two NL teams: the most obvious choices are the Cubs and Mets, both grouped together and around $20 million higher than the Phillies. I'll show at the end that the removal of these four teams leads to little change in the correlation for the data. Last, I chose to run the trend line through the origin: if you didn't spend a dime, you'd have no team and win zero games. The result is a trend line that fits the data very well for 26 of the 30 teams.

Any measure of performance should, by definition, include roughly the same number of teams "over-performing" and "under-performing". By "over-performing", we mean that the team is getting more wins than the team salary should suggest; by "under-performing", we mean that, for this salary, the team might be expected to have more wins. Thus, investigating the distribution of AL and NL teams above and below the trend line, we see:


Seven AL teams above the line (over-performing) and seven AL teams below the line (under-performing); a note about the Yankees later. Nine NL teams are over-performing, one is on the line and six are under-performing. So, the teams are spread fairly evenly on each side of the line; the NL is slightly over-performing, but we'll see in the next graph that this is mostly because of one division (the NL West).

Finally, we look at the six divisions of MLB. Do all teams in a division follow a similar plan? Or are any of the divisions being won by a team willing to pay a higher salary? Grouping the teams by divisions looks like this:


If we count the number of over-performing teams in the AL, we find three (out of five) in the East, two (out of five) in the Central and 2 (out of 4) in the West. Meaning, the AL is not just evenly divided as a league, it's also fairly even within each of the divisions; in other words, in the AL, you get what you pay for. The exception, of course (and when is it not), is the Yankees whose salary is so outlandishly high they would have to have a record of 100 wins and negative 20 losses to sit on the trend line; clearly, the Yankees are proof of the Law of Diminishing Returns (LoDR), you can't buy more wins than the number of games you play.

In the NL, however, the situation is different. Counting the number of over-performers, we find one (out of five) in the East (the Marlins*, of course), four (out of 5) in the Central and five (out of five) in the West. The West makes for certainly the most interesting division right now in baseball, with no team spending more than $100 million but three of the five on pace for 90+ wins and all five with a mathematical chance at a winning record.

Returning to the LoDR, we see all six division leaders grouped together near the $100-$110 million range (boxed section), with around 50-55 wins at the All-Star break. Thus, while one would expect there to be a minimum amount of payroll needed to compete (unless you have the best scouts in the majors like the Marlins), this data appears to support the idea that there's also a maximum payroll; beyond $120 million, you don't really get any extra value and may be filling your team with stars that can't find playing time and egos that clash. Smart money would say, emulate the Cardinals, whose $89 million salary (13th highest and nearly a third of the Yankees) has them ~50 wins and the lead in the Central. Also don't overlook: the Rays, the Rangers, the Giants, the Rockies (who also fired their manager) and the King of More With Less, the Marlins.

But, the question remains: were the Nationals under-performing? Not even close. In fact, they fall nearly perfectly on the trend line: 26 wins is about what you should expect with a salary of $59 million.

And how strong is the correlation between salary and wins (where 1 means very large positive correlation and 0 means no correlation)? If we exclude the four outliers, the correlation is around 0.65; with the outliers, around 0.45 -- both of these values suggest a strong correlation between salary and wins. So my advice to the Nationals owners and GM would be: first, lay off Acta, he wasn't the problem (or, at worst, one in a sea of many); second, open your wallets. $59 million these days is merely the ante, you better chip in another $20 or $40 if want to still be sitting at the table come October.

UPDATE: Originally, this said "the Braves" by mistake.

Read More......

Unemployment and the real culprit

Since the spring of last year, the unemployment rate has risen from less than 5% to 9+%; millions of Americans out of work -- 13.7 million to be exact -- is obviously dire news. So, when last month's numbers showed a drop in layoffs, one might be inclined to think the economy was beginning to turn around. But, the problem is, we're looking at the wrong data.

I have always seen the current economic downturn from the point of view of a "headhunter". In my job, I am surrounded by undergraduate seniors (and after them, even more students waiting their turn) and by this point in the year, most students will have gone through the hiring process with several companies to ensure a job upon graduation. University departments crank out tens or (in the good days) hundreds of graduates every year, regardless of the economy, because people recognize that these days most jobs require some college experience. Thus, every semester, new applicants join the job pool whether the industry is hiring or not.

Lately, as the department looks for jobs for these students by planning career fairs or inviting speakers, we hear the same rejoinder over and over: We're not hiring right now. Typically, this means "we're done with the bulk of our layoffs, but we're not economically sound enough to hire additional personnel right now." Companies instead spread around the existing personnel into departments or jobs the employee was not necessarily hired for in an attempt to ride out the downturn or new avenues of research are suspended and instead the focus shifts to existing product lines. These are fine practices since it means people keep their jobs but it means the applicants start piling up.

On that note, we can turn to the U.S. Department of Labor. At the bottom of the page you can read the latest statistics released (i.e. consumer price index, unemployment rate) or search through the massive archive of data. One interesting number to look at is the number of unemployed workers per job opening (UPJO). This can be calculated by combining the data from the following spreadsheets: job opening rate, unemployment and employment totals. In the end, we can produce the following graph:


Thus, we've gone from approximately 1.5 theoretical applicants per job (if everyone wanted to apply) back in 2007 and 2008 to currently nearly 5 applicants per job opening. In other words, while the unemployment rate has nearly doubled over the past year, the UPJO rate has fully tripled. Why? A decline in job openings. Of course, one might argue the situation is worse in specific jobs (i.e. manufacturing, construction) and better in others (nursing). Similarly, one might argue whether the downturn has hit men worse than women (read: unintended glass ceiling effect). But what is clear: in general, job openings are disappearing.

How does the U.S. turn this around? We should focus on encouraging both businesses to hire and citizens to invest in those companies. Instead, the current administration has decided to employ cap-and-trade schemes (which inevitably act as regressive taxes on the poor), constrain venture capital (out of some misguided attempt at "fairness") and refuse repayment of TARP money (so the Federal Government can maintain control). The end result is that companies are either government-controlled and therefore unable to dictate hiring practices (in GM's case, they were forced to fire thousands of employees and shutter dealerships) OR they are too scared or simply unable to bring on new employees out of uncertainty as to what the future economy will look like.

In the meantime, the applicants are piling up.

Read More......

The trouble with computer models

From a story today in the L.A. Times:

"We're looking at a scenario where there's no more agriculture in California." And, [Chu] added, "I don't actually see how they can keep their cities going".

Strong words indeed: the collapse of agriculture in California, the downfall of cities, water shortages in the Midwest and so forth. Perhaps we really should look at alternative energy sources if all of this is going to happen. And the proof for this collapse? Computer models.

Computer models. Ugh. This is my life every day, trying to simulate the phenomena we can't exactly perform as an experiment in a lab. Things that go on at the nanoscale and, ironically enough, macroscale levels about whose inner workings we're (as scientists) unsure. Why is the universe expanding? What causes gravity? What is the conscious "me" inside of me? The truth is, we don't really know some of the answers to Life's questions -- that's another question we could ask, why don't we know everything? -- so, what we don't know, we estimate. We guess as best we can.

And since no one has any idea what the year 2100 is going to be like (much less the weather next week), they have to "simulate" it: essentially, they create a model of the world (or just the U.S.) and estimate the factors effecting this piece of land and they run it, simulating seconds, months, decades even. The problem is: it doesn't always work. And I say that, knowing full well that this is an indictment of my own PhD in computer modeling.

The problem is, as Douglas Adams would certainly find amusing, to simulate something as complex as the Earth's weather, to factor in everything that determines tomorrow's rain or shine (e.g. spin of the Earth, solar activity, atmosphere), it would take a computer with the complexity of, well, the Earth. Of course, we don't have the time or space on Earth to build the Earth, so we neglect some things. Accidental spill of iron into the Atlantic in the year 2047? We have no clue that's gonna happen, leave it out. Europeans stop having children, altering our estimates of future population? Possible, but we can't count on it, so leave it out. You get my point: the only way we can be sure what 2100 will be like is to, well, live until 2100 and see.

But that doesn't sell when you're trying to get funding for your project. If I tell my funders that I think there's a 50/50 chance of my research producing uninteresting results, you bet I'm not going to get many backers. If, instead, I tell them I bet I can produce colloids that will not only cure all of the world's problems but also smell like cinnamon, watch the money roll in. So then how do I simulate success? I program it in. When I estimate the factors affecting my colloids or polymer formations, I try only the situations that I predict will be successful; why waste time on the trivial non-success? Lack of success doesn't sell papers (or journals for that matter). Climate models are just as guilty. If predictions show that 2100 will look kinda like 2009 -- a little overcast with 30% chance of precipitation -- would Congress invite them to come speak? Of course not.

So they factor in change. They factor in "something happening". But, most importantly, they always factor in something drastic or catastrophic happening. Results which are less than exciting or, heaven forbid, show a reverse effect (say global cooling) are tossed out. And if one model says the Earth's temperature is going up 2 degrees and another says 12 degrees, then you report the 12 degree one; that gets people's attention. Now, the point is, it doesn't matter whether the model shows 12 degrees of heating or 1.9 or zero or, as six models at the link above, shows actual cooling, the point is: they don't agree with each other. Why? Because it's all a guess. Yet, government wants you to fear something drastic could happen because you'll pay them to keep you "safe" via more taxes. But don't buy it. Educate yourself and don't buy into the myth of computer models "predicting" future weather.

Read More......

Alice in the Cities: Possible Itinerary

[This is a continuation of the Alice in the Cities series, part one is here.]

In a previous post, I listed (in film chronology) the spots which Wim Wenders filmed along the east coast of the U.S. (or at a minimum, my best guess). We can also look at this pictorially by creating a Google Map of the trip. This is below the fold.

View Larger Map

By plugging in all the spots pictured in the film (some more positively ID'd than others), we can see a clear pattern. Wenders and Vogler (who played Philip Winters) more than likely landed in New York. They began to drive the east coast, maybe with only Robby Mueller as crew, looking for "American" places to film. They were essentially filming without a script. From DVD Times,

Freeing himself also from any strict script, Alice in the Cities then grew along with the relationships between its characters, largely improvised and scripted as they journey through the movie.

There's also evidence of this in the film itself. At one point, Winters is staying in Manhattan with a friend as he waits for the flight back to Germany/Amsterdam. He tries to explain the trip he has just taken to the woman:

Soon as you leave New York everything looks the same. . . All I could imagine was things going on and on. Sometimes on the road I was sure I would turn back next morning. But still I went on and on and listened to that sickening radio.

Winters is probably channeling what Vogler felt as he rode in the car every day with Wenders, listening to the "blasted" radio, staying in less-than-stellar motels, thinking to himself each morning: Maybe he'll [Wenders] want to turn back today.

Returning to the itinerary. Viewing it above as such does lend credence to the possibility of Everetts, NC being the home of Chester's Grocery. It would have been along the route of a person driving HWY 17 from Surf City, NC to the Richmond (VA) area. At a minimum, it suggests that Chester's is probably along the coast like the other spots, since there's no reason to believe at this point that Wenders deviated much from this road. It would make a natural North-South path along the coast and would provide cheaper restaurants and motels for the two to visit.

One drawback to plotting the trip on a map, though, is that it contradicts the above quote from the film: "Soon as you leave New York everything looks the same." At some point, the trip more than likely took them through Washington, D.C. The trip probably went through Philadelphia. To say that these places look no different than Low Country, South Carolina is absolutely absurd. But these spots are absent from the film because they negate the message that Wenders was trying to convey with the opening sequence of Alice in the Cities: that Winters could not write because he was changed by the homogeneity of America outside of NY.

My discussion on this topic (homogeneity of the U.S., part 3) to follow as soon as possible.

Read More......

Alice in the Cities: Unknown locations

[This is a supplemental to the Alice in the Cities series, part one is here.]

Two locations in the opening segment of the film are difficult to place geographically: the diner and the gas station. The former, the diner, suffers from lack of signs (you are never shown the name of the restaurant you're inside); the latter is difficult due to the fact small-town gas stations from the 1970s (most likely locally owned) rarely make the internet.

First, the diner. After getting a drink at the counter, Winters looks out the open front of the diner to see one building on the right with garage doors and a building on the left with "EPUBLIC SHOE CORP" (sic).

One possible location, (h/t Emily), is Far Rockaway, NY where the Republic Shoe Corporation was founded in 1957. Certainly the building in the movie could have been moved by now but it might make sense that this shot is in NY or NJ near where the company was founded. Suggestions will be appreciated.

The second difficult location is the gas station. Winters stops for gas and then walks out the road (heavily trafficked behind him) to take a picture of the store front.

We're shown the picture that Winters takes (above) with his car at the pumps and a boy he talked with to the right: Whatchu takin' pictures for? The owner won't like it. One possible longshot is Everetts, NC which is referenced in this book, Hugh Smithwick's Descendants. Everetts sits right on HWY 17 in eastern NC and would be on the way from Surf City, NC to the Richmond, VA area for someone travelling the backroads or HWY 17. Again, suggestions are welcome.

A Google map of the itinerary has been added now, follow the link here.

Read More......

Alice in the Cities: U.S. Connection (Part 2)

[This is a continuation of the Alice in the Cities series, part one is here.]

In this part, I will show screen captures from the 1973 German film Alice in the Cities (Wenders). The movie opens with 20 minutes of "on-the-road-in-the-U.S." footage of which roughly the first 7 minutes are most relevant to us. Wenders, like his lead character Philip Winters, is essentially driving through the U.S., taking footage of what he sees out the window of the car.

The scenes happen in the following order in the film -- note, this is not necessarily (and probably not likely) the order in which the scenes were shot.

1. Opening scene is of a plane in the sky pulling down to a road sign near the beach and a jut of rocks into the ocean.

The sign says "67th St.", the preceding "B" meaning "beach" (the shot of the ocean backing this up). My guess is this is near Rockaway in Queens. Later in the movie, Philip sells his car in the shadow of Shea Stadium not too far away from this likely site. A link to Google maps can be accessed here. Also, Google has linked a photo someone took on these rocks of the waves crashing in.

2. Driving through a small town, Winters stops the car to take a picture, then turns to the right. Seemingly in the same town, there is a close up of a water tower labeled "Surf City".

"Surf City" is a real place: Surf City, NC. Searching for "Ward Realty" delivers the site Ward Realty Corporation, in business since 1951 on Topsail Drive in "downtown" Surf City.

3. View through the front window of a motel named "The Driftwood", then another motel named "Roxanne" in front of which Winters turns the car to the right to pull up next to the ocean.

These two motels are still in business today on the strip in Myrtle Beach, SC. The Driftwood started 75 years ago and can be found at 1600 N. Ocean Blvd; it sits right next door (1604 N. Ocean Blvd) to the Roxanne. 35 years ago, Wim Wenders pulled up between these two motels and took a picture (above right) that set me off on this project. A kid growing up in South Carolina needs no sign to tell him he's looking at Myrtle Beach.

4. A scene in a diner, location unknown. 

5. Driving down a road, presumably in the Low Country. There are moss covered trees and then a sign for HWY 41.

My best guess is that this is the Low Country, SC, outside of Charleston -- mostly due to the moss-covered trees and plantation-style homes in this part of the film. There's a point along HWY 17, near Mt Pleasant, where HWY 41 comes to join HWY 17 (above left reads "JCT") which would have been the major road of travel in those parts 35 years ago.

6. A scene at a gas station, Chester's Grocery & Gas, location unknown although probably on a main thoroughfare since large trucks are using the road. 

7. Driving through the next town, Winters reads the call letters from a sign near a building, "Channel 6 W-T-V-R", he says.

This is a very easy find -- WTVR Channel 6 in Richmond, VA. 

8. It is now night time, Winters passes under a sign on the interstate for Emmet St. and then (after a cut in filming) spies a motel called the Skyway Motel.

A Google search for "Emmet St." (above left) brings up about 10 possibilities only two of which are remotely close to the east coast: Daytona, FL and Essex, NJ. Searching for "Skyway Motel" (above right) shows there's one near Jersey City, NJ about 4-5 miles from Emmet St., Essex, NJ. Both are right off HWY1 and I-95 so this could very likely be its location.

A supplemental section is next, containing two locations I could not place (at this time).

Read More......

Alice in the Cities: Introduction (Part 1)

I have not blogged in a while, I know, thanks for returning. I've been hesitant to blog as I apply for jobs but I also want to get this post in before I forget too much.]

A quick background is required: One of the foreign films that I was lucky to see as a student in college was the 1973 Wim Wenders' road movie, Alice in den Staedten (Alice in the Cities). It was part of a 300-level German class, including other gems such as Das Boot, Der Himmel über Berlin (Wings of Desire) and Lola Rennt (Run Lola Run). Wenders is a well-known German filmmaker, with over 40 years of films and plenty of awards. In 1973, he began a series of road movies -- Alice in the Cities, The Wrong Move, and Kings of the Road. The first installment (Alice) centers around a photojournalist/writer Philip Winters (played by Rüdiger Vogler) who has a serious case of writer's block. He has been commissioned to write an article about America and, in preparation for that piece, drives around the U.S. taking Polaroid pictures. He is distraught, however, when the pictures develop -- they never seem to show the real images he was looking at. In the end, he can not finish the writing in the States so he sells his car to afford a ticket back to Germany (after a detour in Amsterdam because of an airline worker strike) where he promises his friend he will finish the article.

When I sat down in front of YouTube on Monday to watch this film, I had no idea that I would discover that the places and pictures Wenders cataloged in the opening section of the film would be so close to home -- South Carolina, North Carolina, Virginia, etc. In part 2, I'll attempt to prove what I feel are known spots along the east coast, specifically in the Southeast, by comparing screen captures to information readily available to us now online. In part 3, I'll discuss what current literature says about this Wenders piece and how it may differ for those of us who grew up in the Southeast.

Part 2 is posted here.

Read More......

A quick primer on Subprime

This has to be my favorite introduction/primer to the current economic crisis. For one, because it's essentially non-partisan and secondly, because it's so funny. Link comes after the fold and contains some expletives so consider this your warning.

Read More......

Early voting nearly here

For the campaign season that seems to have gone on forever, the end is nearly here. Early voting has begun in some states and we're still six weeks out. Those of us in NC have a few days waiting left: in general, we can start on Oct 16th. Guidelines and information can be found here.

Whenever I do cast my ballot (and I'm nearly certain it will be early), I will be ecstatic. Not because of the historical moment I'll be participating in: the first election of a minority if either party wins -- I could less about this pointless narrative that partisans want to thrust upon us. No, I'll be happy because of the sweet relief and the sound of silence.

Read More......

How a hurricane changed the MLB season

From August 1st to September 11th, the Houston Astros weren't just good, they dominated MLB by posting a 30-10 record. Over the two weeks leading up to September 11th, they lost only one game (a close one on the road in Colorado), going 14-1. They were screaming up the standings in the NL Central, blowing away the competition. Then Hurricane Ike came, and it blew away any chance the Astros had.

On September 13th in the pre-dawn, the atmospheric wave that left Sudan nearly a month earlier slammed into the Texas shoreline as the strong Category 2 hurricane known as Ike. It devastated Galveston, whose residents are only now returning. Ike continued through Texas, across Houston, up into the rest of the U.S., to dissipate a few days later. In its wake, homes, businesses and lives were destroyed.

In Houston, the red-hot Astros were forced to find a new home for their upcoming baseball games. After playing (and winning) on Thursday the 11th at home against the Pirates, they hit the road, driving to Wisconsin. The first two games (Friday and Saturday) of a series against the Cubs were postponed. On September 14th, they stepped on their "home" turf of Milwaukee Miller Park, having not played in three days, three days mostly spent moving an organization countless miles north. Zambrano and the Cubs promptly no-hit the Astros, the first neutral site no-hitter ever.

It would only be the beginning. The Astros dropped their first five games post-Ike and all told went 2-7 since September 14th. The day they hit the road for a series of home games in Wisconsin they were 3 games out of the Wild Card in the NL and, having gone 14-1, were probably the best bet in the league to win the spot. Two weeks later, they sit 4 1/2 games back with no room for error: another loss or a win by the Mets and the post-season is out.

What happened? Owner Drayton Mclane stated that it wasn't the change of venue, but the idea of leaving behind homes and loved ones. Rather than focus on the Cubs, they were worried about the devastation happening at home. The outcry was so loud, MLB took out a full-page ad in the Houston Chronicle explaining to Astros fans that there was really no other option. You can read it for yourself at the link, but the thrust of the arugment is that there was no other option: they wanted a domed stadium that was available (no conflict of schedule) and they wanted to keep in somewhere from Texas east. The result: Milwaukee. Of course, with millions of Cubs fans next door, it was bound to cause controversy. And in the end, it probably was the reason the Astros collapsed, the timing and the change of venue.

Houston played their first home game in weeks last night. Fittingly, they lost to the Reds. Surely, they've given up a little bit, soured by the whole experience from Ike to Milwaukee. Could MLB have held onto those games until a later point in the season? If they could delay one game until September 29th (after the regular season), why not a double header? Why not make an exception for the Astros and wait until the end of the season to see if those game would have even been necessary to decide the standings. Instead, MLB made its decision to send the Astros packing and the rest is history. And we the fans are left to wonder what if.

Read More......