Episode Summary
Executive Summary: Sean Carroll interviews Daryl Morey about the modern NBA as a decision-making and optimization problem: how data, analytics, human judgment, and incentives shape team building, trades, and player evaluation. They also discuss the limits of interviews, the rise of tracking data and machine learning, roster chemistry, tanking, rule changes, and—at the end—Morey’s skepticism about dark matter, which Carroll strongly challenges.
Main Topics: The GM role as optimization under constraints: Morey frames general management as a decision-making job focused on maximizing championship probability across a multi-year window while surviving year to year. Salary cap, trades, drafts, and free agency are the main levers. Analytics and data replacing intuition alone: The conversation traces the NBA’s shift from scouting-only evaluation to mixed human/data analysis, emphasizing behavioral economics, predictive modeling, and using data to correct bias rather than replace judgment. Player evaluation, tracking tech, and machine learning: They discuss the explosion of tracking data from cameras and league systems, automated recognition of actions, and how AI helps identify patterns in draft models and on-court tactics. Biases, interviews, and human judgment: Morey argues interviews are weak predictors, while scouts and executives are vulnerable to endowment effect, anchoring, and confirmation bias. Still, experienced humans can notice things data misses. Team chemistry, game theory, and roster construction: The discussion covers the importance of fit, chemistry, and having the right mix of personalities, plus the limited but real role of game theory in how teams target stars and structure possessions. Rule changes, tanking, and league incentives: Morey proposes changes like Elam Ending, simpler free throws, fewer live-ball timeouts, less replay, and better draft incentives. He defends tanking as rational asset-building when rebuilding. Dark matter debate: The podcast closes with a physics detour: Morey is skeptical dark matter exists, while Carroll argues the evidence—especially from the CMB and lensing—is unusually strong and predictive.
Key Arguments: A GM’s job is to maximize championship odds, usually over a multi-year horizon, not just win the next game or appease public opinion. Data should be used to ground and challenge human intuition, not to eliminate it; the best decisions combine models, scouts, and experience. NBA analytics once had a big edge by recognizing the value of threes, but that edge has mostly been eroded as the league caught up. Interviews with draft prospects are near-zero value as predictors because they mainly select for interview skill, not NBA performance. The most valuable NBA players generate disproportionate team value because they can influence possessions repeatedly in a game. Game theory matters, but less than people think; elite players can often be identified and attacked repeatedly without the “mixing” being decisive. Tanking can be rational and beneficial when rebuilding because it improves access to high-end talent and future championship probability. Morey doubts dark matter because unseen placeholders often turn out to reflect bad theory or measurement, though Carroll argues the accumulated observational fit is unusually strong.
Data Points: Rockets overall record rank under Morey: 2nd-best in the NBA - Carroll cites Houston’s record during Morey’s tenure, behind only the Spurs. Current Sixers title odds (Vegas): 6% - Morey says Vegas had Philadelphia at about a 6% chance to win the title. Internal title odds: higher than Vegas - Morey says teams almost always rate themselves above external market estimates. Expected edge from analytics: 3% to 5% - Morey says this is the kind of advantage decision science can create over time. Tracking frequency: 25 times a second in 3D - Camera systems track ball and players on the floor at high resolution. Draft interview predictive value: near zero - Morey says draft interviews added essentially no marginal predictive power in Houston. NBA team share of value from stars: about 90% - Morey argues the best players drive the vast majority of value in basketball. Trial game endgame example: last 1:17 took 19 minutes - Morey uses a game example to illustrate how slow the end of NBA games can become. Fans in Philadelphia during pandemic-era games: 15% - Morey notes the team was playing with limited fan capacity at the time.
Pivotal Quotes: "the theory of being a general manager" — Sean Carroll: Carroll frames the episode as a broader lesson in optimization, not just basketball. "we’re trying to optimize our championship probability over that three-year window" — Daryl Morey: Morey explains how he thinks about team-building goals and planning horizons. "I wish it was all science. It’s not. It’s art and science like most things." — Daryl Morey: Morey summarizes the balance between analytics and human judgment in roster decisions.
Implications: The episode shows how high-stakes sports management has become a testbed for data science, organizational design, and bias correction. It also highlights how science-like inference works in public debates—especially when evidence is strong but incomplete, as with dark matter.
From the Episode
That for the greater good of the Mindscape audience. And instead, what we're talking about today is the theory of being a general manager. Okay, and I think that whether or not you care at all about basketball, this is an idea that has much broader application. There's no basketball knowledge or interest required, but what we'll be talking about is a combination of science, art, and human relations. Okay, if you're a general manager of a basketball team, you have an extremely quantified Goal. You want to win games and ultimately win championships, okay? You know whether you have succeeded or not in ways that are highly quantifiable, as I said, much more clear-cut than in many other areas of human endeavor. And you also have resources, abilities that are very quantifiable. You have players, they can score, they do score at a certain rate, they can pass, they can rebound, they can play defense, etc. Then the question becomes: how do you optimize? How do you give in your resources?
I've always had owners who are like just championship robust, like that is your goal. But even within that, to your point, if you just focus on winning in one year, you'll create all these weird SKUs. So we look over about a three-year window and we're trying to optimize our championship probability over that three-year window. And we use a lot of outside measures to have internal measures and outside measures. We, like right now, Vegas, I think, has us about a 6% chance to win the title. And that Sounds like really small, but it turns out to be pretty high. That's, I think that's fifth or sixth in the league. Yeah, okay. And you personally, where do you put your chances? Yeah, we have them internally higher, but that's true almost always, right? So, whatever process you have, whatever team you are, if you don't think you have a higher chance than Vegas, then that means you're not actually following any sort of proprietary process.
Makers. I wish it was all science. It's not. It's art and science like most things. That's kind of fun, though. If you have 30 minutes free, you never have to worry about a break-in at your home ever again. That's how quick and easy it is to set up a security system from SimplySafe. It's the kind of thing that's so easy to do, you can do it during a Netflix binge, watching a game, or listening to a certain podcast. I have a SimplySafe system myself, and look, I'm a theoretical physicist. Setting up electronic systems is actually not. Close to my core competency, but I had no trouble putting together the SimplySafe system. You can choose exactly the sensors you need: video cameras, motion detectors, window sensors, etc., and you can program those motion detectors to not detect your pets, which is crucially important here at Mindscape World Headquarters. It's easy to assume that everyone in your house already feels safe, but they might not, and it's worthwhile to talk about it. SimplySafe is a small, easy step to make sure everyone feels safe at your home. So go to simplysafe.com/slash.
About Sean Carroll MindScape
Ever wanted to know how music affects your brain, what quantum mechanics really is, or how black holes work? Do you wonder why you get emotional each time you see a certain movie, or how on earth video games are designed? Then you’ve come to the right place. Each week, Sean Carroll will host conversations with some of the most interesting thinkers in the world. From neuroscientists and engineers to authors and television producers, Sean and his guests talk about the biggest ideas in science, ...