I've been listening recently to Michael Lewis's podcast Against the Rules, in which he recently revisited the subject of his famous book, Moneyball. I am a huge sports data guy, and I enjoy Lewis’ storytelling, though what is fascinating about this podcast is that it cuts against many of the common understandings that have arisen about Lewis’ original work:
The popular understanding of Moneyball, especially the Brad Pitt movie version, is that Moneyball is a story about how data triumphs over old-school intuition or how a bunch of math nerds proved that the baseball establishment was hopelessly backward. For many people, the assumption is that, except for a few old-school former player holdouts, the story ends with the data nerds gradually winning over the last 20 years. In some ways, this understanding is proper: Moneyball won. Every team in baseball is spending far more on data analytics than they were 20 years ago. But in many ways, I think this also artificially cheapens the full context of the “Moneyball” impact over the last 20 years. While data certainly has changed baseball (and all sports), only been a small part of what has been a broader and revolutionary rethinking of how baseball worked.
Baseball has always used data. I grew up collecting baseball cards, which I loved most because the back of the baseball card was a treasure trove of data. You could use this data to mentally track how good a particular player was, as Baseball is a very quantifiable game. Every plate appearance in baseball involves a batter getting on base or getting an out. Every hit includes a batter advancing a certain number of bases and a certain number of runs scored. Moneyball did not introduce the idea of thinking of players in terms of their stats because baseball has always been a sport that appealed to a certain kind of statistically minded kid. There have always been baseball stat nerds. What Moneyball did was shift our understanding of which statistics matter most. As a kid, my baseball knowledge was based heavily on statistics; it just happened to be a very flawed understanding of what mattered. For instance, as a kid, I would always gravitate towards home runs as the single most crucial explanatory statistic. While home runs are significant, they should not be used as the single metric to determine a player's overall value. Other revered statistics like Batting Average and RBIs have gradually lost cache, as their value in predicting wins is more diminutive than believed.
The Moneyball revolution had its seeds in the work of Bill James in the late 70s when he used the tools of economics and empirical data to demonstrate that many overlooked statistics that are far more explanatory of how much a player contributes to winning. Walks and on-base percentage are two statistics featured prominently in the Moneyball movie, but countless other statistics, some longstanding and some new, have proven vital. The revaluing of statistics inspired by Moneyball led to the creation of Wins Above Replacement, or WAR, which tries to take all the insight of what statistics matter and turn it into a simple data point that we can use to compare players. Of course, WAR is not a 100% perfect measure of how good a player is but rather a better version of the simple statistical model of the previous generation (i.e., looking at a player’s home runs, batting average, RBIs).
We can tell similar stories about other sports, including basketball, where people have always relied on data to supplement their subjective judgment of how good a player was. When I played in high school, I would always check the scoring book after the game to see how many points I had scored because this was the best data I had to judge my performance. This perception was not 100% wrong: every advanced statistical basketball model finds scoring central to impact. We can expect a player who can score 15-20 points a game to be a net positive to their team’s success. However, if you treat points scored as a default model, you miss essential information about what a player contributed. Defense and passing are two big ones, neither of which was my particular strength in high school. Coaches and players who rely on points per game are cheating themselves on an opportunity to win and get better. Contemporary basketball analytics doesn’t give coaches all the information they need; however, they provide a suite of more sophisticated measures of a player’s impact than how many points a player scored (for the curious, a good overview of current NBA metrics can be found here).
The Oakland Athletics continued to consistently overachieve compared to their limited resources during Billy Beane’s tenure. But the A’s succeeded over the last 20 years not because they “solved” baseball in 2002, but because they spent those 20 years constantly hunting for inefficiencies in EVERY aspect of baseball. The movie portrays one of Oakland A’s greatest successes in their 2002 season as unearthing Scott Hatteberg (played by Chris Pratt), whose ability to walk was underrated, and defensive shortcomings were overrated. But as it turns out, while Hatteberg was a valuable player on the 2002 Athletics, he was only their 7th most productive player according to FanGraphs's WAR rankings, far less valuable than their pitching staff (whose contributions are absent from the movie). Hatteberg's 2002 success would be a bit of a fluke in the years following. From 2003 to 2005, Hatteberg would be a below-average player for the A’s. While his ability to draw walks kept his offense, his defense got worse and neutralized any benefit his offense gave the team (which is why his playing time gradually decreased over the coming years.
Bill James, the OG sabermetrician of baseball, always pointed out that statistics missed incredible nuances. His go-to example was that box score statistics made two significant mistakes when it can to defense: they both obscured the importance of defense while also not measuring the defensive value of baseball players very well. The primary defensive statistic tracked by box scores is “Errors,” which count mistakes that players make. But errors are not a meaningful statistic if they are not compared to the number of positive defensive plays a player makes! James was fond of saying that the shortstop whose athletic prowess leads him to get a glove on more hard-hit balls likely will end up with more errors recorded than the slower, less capable defender who never has a chance at getting to the ball in the first place. Derek Jeter never committed many errors, making some believe he was a good defender. Over time, however, analysts increasingly recognized that his minimal defensive range in his mid to late-career made him a far below average defensive shortstop (a subject very unpopular with the many Yankees fans out there).
The A’s learned this quicker than anyone: as 538 points out, the As actually had quite a mediocre On-Base Percentage through Beane’s 20 years. Where they excelled was spotting excellent defenders and developing great pitchers. Some of these inefficiencies were spotted and developed through data: new and better statistics, like defensive runs saved, have been created over the last 20 years to do more justice to defense. But even these flaws are widely understood to be flawed in various ways, so many teams are increasingly trying to craft a more complex statistical model that layers a complex array of contextual statistics on top of the more detailed statistics. Part of the Moneyball approach is to see baseball the way an economist would: looking for inefficiencies and underutilized resources, which means constantly turning over new leaves for data in search of an edge.
But it's also worth pointing out that rather than displacing old-school scouting or skills coaching, teams like the As that have overachieved have invested more in scouting, player development, and injury prevention. The two are complementary: Analytics staff can use data to highlight where scouting and coaching can make the most impact on performance. Striving for the most accurate understanding of baseball possible leads to teams maximizing the quantifiable parts of baseball AND optimizing the elements of the game shaped by human judgment. Even former skeptics like the Washington Nationals General Manager Mike Rizzo have grown to appreciate Moneyball precisely because they could see how this approach seamlessly could integrate with their old-school technique. Rizzo attributes the National's success from 2012 through 2019 to using a hybrid approach that incorporated data while still valuing old-school scouting and coaching.
The Moneyball revolution is best seen as a broad effort to create a better mental model of how sports work rather than some narrow triumph of a few statistical measures. Author Shane Parrish describes a mental model as follows:
A mental model is simply a representation of how something works. We cannot keep all of the details of the world in our brains, so we use models to simplify the complex into understandable and organizable chunks.1
Parrish goes on to describe the motivation for building a better mental model as follows:
In life and business, the person with the fewest blind spots wins. Removing blind spots means we see, interact with, and move closer to understanding reality. We think better. And thinking better is about finding simple processes that help us work through problems from multiple dimensions and perspectives, allowing us to better choose solutions that fit what matters to us. The skill for finding the right solutions for the right problems is one form of wisdom.2
The best way to understand the progress of the Moneyball revolution: it pushed baseball to try to make use of all the tools available to eliminate the blind spots of previous generations better. Yes, better utilizing data and creating new statistics are crucial to developing that mental model. It is impossible to find successful baseball teams in the modern MLB (or the modern NBA) that do not leverage analytics. That said, framing the revolution as one where nerds are displacing jocks in a competition for jobs and status needlessly pits different sources of information against one another. Teams do not use Analytics for their own sake; they are beneficial because they have the potential to improve the game. As a high school basketball coach, I quickly realized that while using data is valuable, there is far more to understanding basketball deeply than data. Modern sports teams are better at data, but they are also better at leveraging the wisdom of scouts and the work of coaches. And within the realm of successful franchises, some have a slightly different mix of using data, scouts, and coaching. And we should embrace that diversity as needed to push things forward into the future.
My full plunge into the world of housing policy came in the Spring of 2014 when I read a short book written by Ryan Avent called The Gated City. I was only 2 or 3 months away from moving to Los Angeles to start living and working on the Eastside. While I had a background in economics, I wasn’t familiar with the vast landscape of economics research on urban economics and the housing market's role within cities. I was more deeply concerned with narrow conversations about gentrification and economic justice, as I was on the verge of moving into a neighborhood where those conversations were very salient. Around the time I moved into the Eastside, a small group of radical young activists vandalized art galleries and coffee shops because they felt like those businesses were causing gentrification in the neighborhood. One of the programmatic leaders of the non-profit I was volunteering at in my first few years in the area spoke favorably about the group’s actions because he thought gentrification was an existential risk to the community. One leader at my church publically said that it was a mistake to allow interns like myself to come to the church because we were contributing to gentrification and colonization. It took some time to process all of this, as you can imagine!
Avent’s book would be only the first in a long line of housing policy materials I have read in the last eight years. The subject has been fundamentally transformative as I grapple with what I saw around me here on the Eastside. It reshaped how I thought about cities; it helped me realize how powerful “the city” tool was in the broader economy. Avent also raises how fundamentally flawed housing policies slowly killed the most economically productive cities. But maybe, more importantly, it helped me reframe the gentrification debate that was raging around me.
Gentrification became a conflictual conversation because it presented the issues as a zero-sum tradeoff between economic development in the neighborhood and economic inclusion. Those fighting gentrification feared that any amenities that could appeal to young college-educated white people (like myself) would necessarily come at the expense of longtime Latino residents of the neighborhood. Not all Eastside residents agree with this perspective; many publicly and privately confided that they had long wanted coffee shops and excellent amenities in the area! But there was widespread discontent that those amenities were coming in simultaneously as rents were going up astronomically, and people were leaving the neighborhood as they could no longer afford their housing. From what was happening on the ground, the perspective that the only way to be inclusive was to stop economic development seemed entirely rational!
The problem with this framing of gentrification was that it did not consider what was happening in the broader California housing market. Housing prices were not just going up in the Eastside, but across California, including many wealthy areas like Beverly Hills that build virtually no housing. Prices are going up broadly because of fundamentally broken land-use policies that have artificially created scarcity across the state, especially in wealthy communities on the Westside and suburbs of Los Angeles. When you frame the problem of housing as just about what is happening in one neighborhood, you miss this. My point is not that the debate over gentrification is unimportant; I think it is crucial to debate the issue! However, to correctly identify the root causes of gentrification, one has to contextualize the narrative in the broader conversation about housing and cities. The empirical literature on the impact of market-rate development in gentrifying communities is complex and mixed and even suggests that often market-rate housing brings down rents in gentrifying neighborhoods. This research has convinced me that if you want to stop gentrification, your most effective strategy is to broadly build new housing in our city, especially in wealthy neighborhoods. This strategy is likely to be far more effective than stopping all change on the Eastside, as some activists are inclined to do.
In the last ten years, a lot has happened to the forefront of the conversation around cities. Around the time I read Avent’s book, Sonja Trauss founded SF BARF and started showing up to San Francisco Planning meetings. As Conor Doughtery chronicles in Golden Gates, this marked the beginning of pro-housing sentiments jumping from economics papers into a full-blown social movement. That social movement is still very young, much like the “Moneyball” revolution was still very young when Michael Lewis’ book came out in 2003. So it is almost certainly too early to make a judgment call about whether it will be successful in its aims. I am not an unbiased observer of this but a semi-active participant, so I will refrain from making too definitive judgments about the relative success of the project.
I am hopeful this movement will, albeit slowly, succeed. Why do I think that? The Moneyball revolution succeeded because even though it met initial skepticism, the imperative to win baseball games was too powerful for teams to resist over the long run. Likewise, the economic inefficiencies in our current housing regime are too significant for our cities to ignore the problem. By some estimates, misguided housing and land-use policy are reducing the wages of American workers by 1.32 Trillion dollars a year, or about $10,000 a year in income for every working adult. We cannot ignore these wasted benefits forever, even if there are powerful incentives in our political institutions that cut against this change.
But like Moneyball, I hope that this re-evaluation of how we see housing and the underlying economics of cities would not just end with how we see housing but would involve a wholesale re-evaluation of the relationship between the city and the economy. After all, if we could miss the massive impact of bad housing policy on our cities, what else are we missing? Once you realize American cities function far below their potential, you start hunting for inefficiencies in our roads, transit, sidewalks, and parking spaces. Then you begin to look at our city’s economic development incentives, business regulations, and labor market regulations?
One of the best explorations of this came in a book called Order Without Design, written three years ago by Alain Bertaud. Bertaud is a French urban planner who started working in Algeria in the years after French Colonialism ended, only to discover that the paternalism of colonialism was still thoroughly pervasive through the Algerian urban planning code. From this experience and numerous others working in developing cities worldwide, from Yemen to China, he became convinced that there was much to be gained by putting urban planning and urban economics in conversation:
Bertaud found that planners were generally hostile to economics, and economists think working for cities is too “low-status” of a job for them. Of course, housing policy has been the most prominent example because the unintended consequences of bad housing policy are massive. But housing is the only area where our cities operate far below their economic potential. Thinking critically about places we can further develop our mental model of cities is crucial to inclusive prosperity in the 21st century. There is a lot to be gleaned from this approach, a stream of thought that has broadly been called “Market Urbanism” (some have jokingly called Bertaud’s book to be the “Market Urbanist’s Bible”). Expect to see some pieces in the next few months that highlight some of what I think Bertaud’s most important observations are
Not everyone will find market urbanism agrees with their ideological priors, either because you are on the political left and are distrustful of markets or are on the political right and skeptical of cities. And I think both of those critiques have merit! However, I believe in creating accurate mental models of how our economy and our cities work. To do that, even skeptics must understand this perspective by reading a little bit more into this. Becoming familiar with market urbanism gives skeptics the tools to provide critiques that market urbanism needs to hear if we will journey towards a better understanding. If we cannot debate and reason our way to better cities, inclusive prosperity will remain an elusive goal for 21st century America (and, for that matter, many other countries). And that would be a terrible future to accept as inevitable. My ultimate hope is that we will look back in 20 years and be able to draw one continuous line around the social movement to try to better our society by improving our cities economically. But that will require a lot of intellectual and practical work, something I hope the readers of this blog will play a role in the future!
Parrish, Shane; Beaubien, Rhiannon. The Great Mental Models Volume 1: General Thinking Concepts (p. 17)
Parrish, Shane; Beaubien, Rhiannon. The Great Mental Models Volume 1: General Thinking Concepts (p. 16)