Jul 25, 2018 · 1h 55m · knowledge-project
#37 Annie Duke: Getting Better by Being Wrong
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The Knowledge Project, Shane Parrish interviews author and former professional poker champion Annie Duke about decision-making under uncertainty, cognitive biases, and probabilistic thinking. Duke explains how individuals and organizations can overcome resulting bias, build truth-seeking peer groups, and apply practical frameworks like premortems to make smarter decisions with incomplete information.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Shane holds 12.1% of the talking time here. How this is scored →
Shane as informed peer 4.0 Guest teaching 3.1 Guest disagreement 1.1 Shane pushing back 1.1
speaking balance: gold is Shane, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:05:07 Duke rejects short-term downside framing in innovation
Duke forcefully pushes back against the premise that conventional strategies have lower downside by arguing that avoiding single-trial risk guarantees long-term institutional obsolescence.
Hardest push from Shane ▶ 1:40:57 Shane challenges the real-world viability of admitting uncertaintyShane directly challenges Duke's assertion that admitting uncertainty makes communicators more believable, pointing to strong systemic and political incentives that reward unwarranted overconfidence.
Biggest teaching moment ▶ 11:32 Duke disproves Psych 101 learning dogma with poker persistenceDuke dismantles the foundational textbook rule that immediate feedback and financial pain automatically produce learning, citing veteran poker players who repeat identical strategic blunders for decades.
Shane holds their own ▶ 1:47:10 Shane formulates the foresight-hindsight temporal maximShane articulates a sharp temporal decision-making aphorism ('letting the hindsight of your future self become the foresight of today's self') that Duke immediately applauds and adopts.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Shane as informed peer | Guest teaching | Guest disagreement | Shane pushing back | Why |
|---|---|---|---|---|---|---|
| Transitioning from Cognitive Linguistics to Professional Poker | 4 | 2 | 1 | 1 | Shane opens by demonstrating specific background research on Annie Duke's academic thesis regarding syntactic bootstrapping. Duke confirms the details and explains her transition into professional poker via Montana and her brother's mentorship. | |
| Public Perception and Complexity of 1990s Poker | 3 | 3 | 1 | 1 | Duke recounts public skepticism toward poker in the 1990s and explains how poker unmasks the complexity of learning under noisy, hidden-information feedback loops. Shane prompts her on the time pressure and complexity of the game. | |
| Experimental Psychology Versus Textbook Learning Models | 4 | 5 | 1 | 1 | Duke distinguishes experimental cognition from clinical psychology and refutes standard textbook models that claim feedback and pain automatically produce learning. Shane contributes concise framing around pain and reinforcement theories. | |
| Mental Models, Self-Narrative, and Ego Defense | 5 | 4 | 1 | 1 | Shane asks about mental model construction and synthesizes the tension between skill and luck attribution. Duke details how the limbic system and identity defense prompt players to pawn off bad outcomes to bad luck. | |
| Eric Seidel and Eliminating Bad Beat Stories | 3 | 4 | 1 | 1 | Duke shares how mentor Eric Seidel aggressively shut down her bad beat stories to enforce strategic thinking over emotional venting. Shane notes how standard social contracts validate unhelpful excuses. | |
| Truth-Seeking Groups Versus Confirmatory Thinking | 4 | 4 | 1 | 1 | Shane asks why high intelligence correlates with overconfidence in decision making. Duke explains how smart people weaponize data to rationalize priors and contrasts Phil Tetlock's confirmatory versus exploratory thinking styles. | |
| Rewiring Social Incentives and Group Accountability | 4 | 3 | 1 | 1 | Shane queries how individuals overcome the instinctive emotional pain of acknowledging error. Duke shows how group norms can rewire dopamine incentives to reward mistake admission and constructive tribal identity. | |
| Establishing Group Charters and Productive Disagreement | 4 | 3 | 1 | 1 | Shane cites the five-person social influence adage, prompting Duke to outline explicit group charters. Duke details how exploratory groups shorten the feedback loop for catching cognitive errors from decades down to minutes. | |
| Expressing Beliefs Probabilistically and Pricing Reality | 4 | 3 | 1 | 1 | Shane asks whether expressing internal percentages acts as an ego-protecting method to maintain accuracy. Duke agrees, referencing scientific hypothesis falsification and sharing an anecdote about pricing lunch orders probabilistically. | |
| Managing Stakes, Risk Aversion, and Emotional Fluctuations | 4 | 3 | 1 | 1 | Shane asks how monetary stakes and broadcast exposure affect player behavior. Duke breaks down bankroll management, treating chips as functional tokens rather than money, and the emotional hazard of tracking the short-term ticker. | |
| Resulting Bias and the Pete Carroll Super Bowl Dilemma | 5 | 3 | 1 | 1 | Shane connects poker broadcast exposure to corporate surveillance and resulting bias. Duke illustrates resulting bias via Pete Carroll's 2015 Super Bowl pass play call, while Shane draws a sharp parallel to Doug Pederson's Philly Special. | |
| Conventional Failure Versus Unconventional Innovation | 5 | 3 | 2 | 2 | Shane quotes Keynes regarding the corporate incentive to fail conventionally rather than succeed unconventionally. Duke extends the point, contrasting IBM's resulting culture with Apple's appetite for high-variance innovation. | |
| Practical Tools for Probabilistic Decision Making | 4 | 4 | 1 | 1 | Shane asks for specific organizational decision-making workflows. Duke outlines scenario probability trees, red team/blue team perspective-swapping drills, and Gary Klein's premortem technique to reframe dissent as teamwork. | |
| Translating Individual Decisions to Team Accountability and Journaling | 4 | 3 | 1 | 1 | Shane contrasts poker's individual accountability with corporate delegation and asks if Duke maintains a decision journal. Duke explains how her drafting and journaling occur through internalized dialogue with peers. | |
| Essential Context for Objective Decision Analysis | 4 | 4 | 1 | 1 | Shane asks what objective information Duke needs to evaluate decision quality in the absence of outcome data. Duke outlines granular technical poker variables and translates those needs to corporate sales pipelines and relationship histories. | |
| Quarantining Outcomes and Preventing Groupthink | 5 | 3 | 1 | 1 | Shane highlights how organizational hierarchies cause employees to withhold critical data from leaders. Duke discusses how knowing an outcome irrevocably anchors analysis and why leaders must quarantine advisors from outcome knowledge. | |
| Bottom-Up Culture Change and the Decision Swear Jar | 4 | 3 | 1 | 1 | Shane asks how an individual employee can spark bottom-up change without executive mandate. Duke proposes a 'decision swear jar' for black-and-white phrasing and adopting moderating stress-test language. | |
| Three Core Takeaways from Thinking in Bets | 6 | 3 | 2 | 3 | Duke summarizes three core tenets: stating uncertainty, forming groups, and time traveling. Shane pushes back, noting societal and corporate incentives reward broadcasted overconfidence, before formulating a memorable maxim on future hindsight becoming present foresight. | |
| The Morality of Gambling and Advice on Mortality | 3 | 2 | 1 | 1 | Shane poses a listener question regarding the morality of professional gambling and asks what Duke would tell her twenty-year-old self. Duke applies market exchange principles and shares advice on managing existential anxiety over aging. | |
| Episode Conclusion and Podcast Resources | 0 | 0 | 0 | 0 | Shane delivers a solo outro detailing show notes, website links, transcript availability, and the weekly newsletter. Host-side scores are zeroed for this monologue segment. |
Statements from this episode (23)
Duke: Mastering poker reveals how little a player actually knows
“So as you get better and better at the game, you realize you know less and less about the game in the, in terms of the percentage of the things that you, there are to know about the game.”
Duke: Decision quality and outcomes are loosely linked in poker
“Outcomes and decisions are so loosely linked in that game. I mean, you know, you have the best hand, you get crushed, you have the worst hand, you win sometimes, and there's so much hidden information, the cards are face down”
Duke: Fast, abundant feedback does not prevent people from repeating mistakes
“As I continued to play poker, I noticed this over and over again, that people really persisted in their mistakes, despite the fact that there was very clear feedback That was very quick, and there was a lot of it.”
Duke: The pain needed to motivate learning also obstructs it
“You need the pain in order for it to matter to you so that you'll learn, but the pain actually gets in the way of learning, and why is that? Because when you're losing, it just lights your limbic system up.”
Duke: Blaming bad outcomes on luck devastates learning
“Now, what you've done there is you've socialized the result to the world. You've offloaded responsibility for this bad outcome to the world. Now, in the short run, that feels good. Because you don't need to do any kind of, like, identity update. You don't need…”
Duke: Smart people aware of cognitive biases are worse at overcoming them
“In fact, there's a lot of evidence that says that smart people who know about them are actually worse.”
Duke: Table perceptions of gender altered bluffing effectiveness in professional poker
“Strategies that might work for Eric Seidel might not work for me because I'm a woman and he's a man, and so the perceptions of us were very different at the table. So, for example, his ability to bluff in a situation might be really different than my ability t…”
Duke: Truth-seeking groups rewire incentives by immediately rewarding mistake admission
“What happens is that when you set up that kind of group agreement and you really live it, the group is actually reinforcing those kinds of behaviors. So, you know, you can think about habit formation, right? It's that you need to have this reward, and the rewa…”
Duke: Van Bavel and Pereira research shows tribes provide feelings of distinctiveness
“There's a great paper actually that just came out recently from Jay Van Bovel and Andrea Pereira out of NYU, which is talking about this. And one of the things they say is that there's a few things that tribes give you but one of them is actually it gives you …”
Duke: Disagreements trigger defensiveness when perceived as identity attacks
“When somebody disagrees with us, we can very often feel it as an attack on our identity. And when that happens, I think that disagreeableness ensues. Because when you feel like you're personally being attacked, that the information that is being told to you is…”
Duke: Basing identity on uncertainty makes disagreement helpful, not threatening
“Your identity now becomes about acknowledging uncertainty, about how good are you at listening to the other side, how good are you at, you know, analyzing those things, and it stops that I, it stops you from having that defensive reaction, and I think naturall…”
Duke: Poker players must treat chips as behavioral tools, not money
“So one of the things that you want to try to do in poker is to separate the chips from the money. You want to view the chips as tools to accomplish a particular goal in a hand and not view it as money at all.”
Duke: Judging decisions solely by outcomes kills workplace innovation
“So what is, if you're a leader and you're a resulter, which we all kind of are, then what you're doing is you're telling employees, don't ever make an unusual choice. Don't do anything innovative or unexpected because if it fails, I will come down on your head…”
Duke: IBM lost to Apple due to a culture of outcome bias
“And it's, I've always felt that it's because IBM is operating under a culture of resulting, and Apple isn't. Apple can't possibly be Operating under they're doing something unconventional by definition. Their whole business model is to make unconventional choi…”
Parrish: Keynes observed that failing conventionally is preferred over succeeding unconventionally
“You know, it was Keynes who said it's better to fail conventionally than succeed unconventionally, I believe.”
Duke: Assigning explicit probabilities makes decision-makers more open-minded
“The desire to make the guess makes you very information hungry, so it actually makes you more open-minded, because now your goal is not, like, I have to have the right answer. It's, I want to try to get as close as I possibly can to what the objective probabil…”
Duke: Premortems fix corporate agreeableness by rewarding failure identification
“When you do a premortem, now you've changed what it means to be a team player, right? Like the people who are going to get pats on the back are the ones who come up with the most creative reasons for failure.”
Duke: Memorializing probability estimates prevents resulting when bad outcomes hit
“And you memorialize it. So now, when the thing happens that's right up there on the whiteboard, you took a picture of it. That didn't work out, and it's got that percentage by it, you know? Okay, well, we thought 33% of the time this was gonna happen, it was g…”
Duke: I do not keep a decision journal
“I actually don't. I'm not a journaler, and I wish that I were because I think it would be way better.”
Duke: Leaders stating opinions early causes teams to unconsciously support them
“If the leader expresses their belief about what's correct, in a sense, that acts like an outcome. And the group, without knowing it, and the leader without intending to have that happen, now causes a situation where disagreement with that belief is less likely…”
Parrish: Never succeed a long-tenured sycophant under an unchanged boss
“Never follow a sycophant or a yes man who's been in their job for a long time if the boss also hasn't changed.”
Duke: Spotting decision mistakes is much harder after winning than losing
“Because it's a lot harder to spot your mistakes when you win. Because, boy, that, it's like not basking in the glow of how great you are. That, that takes a lot. That's really hard, right? In some ways, that's the harder thing.”