Bill Petti

New Year, New Writing Gig

Just wanted to let everyone know that starting January 4th I will be writing a weekly baseball column (sometimes twice weekly if I am feeling especially opinionated) at Beyond the Box Score.

Beyond the Box Score is a fantastic site, examining baseball from an analytical perspective.  The authors definitely embrace sabermetrics, but they don’t beat readers over the head with complex statistics.  As with most things that I do, the subject of my columns will vary quite a bit.

Generally speaking I’ll likely focus on team performance, player valuation, and lots of exploratory questions about the game.  Oh, and you can be sure there will be lots of pretty visuals and laments about the NY Mets.

Be sure to stop by if you are interested.  You can read and subscribe to my entries here, but I encourage you to subscribe to the site as a whole (RSS feed here).

PoliSci-unrelated post of the day: Visualizing Major League Baseball, 2001-2010

This post originally appeared at Beyond the Box Score.  If you are a baseball analysis fan and don’t already read BTBS I highly recommend it.

2010 marks the end of the “ought” decade for Major League Baseball.  I thought I would take the opportunity to analyze the last 10 years by visualizing team data.  I used Tableau Public to create the visualization and pulled team data from ESPN.com (on-field statistics) and USA Today (team payroll).

The data is visualized through three dashboards.  The first visualizes the relationship between run differential (RunDiff) and OPS differential (OPSDiff) as well as the cost per win for teams.  The second visualization looks at expected wins and actual wins through a scatter plot.  The size of each team’s bubble represents the absolute difference between their actual and expected wins.  Teams lying above the trend line were less lucky than their counterparts below the trend line.The final tab in the visualization presents relevant data in table form and can be sorted and filtered along a number of dimensions.

The first visualization lists all 30 teams and provides their RunDiff, OPSDiff, wins, and cost per win for 2001-2010.  The default view lists the averages per team over the past 10 years, but you can select a single year or range of years to examine averages over that time frame.  The visualization also allows users to filter by whether teams made the playoffs, were division winners or wild card qualifiers, won a championship, or were in the AL or NL.  The height of the bars corresponds to a team’s wins (or average wins a range of years).  The color of the bars corresponds to a team’s cost per win–the darker green the bar the more costly a win was for a team.  Total wins (or average for a range of years) is listed at the end of each bar.  In order to create the bar graph I normalized the run and OPS differentials data (added the absolute value of each score + 20) to make sure there were no negative values.  For the decade, run differential explained about 88% of the variation in wins and OPS differential explained about 89% of the variation in run differential.

The visualization illustrates the tight correlation between RunDiff and OPSDiff, as the respective bars for each team are generally equidistant from the center line creating an inverted V shape when sorted by RunDiff.  In terms of average wins over the decade, there are few surprises as the Yankees, Red Sox, Cardinals, Angels, and Braves round out the top 5.  However, St. Louis did a much better job at winning efficiently, as they paid less per win than the other winningest teams (<$1M per win).

5283757704_65097bb2f0_z_medium
(click for larger image)

The viz also illustrates the success of small market teams such as Oakland and Minnesota who both averaged roughly 88 wins while spending the 3rd and 4th least respectively per win.  If you filter the visualization for teams that averaged over 85 wins during the decade, it really drives home how impressive those two teams’ front offices have been at assembling winning ball clubs with lower payrolls.  No other team that averaged >85 wins paid less than $975K per win.  Oakland looks even more impressive when you isolate the data for years that teams qualified for the playoffs.  Oakland averaged 98.5 wins during seasons they made it to playoffs, and did so spending only $478K per win.


5283158827_5cec5c321a_z_medium
(click for larger image)
What about the big spenders?  The five biggest spenders included the Yankees, Red Sox, Mets, Dodgers, and Cubs.  The Yankees spent an astounding $1.8M per win during the decade, but they also averaged the most wins with 97.  Some will say this provides evidence that the Yankees–and other big market teams–simply buy wins and championships.  However, only 17% of the variation in wins was explained by payroll during the decade.  Moreover, while the Yankees occupied 6 of the top 10 spots in terms of cost per win they were the only team to earn a positive run differential.  The Cubs, Mets, Mariners and Tigers all finished under .500 and missed the playoffs while those Yankee teams qualified for the playoffs 5 out of 6 years and won one World Series.  Yes, the Yankees spend significantly more per win, but they spend more wisely than many other deep pocket teams.
Teams that made the playoffs averaged a little over $1M per win in those years they qualified, with Wild Card teams ($1.030M) spending a tad bit more than Division winners ($1.006M)–about $14K per win on average.  World Series winners spent $1.08M per win in their winning years compared to $1.002M for other playoff teams.  Teams that failed to make the playoffs averaged $923K per win.
The best team of the decade in terms of run differential?  The 2001 Seattle Mariners, who amassed an incredible +300 RunDiff.  Even with that total they were only expected to win 111 games–they would go on to win 116.  The Mariners had only the 11th highest payroll that year and so paid a measly $644K per win.  The absolute worst team of the decade?  The 2003 Detroit Tigers, who earned a RunDiff of -337 and actually won less games than expected (43 vs. 47).  Given their ineptitude on the field, the Tigers paid $1.14M per win even though their total payroll for the year was only $49M.
Luckiest team?  The 2005 Diamondbacks who won 77 games despite a RunDiff of -160 (only 64 expected wins).  Hardest luck team?  The 2006 Indians, who only won 78 games with a +88 RunDiff that should have translated into 90 wins.
5283158811_90d96e0457_z_medium
(click for larger image)

There are tons of ways to manipulate the visualizations and cut the data.  Hopefully viewing the data in this way is helpful and illuminates some things we didn’t know and drives home other things we had a hunch about. This is my first attempt to visualize this data, so please feel free to send along any and all comments so I can improve it.

Author’s Note: Due to a very helpful comment by Joshua Maciel, I have updated the visualization.  Here is a link to the original version for those that are interested.

Structural explanations are not always sexy or gratifying, but they typically explain a lot

In the days after the US midterm elections cable news outlets, radio programs, political pundits, newspapers, and activists on both sides of the ideological spectrum have exerted a great deal of blood and sweat to explain the nationwide drubbing of the Democrats. Democrats are predictably covering their behinds—conceding voter anger, but cautioning that the country has not lurched to the right in just two years. Republicans are claiming validation of their position and a greater ideological alignment with the American people. Activists and enthusiasts of all stripes are weaving narratives that use the election results to validate their personal political perspective. The question, of course, is whether any of this is correct or meaningful. Was this election a mass repudiation of Democratic policies? Was it a validation of the Republican platform and/or Tea Party-style conservatives?

Elections are like Rorschach bots—everyone sees something different, and often times what they see is what they want to see. Particularly with elections, people like to place causation in the hands of people—agents—whose efforts, words, thoughts, etc, drive the outcome. And to be sure, individual agents can and do wield a great deal of influence on events. But an overemphasis on agents can lead to spurious conclusions about why something happens. You must also look at structural or environmental factors.

Over at the Monkey Cage, John Snides has a great piece precisely along these lines. Snides and his colleagues looked at which factors where the best predictors of voter choice:

If you had one thing, and one thing only, to predict which Democratic House incumbents would lose their seats in 2010, what would you take? The amount of money they raised? Their TARP vote? Their health care vote? Whether they had a Tea Party opponent? A Nazi reenactor opponent?

Not surprisingly, it’s none of those.

As is typically the case, the partisan makeup of a politician’s district mostly determines which candidate will win.  Snides and his colleagues found that the 2008 Presidential vote in a district explained 83% of the variation in the 2010 vote share (see graph below).

This data does not negate agent-centered factors, but it certainly dulls them.  Additionally, many of the theories being thrown about (the vote was a referendum on Obama, on Democrats, on “Big Government”, etc) just don’t have the explanatory power that the partisan makeup of a district has.

What’s clear is that, structurally speaking, the Democrats were set up for a shellacking.  Historically, the President’s party takes a big hit in the midterms, incumbents are punished in a poor economy (regardless of their control over it), and incumbents in swing districts will be the first to go.  Many of the seats Democrats gained in 2006 and 2008 to take a commanding majority in the House were obtained by targeting vulnerable Republicans in swing districts.  Conservative Democrats ran and won in those districts, meaning they faced a center-right electorate.  Given these structural factors, it is no surprise that the Democrats lost so many seats.

Structural explanations are not very sexy.  They don’t allow a ton of room for debate and analysis after the initial work is done.  By their nature, there isn’t a whole lot that can be done to alter the conditions (i.e. a reduced role for agency).  And they don’t really allow people to indulge in great philosophical and ideological satisfaction.  But, at the end of the day, they can be powerful explanations.  Democrats in 2006 and 2008 were overzealous in their interpretation of what those election results implied, and the same may happen to Republicans in 2010.  Savvy politicians and operatives should take heed.

[Cross-posted at Signal/Noise]

Book Review: Codes of the Underworld

I recently finished Diego Gambetta’s Codes of the Underworld: How Criminals Communicate.  For those looking for a more academic take on signaling (particularly from a sociological point of view) it’s a great find.  As I previously mentioned, Gambetta uses the extreme case of cooperation amongst criminals to tease out more general dynamics of trust, signaling, and communication.  The Mafia can be considered a “hard-case” for theories of signaling trust; given the extreme incentives for criminals to lie and the lack of credibility they wield given the very fact that they are criminals, how is it that criminals manage to coordinate their actions and trust each other at all?  By understanding how trust works in this harsh environment we learn something about how to signal trustworthiness in broader, less restrictive environments.  As Gambetta notes:

Studying criminal communication problems, precisely because they are the magnified extreme versions of problems that we normally solve by means of institutions, can teach us something about how we might communicate, or even should communicate, when we find ourselves in difficult situations, when, say, we desperately want to be believed or keep our messages secret.

The book is a great example of studying deviant cases or outliers, particularly when the area of study is not well worn.  This is a valuable general methodological lesson.  We are typically taught to avoid outliers as they skew analysis.  However, they can be of great value in at least two circumstances: 1) Generating hypotheses in areas that have not been well studied and 2) Testing hypotheses in small-N research designs, where hard cases can establish potential effect and generalizability and easy cases suggest minimal plausibility.

Gambetta takes a number of criminal actions and views them through the lens of signaling.  This allows readers to see actions, in many cases, in completely new ways, highlighting the instrumental causes of behavior.  For example, Gambetta looks at how criminals solve the problem of identifying other criminals by selectively frequenting environments where non-criminals are not likely to go.  Since criminals cannot advertise their criminality, they face a coordination problem.  Frequenting these locations acts as a screening mechanism since only those that are criminals are likely willing to pay the costs to frequent these locations.  (This ignores the issue of undercover law enforcement, but Gambetta deals with that as well).  Gambetta also makes the reader look at prison in a new light.  Criminals derive a number of advantages from serving time in prison, not the least of which is providing them with a signaling mechanism for communicating their credibility to other criminals (as prison time can be verified by third parties).  Additionally, many criminal organizations will require that new members have already served time before they are allowed to join.  Moreover, Gambetta explores how incompetence can work to a criminal’s advantage, since it can signal loyalty to a boss who provides the criminals only real means of income (a topic I discussed here).

Gambetta also looks at the conspicuous use of violence within prisons.  This isn’t a new topic, as any law enforcement drama will undoubtedly portray the dilemma of a new inmate who must establish their reputation for toughness and resolve or else suffer constant assaults by other inmates.  However, Gambetta makes it interesting by embedding the acts in a signaling framework.

First, Gambetta’s hypothesis regarding the importance of non-material interests is borne out by various studies.  Among others, he cites one study of prison conflict that found:

“[n]on-material interests (self-respect, honour, fairness, loyalty, personal safety and privacy) were important in every incident.”  While only some violent conflicts occur for the immediate purpose of getting or keeping resources, all of them have to do with establishing one’s reputation or correcting wrong beliefs about it.  Even “a conflict that began over the disputed ownership of some item could quickly be interpreted by both parties as a test of who could exploit whom.”

Second, Gambetta hypothesizes that we should expect to see more fights when prisoners do not have enough of a violence track record when they first arrive in prison.  One observable implication of this is higher rates of prison violence among female prisoners and younger prisoners.  In fact, the empirical record bears this out quite nicely.  Rates of violence are inversely related to age, providing ” a plausible social rather than biological explanation” for youth violence.  Additionally, Gambetta finds that, although less violent in the outside world, “women become at least as violent and often more prone to violence than men”.  Interesting, women are less often convicted of violent offenses, suggesting that the results are not simply the result of selection effects.

Both points have implications for political science and international relations, given the growing use of signaling models to explain political behavior.  The issue of reputation in international relations is one that is still growing and Gambetta’s hypothesis about lack of “violence capital” fits right in to much of the current work in conflict studies.

Overall, Codes of the Underworld is unique and thought-provoking work.  For those with a strong interest in communication and signaling, it is a must read.

[Cross-posted at Signal/Noise]

U.S. Midterm Election Prediction Fest 2010

At Gallup, we are officially predicting–regardless of turnout level–at least 40 seats for Republicans.  Based on the numbers and our historical model, Republicans should land about 60+ House seats, easily gaining the majority.

Personally, I’ll say 65 just to be (arbitrarily) specific.  I’ll also predict that Republicans pick up 7 seats in the Senate, 3 short of a majority in that body.

What do you think? Feel free to leave your own predictions in the comments section.

UPDATED: So you want to get a PhD in the humanities/political science

Makes you laugh, but also cry a little bit inside…

Update: and for those polisci folks out there

Social Science: Apparently Too ‘Sciencey’ for the Iranian Government…

so says a senior Education Ministry official:

“Expansion of 12 disciplines in the social sciences like law, women’s studies, human rights, management, sociology, philosophy….psychology and political sciences will be reviewed,” Abolfazl Hassani was quoted as saying in the Arman newspaper.

“These sciences’ contents are based on Western culture. The review will be the intention of making them compatible with Islamic teachings.”

The Ayatollah added:

“Many disciplines in the humanities are based on principles founded on materialism disbelieving the divine Islamic teachings,” Khamenei said in a speech reported by state media.

“Thus such teachings…will lead to the dissemination of doubt in the foundations of religious teachings.”

I have no doubt that this will only serve to elevate Iranian social science to new heights.  I mean, it isn’t like doubt is fundamental to the scientific enterprise or anything…

[via The Monkey Cage]

IR makes it to TED

Joseph Nye gives a TED talk:

To my knowledge, this is only the third political scientist and IR specialist to give a TED talk (Bruce Bueno de Mesquita and Samantha Power being the other two).  Hopefully we’ll see more as I think the TED series is a great way for polisci and IR scholars to make their knowledge relevant and understandable outside of the discipline.

In Praise of Falsification

For those that have not read it yet, The Atlantic recently featured an article profiling Dr. John Ioannidis who has made a career out of falsifying many of the findings of medical research that guides clinical practice.  Ioannidis’ research should cause us all to appreciate the various bias we may bring to our own work:

[C]an any medical-research studies be trusted?

That question has been central to Ioannidis’s career. He’s what’s known as a meta-researcher, and he’s become one of the world’s foremost experts on the credibility of medical research. He and his team have shown, again and again, and in many different ways, that much of what biomedical researchers conclude in published studies—conclusions that doctors keep in mind when they prescribe antibiotics or blood-pressure medication, or when they advise us to consume more fiber or less meat, or when they recommend surgery for heart disease or back pain—is misleading, exaggerated, and often flat-out wrong. He charges that as much as 90 percent of the published medical information that doctors rely on is flawed. His work has been widely accepted by the medical community; it has been published in the field’s top journals, where it is heavily cited; and he is a big draw at conferences. Given this exposure, and the fact that his work broadly targets everyone else’s work in medicine, as well as everything that physicians do and all the health advice we get, Ioannidis may be one of the most influential scientists alive. Yet for all his influence, he worries that the field of medical research is so pervasively flawed, and so riddled with conflicts of interest, that it might be chronically resistant to change—or even to publicly admitting that there’s a problem. [my emphasis]

Unlike most famous researchers, Ioannidis is not famous for a positive discovery or finding (unless you count his mathematical proof that predicts error rates for different methodologically-framed studies).  Instead, his status has been obtained because of his ability to falsify the work of others–to take their hypotheses and empirical research and show that they are wrong.

This is highly unusual, not only in the area of medical research, but in most academic disciplines.  The article notes that researchers are incentivized to publish positive findings–preferably paradigm altering ones–and this leads to a breakdown in the scientific method.  As Karl Popper so famously argued, knowledge accumulates based on the testing of theories that are then subjected to replication by other researchers.  If the original findings are falsified–meaning that the evidence does not support the theory–the theory is scrapped and replaced with a new theory that has greater explanatory power.  Knowledge is built through the cumulative falsification of theories.  One can think about falsification as the successive chipping away at a block of stone–the more we chip away the closer we get to an actual form.  If researchers are not incentivized to pursue falsification we all lose as a result, since incorrect findings are not vigorously retested and challenged.  According to Ioannidis, if they are challenged it is often years–if not decades-after they have been generally accepted by research communities.

It would appear that Theodore Roosevelt was not entirely correct.  The critic should, in fact, count a great deal.

[Cross-posted at Signal/Noise]

Friday Signaling Roundup

Here are a few quick signaling items for your perusal.  I will try to do a similar roundup each Friday if I’ve stumbled on enough items throughout the week.  Enjoy!

  • How to Signal That You Are Marrying for Love? It’s tougher than you might think.  Some suggest using a pre-nuptial agreement to signal one’s love and affection instead of their love of money.  If one is truly marrying for love and not money they should have no problem signing a pre-nup if they are the less-wealthy of the pair.  However, the pre-nup may act as a signal from the wealthier of the two parties that they have reason to believe that the marriage will not last.  Therefore, pre-nups are likely only an optimal signal when they are suggested at first by the least wealthy member of the couple. (via Cheap Talk)
  • Tyler Cowen asks the questions “Which ingredient most signals a quality dish?”:  I can’t think of one off the top of my head.  Scallions is noted in the post, and that’s a pretty good one.  I’d think that ingredients that are financially costly and/or time consuming to prepare would also signal quality.  So, higher quality cuts of meat or dishes that are slow roasted or smoked, etc.  A friend of mine once remarked, “Ah, Bean salad.  If you’ve got bean salad then you know there is going to be great desert.”  He was using the quality of an earlier dish to predict the quality of a later one.  (via Marginal Revolution)
  • Can Cheap Talk Deter (PDF)? Potentially in an entry-deterrence situation, according to a draft paper by Dustin Tingley and Barbara Walter.  Tingley and Walter find that in an experimental setting, contra the expectations of their formal model, when participants were able to make a verbal threat to the first potential market entrant it decreased the instances of conflict from 83% (where communication wasn’t allowed) to 38%.  This is interesting, since the verbal threats by the defender where by definition costless (since they wouldn’t not face the challenger again and additional challengers would not know if they followed through on the threat)–meaning, they shouldn’t have revealed any additional information to the challenger.  My first thought is that in an experimental setting subjects might be revealing information through their body language or micro-expressions (which can’t be captured by a formal model) and that these signals conveyed additional information to the challenger.  But defenders where only allowed to communicate their threats to challengers through email.  The authors offer some potential reasons for the discrepant results, such as the unexpected success of early round costless threats actually signals that the defender is a savvy player and understands the game (i.e. fighting early in early rounds to deter future entrants makes sense, and therefore they are likely to follow through on the threat since future entrants will see that they fought).

[Cross-posted at Signal/Noise]

The Individual Utility of Incompetence

There are many reasons why organizations (government, businesses, etc) grow dysfunctional and stagnant.  One major reason lies with the promotion and retention of less capable workers.  There have been a number of studies that explored this dynamic (for example, The Peter Principle, which theorizes that people are promoted as long as they are competent, which means at some point they reach a position of incompetence).  In general, though, the promotion and retention of incompetent workers would seem to run counter to the rational interests of the larger organization.  So why does this behavior persist?  Why are less competent workers able to retain their positions and, in some cases, obtain promotions?

One potential reason is that it is their very incompetence that is valued.  Incompetence acts as a credible, costly signal that they can be trusted by superiors looking to accumulate a power base.

Sociologist Diego Gambetta is a pioneer in the study of signaling.  In his 2007 book Codes of the Underworld: How Criminals Communicate, Gambetta uses the extreme case of cooperation amongst criminals to tease out more general dynamics of trust, signaling, and communication.  The Mafia can be considered a “hard-case” for theories of signaling trust; given the extreme incentives for criminals to lie and the lack of credibility they wield given the very fact that they are criminals, how is it that criminals manage to coordinate their actions and trust each other at all?  By understanding how trust works in this harsh environment we learn something about how to signal trustworthiness in broader, less restrictive environments.

Gambetta theorizes that one way that a criminal can signal their trustworthiness to another is through their own incompetence:

The mobsters’ henchman, so often caricaturised in fiction as an énergumène, epitomizes the extreme case of this class. If he were too clever he would be a menace to the boss. Idiocy implies a kind of trustworthiness.  […] One way of convincing others that one’s best chance of making money lies in behaving as an ‘honourable thief’, is by showing that one lacks better alternatives.  […] Incompetence is one way of telling people “You can count on me for even if I wanted to I would not be able to cheat.”

Through this mechanism, lower-level criminals can signal their trustworthiness to their bosses, since they are essentially dependent on their bosses for their economic gains given their lack of independent skill and intelligence.  This pervasive logic means that criminal organizations are likely to employ mostly incompetent criminals and that leaders will likely surround themselves with less competent lieutenants over time.

It is not hard to see this same logic play out in businesses, schools, and government.  If organizations are set up in such a way where the accumulation of loyalists is incentivized instead of performance, we should expect to see a greater number of incompetent employees relative to competent ones.  Additionally, we should see more incompetent employees advance as their “sponsor” advances.

[Cross-posted at Signal/Noise]

Happy 75th Anniversary to the Gallup Poll

“A unique anniversary is upon us. Seventy-five years ago today — Oct. 20, 1935 — the Gallup Poll published its first official release of public opinion data.

Here we are three-quarters of a century later, still working to fulfill the mission laid out in that first release: providing scientific, nonpartisan assessment of American public opinion.

The subject of that first release? Well, given the fact that 1935 was smack dab in the middle of the Depression, it may come as no surprise that the topic focused on public opinion about “relief and recovery,” or in other words, welfare. President Franklin Delano Roosevelt was at that time heavily involved in creating a number of relief, recovery, and work programs designed to help people whose lives were being affected by the Depression. Figuring out what the public thought about all of this became Dr. George Gallup’s first official poll question.”

You can read the rest of Frank Newport’s write up of the first poll here.

[Cross-posted at Signal/Noise]

And the AL Cy Young Award Should Go To…

Time for a little baseball blogging.

There is quite a lot of buzz surrounding the AL Cy Young award this year. While there are a number of pitchers that possess a high number of wins (17, 18, 19, and even 20 games), there are many who believe the award should go to Seattle’s Felix Hernandex.  Despite only winning 13 games and losing 12, Hernandez’s performance this year has been nothing short of amazing.  His problem is that he played on one of the worst teams in the league.  He was 8th in the league amongst starters in terms of runs support (86 runs over 34 starts) and was actually dead last in terms of runs support per nine innings (3.1).  If you look beyond wins to the other two orthodox statistics that make up the pitching triple crown, Hernandez finished first in ERA (2.27) and second in strikeouts (233).  It is his performance in these other two categories that have many arguing for Hernandez to win the award, since he shouldn’t be penalized for his team’s lack of ability to score runs to support his dominance.

If someone like Hernandez wins this year it would truly represent a paradigm shift in the way baseball writers evaluate player performance.  In the history of the AL Cy Award, no starting pitcher has ever won with less than 16 victories (Zach Greinke won last year).  In the NL, only Fernando Valenzuela managed to win the award with as few as 13 wins, and that was in 1981, and no winner from either league had a record as close to .500 as Hernandez does.

That being said, I would actually argue that Hernandez is not the only “non-orthodox” contender.

There is only so much control a pitcher has over the outcome of a game.  And while starting pitchers have more control than most, they still must rely on their defense to play well and on their offense to score runs.  So rather than focus on statistics such as wins (which are heavily dependent on a team’s offense), we should evaluate starting pitchers on their performance independent of their offense and–to the extent possible–their defense. Doing this means focusing on how often hitters deny batters the chance to put the ball in play (strikeouts), how often they give a batter a free pass (walks), how many base runners they allow (WHIP), and how deep into a game they pitch, which gives their bullpen rest and allows their manager to use only the team’s best relievers (thereby, giving the team the best chance to win).

So let’s look at a few statistics:

K/9 – Strikeouts per 9 innings: The more batters a pitcher strikes out, the better.
K/BB – Strikeouts to Walk Ratio: The more strikeouts relative to walks, the better.
WHIP – Walks + Hits per Inning Pitched: The fewer baserunners a pitcher allows to reach base, the better.
FIP – Fielding Independent Pitching: Measures a pitchers performance independent of the quality of their defense.  Lower the better.
RS/9 – Run Support per 9 innings: How many runs a pitcher’s offense scores for them per nine innings.
IP/GS – Innings Pitched per Game Started: The more innings pitched per start, the better.

I’ve created a table with non-counting statistics for the top 10 pitchers in the AL this year, but I have not included their names or their traditional statistics (Wins, ERA, or K’s).  Take a look and think about who jumps out as the best pitcher:

Now, all of these guys are good, but there is one whose performance really jumps out.

First, it’s hard to miss the obvious gap between Pitcher A and their K/BB ration of 10.28 and the rest of the field.  For every 1 batter Pitcher A walks he also strikes out 10.  That is more than double the next closest pitcher (Pitcher B at 4.31).  That ratio of 10.28 is the second highest in the history of baseball and only the third time we’ve seen a double-digit ratio (the other other two times-1994 and 1884).  Pitcher A also had the lowest WHIP, the lowest FIP, and the highest IP/GS.  The only two areas he didn’t finish first is K/9 (10th) and RS/9 (4th fewest).

So who is Pitcher A?  Felix Hernandez?  Nope.  It’s Cliff Lee.

Here’s the chart with the names included:

In terms of the traditional statistics, Lee only went 12-9 with a 3.18 ERA (6th in the AL) and 185 strikeouts (10th in the AL) in 28 starts.  At first blush, his body of work doesn’t look that impressive.  But if you go beyond mere “counting” stats, Lee’s dominance becomes more evident and Hernandez-esque.  His higher ERA (still 6th best) can be explained by an unusually high .302 batting aver for balls in play (BABIP), meaning when batters actually managed to put the ball in play they reached based 1/3 of the time.  BABIP is strongly correlated to ERA.  My guess is that Lee’s high BABIP can be explained by the fact that the defense behind him wasn’t the greatest, reflected in the fact that he had the best fielding independent pitching in the AL amongst starters.

Hernandez had less run support (3.10 to 4.45) and more strikeouts per nine innings (8.36 to 7.84), but otherwise Lee was better than Hernandez in every non-counting category (and he was better than every other contender).

Will Lee win the AL Cy Young?  I doubt it.  My guess is it will either go to Hernandez or CC Sabathia (since he had 21 wins and played for the Yankees in the AL East), but it is hard to argue with how dominant he was over the course of the regular season.

[Cross-posted at Signal/Noise]

You can’t create national security policy in a vacuum

Stephen Biddle has a spot-on piece over at Foreign Policy on how Presidents, and Obama in particular, must take into account domestic politics when setting national security strategy.  With the release of Bob Woodward’s latest book, Obama’s Wars, many have jumped on the President’s alleged quote that he can’t lose the entire Democratic Party to justify the need to set a troop draw-down date for Afghanistan as evidence that he’s putting politics above national security (as if anything can be separated from politics).

Biddle responds:

…I do know that it’s no sin for a president to consider the domestic politics of military strategy. On the contrary, he has to. It’s a central part of his job as commander in chief.

Waging war requires resources — money, troops, and equipment — and in a democracy, resources require public support. In the United States, the people’s representatives in Congress control public spending. If a majority of lawmakers vote against the war, it will be defunded, and this means failure every bit as much as if U.S. soldiers were outfought on the battlefield. A necessary part of any sound strategy is thus its ability to sustain the political majority needed to keep it funded, and it’s the president’s job to ensure that any strategy the country adopts can meet this requirement. Of course, war should not be used to advance partisan aims at the expense of the national interest; the role of politics in strategy is not unlimited. But a military strategy that cannot succeed at home will fail abroad, and this means that politics and strategy have to be connected by the commander in chief.

State leaders must always balance the domestic and international when formulating policy.  What may be possible internationally may not be sustainable domestically, and vice versa.  Ignoring either one typically leads to disaster.  Political scientists have long argued that outcomes are the result of simultaneous negotiations between domestic and international audiences, as well as the difficultly states face when trying to sustain public supporter for wars of choice.  Condemning leaders for being prudent may make for good copy, but it makes no sense given all we know about policymaking.

Applied Signaling: Pajamas and 3-year olds

Every night, about 15 minutes or so after we’ve put my 3-year old daughter to bed, we inevitably hear a knock at the door.  She’s typically knocking because she needs to go the bathroom.  She’s also knocking because she wants to scope out what we are doing, find out if she is missing anything.  One thing that bothers her is if me or my wife leaves the house after she goes to bed.  In order to go to sleep she needs some kind of guarantee that we aren’t leaving and are getting read to go to bed just like her.  It appears she’s found one–whether me or my wife have gotten changed into our pajamas.

If we come to her door in our pajamas–or at least different clothes (e.g. sweatpants, etc) than when she last saw us–she takes it as a signal that we are in for the night.  If we were going out or not going to bed soon we would still be in our regular clothes that we wore earlier.  If we haven’t changed, she probes–“why aren’t you in your jammies?”  This let’s us know that she suspects we aren’t in for the night.  It also means that she will likely spend a fair amount of time looking out her window to see if our cars stay in the driveway before she will settle in and go to sleep.  Now, putting on pajamas isn’t that costly of signal–there is nothing stopping us from putting them on and then changing back into regular clothes to leave the house or host guests.  (However, in all honestly this isn’t likely to happen.)

The lesson here is that a) the idea of seeking out signals is intuitive for people and we start at a very early age, and b) rather than fight with our daughter about going to bed we might be better served just changing into our pajamas out the outset to demonstrate to her that we aren’t leaving the house, no one is coming over, and we are also getting ready for bed.  She may not believe our words, but she seems to believe the signal that she’s identified.  Leveraging that signal can lead to better communication and the outcome that we want.

[Cross-posted at Signal/Noise]

Revolving Doors, Lobbyist Edition

Via Marginal Revolution, an interesting new paper that explores what happens to an ex-staffer’s lobbying revenue when the politician they worked for leaves office.

Our main finding is that lobbyists connected to US Senators suff er an average 24% drop in generated revenue when their previous employer leaves the Senate. The decrease in revenue is out of line with pre-existing trends, it is discontinuous around the period in which the connected Senator exits Congress and it persists in the long-term. The sharp decrease in revenue is also present when we study separately a small subsample of unexpected and idiosyncratic Senator exits. Measured in terms of median revenues per ex-staffer turned lobbyist, this estimate indicates that the exit of a Senator leads to approximately a $177,000 per year fall in revenues for each affiliated lobbyist. The equivalent estimated drop for lobbyists connected to US Representatives leaving Congress is a weakly statistically signi cant 10% of generated revenue.  The equivalent estimated drop forlobbyists connected to US Representatives leaving Congress is a weakly statistically signi cant 10% ofgenerated revenue.We also find evidence that ex-sta ffers are more likely to leave the lobbying industry after their connected Senator or Representative exits Congress. (emphasis mine)

They also show that ex-staffers revenues has grown at a faster rate than non ex-staffers since the late 1990’s.

Here’s a graphical representation of the findings from the paper:


[Cross-posted at Signal/Noise]

To explain the origin of exchange, you must explain the origin of trust

I’ve just started reading Matt Ridley’s The Rational Optimist.  So far, it is an excellent, through-provoking read.  A key to Ridley’s argument is that the innovation of exchange–the trading between two parties of separate items or services that both parties value–that led to mankind’s dominance of the planet and the explosion of knowledge and technology.

Ridley explains how exchange–or barter–is qualitatively different from reciprocity (an activity that can be found in other species):

at some point, after millions of years of indulging in reciprocal back-scratching of gradually increasing intensity, one species, and one alone, stumbled upon an entirely different trick. Adam gave Oz an object in exchange for a different object. This is not the same as Adam scratching Oz’s back now and Oz scratching Adam’s back later, or Adam giving Oz some spare food now and Oz giving Adam some spare food tomorrow. The extraordinary promise of this event was that Adam potentially now had access to objects he did not know how to make or find; and so did Oz. And the more they did it, the more valuable it became. For whatever reason, no other animal species ever stumbled upon this trick – at least between unrelated individuals.

As I read this it occurred to me that Ridley is likely right, but also that exchange is just as dangerous an activity as it is a transformative one.  Why?  Because to base one’s existence on exchange means making oneself vulnerable to and dependent on others for what one needs.  As Ridley notes, earlier humans were self-sufficient.  But moving from self-sufficiency to exchange means trusting others that they will provide what you need and will honor the exchange.

In the present, we take this somewhat for granted.  I assume that my local grocer will have the fruits, vegetables, etc, that I need to feed myself and my family.  I don’t worry about the possibility that they either won’t have my food or that they will refuse to provide it to me in exchange for the money that I have.  But imagine back about 100,000 years ago.  At some point, someone had to take a very big leap and become dependent on someone else for what they required for survival.

As a political scientist, my initial reaction is that trust both emerged from repeated interactions with barter partners and was then institutionalized through the emergence of government.  Too often government is derided as an impediment to economic growth, but we often forget that without it one is hard pressed to explain sustained progress.  A capitalist economic system cannot function without a robust legal system that includes rules for exchange and a system that monitors and enforces violators.  How else can a society become so utterly dependent on anonymous, non-local actors to provide that which is crucial for survival?  That isn’t to say that government can’t also play a negative role–often it has.  But ignoring the positive, necessary role that it plays is quite dangerous in my view.  It also requires us to ignore the lessons of history.

I am only up to Chapter 2, and it appears that Ridley will take up this question of how trust emerged in Chapter 3.  I’ll be curious to see how he deals with this question and what answer he proposes.

[Cross-posted at Signal/Noise]

“We should not see moving out of academia as a failure”

Via Drew Conway, a great quote this morning from Stephen Curry, a professor at Imperial College London:

Students should think more broadly about what a PhD could prepare them for. We should start selling a PhD as higher level education but not one that necessarily points you down a tunnel…We should not see moving out of academia as a failure. We need to see it as a stepping stone, a way of moving forward to something else.

Curry was commenting here on changing the mindset of the students, but I would argue in many disciplines the problem isn’t the students, but the professors.  There are still large groups of people in academia that not only disagree with this sentiment, but actively work to undermine students who choose to take their education and apply it outside of academia.  My experience has been in the realm of political science, but certainly know others that have had similar experiences in other disciplines.

The skills one learns in graduate school are absolutely applicable outside of academia.  In many cases, students may be better positioned to apply what they’ve learned and have a more fulfilling career in either government or business.  Not everyone is cut out for this type of career, but then again not everyone is cut out for a life in academia either.  In many cases, it takes a different set of talents to thrive in either environment.  And when we take into account the utter dysfunction of the academic labor market, I don’t think pressuring students to seek a career in that market is the most responsible thing to do.

Bottom line: the focus should be on the students and what will be the best move for them, not what professors think is the ‘proper’ career for those pursuing and holding a Ph.D.

[Cross-posted at Signal/Noise]

Predicting flu outbreaks, fashion trends, and political unrest with social networks

[Cross-posted at Signal/Noise]

Nicholas Christakis and James Fowler have released a new paper that looks at the potential predictive power of social networks.  They claim that current methods of contagion detection are, at best, contemporaneous with the actual epidemic.  What is needed is a true early detection method, one that would actually provide an accurate prediction of a coming epidemic.

Christakis and Fowler claim that social networks can be used as sensors for various types of contagions (whether biological, psychological, informational, etc).  In an inventive twist, they leverage what is known as the Friendship Paradox–the idea that, for almost everyone, a person’s friends tend to have more friends than they do.   Contagions tend to appear sooner in those individuals that are closer to the center of a social network.  The logic goes that if you ask a group of people to name one of their friends, those friends will be closer to the center of the network than the people you asked.  Rather than map and monitor an entire social network, simply monitoring these friends should allow researchers to detect the outbreak of, say, H1N1 much earlier.

They tested their theory using Harvard College undergrads, attempting to detect the outbreak of the flu.  (You can watch Christakis discuss the paper and research during a recent TED talk in the video embed below).  What did they find?

Based on clinical diagnoses, the progression of the epidemic in the friend group occurred 13.9 days (95% C.I. 9.9–16.6) in advance of the randomly chosen group (i.e., the population as a whole). The friend group also showed a significant lead time (p,0.05) on day 16 of the epidemic, a full 46 days before the peak in daily incidence in the population as a whole. This sensor method could provide significant additional time to react to epidemics in small or large populations under surveillance. The amount of lead time will depend on features of the outbreak and the network at hand. The method could in principle be generalized to other biological, psychological, informational, or behavioral contagions that spread in networks.

That is a pretty impressive result.  By simply tracking those individuals located closer to the center of the network, Christakis and Fowler were about to detect the progression of the flu a full 2 weeks before the general population.  They were also able to derive an early warning signal over a month before the peak of the outbreak in the general population.

If this result can be replicated and validated there are various ways it can be utilized.

Here are a few off the top of my head:

  1. Product Launches: Particularly in the tech industry–where so often we now see product launches as proto-typing–we could use this method to very quickly gauge the awareness and adoption of a new product and predict the extent to which it will spread throughout the general population.  Companies would have better early-warning systems, which would allow for killing dud products or boosting marketing for those products that are poised to explode.  I would assume this would be particularly applicable to products that benefit/rely on network effects.
  2. Political Indicators: One can think of political unrest as a contagion–discontent starting earlier with a core group within a social network and then, over time, spreading to those on the outskirts of the network.  Tracking the population as a whole may not give you an early warning of unrest, but rather a snapshot of a problem at a time when it is too late to do much about it.  Focusing on those closer to the core of a social network could provide enough lead time to diffuse tensions or intervene in other ways to avoid a full-scale upheaval.  Moreover, businesses and investors could also use the early warning as a signal to make adjustments in supply chain and their portfolios to take into account the potential unrest.  Finally, citizens within those countries could benefit by having more lead time to evacuate conflict zones, etc.
  3. Economic Indicators: Investors, businesses, and politicians are always looking for better economic indicators–those signals that are leading indicators of larger economic trends.  I wonder if adjusting the sampling frames of various polls to incorporate the Friendship Paradox might give us an even earlier warning for mortgage defaults, consumer confidence and spending, manufacturing activity, etc.  Not as sure about this one, but certainly much of economic activity takes place in a networked structure.

Would love to hear other thoughts.

Book Blegging

Loyal Duck readers, I was hoping you might be able to help me out.

Do you have any recommendations for books about the inventive ways that people (scientists, designers, business folk, etc) have evaluated hard to test subjects? I am looking for something that is less about methodology, per se, and more about testing ideas in a practical way where either the environment or subject matter makes testing difficult (thinking here of astrophysics, for example). I am not looking for something that looks at the subject from a philosophical standpoint, but is more of a collection of examples that highlight the inventive ways people have gone about testing hypotheses in practical ways.

For example, I am thinking here of Shapiro’s famous observational test of general relativity (the Shapiro Delay), or the discovery of Neptune.

Hopefully this makes some sense. Any suggestions?

Thanks in advance!

Older posts

© 2017 Duck of Minerva

Theme by Anders NorenUp ↑