Thinking in Bets

Our belief-formation process is backwards. Harvard psychologist Daniel Gilbert demonstrated that our default is to believe what we hear is true — even when information is explicitly labeled false. Under time pressure, subjects in his experiments presumed all statements were true regardless of labeling. Evolution built us to trust our senses quickly; when language evolved, that same "believe first" wiring carried over to secondhand information.

The consequences are enormous. Most people "know" baldness comes from the maternal grandfather (it doesn't) and that dog years equal seven human years (made up in the thirteenth century). More dangerously, research secretly funded by the sugar industry convinced Americans to replace dietary fat with carbohydrates. Obesity tripled, diabetes surged, and cardiovascular progress stalled — all because a haphazardly formed belief became national policy.

Smart people are better spin doctors. Psychologists West, Meserve, and Stanovich tested blind-spot bias — our ability to see bias in others but not ourselves — and found that more cognitively sophisticated participants showed larger bias blind spots in six of seven biases tested. Awareness of bias didn't help either.

Dan Kahan's research drove this home. When subjects analyzed data about a neutral topic (skin cream), math skill predicted accuracy. But when the same data was relabeled as being about gun control, the more numerate subjects made more errors than less numerate ones, spinning identical numbers to confirm their political beliefs. The polarization didn't shrink among the smartest participants — it grew. This means we can't simply think our way out of bias, no matter how intelligent we are.

Binary thinking cripples decisions. Imagining you're either 100% right or 100% wrong is like using a medical scale with only two settings: 50 pounds and 500 pounds. Instead, Duke advocates rating confidence on a zero-to-ten scale. "I'm 60% that Citizen Kane won best picture" acknowledges what you know and don't know, making you less defensive when new evidence arrives.

This shift has cascading benefits. Updating from "I was 58% but now I'm 46%" feels far less threatening than admitting "I was right but now I'm wrong." It invites collaboration — people share more freely when you signal uncertainty. It makes you a more credible communicator, because listeners trust someone who quantifies their confidence over someone who claims absolute certainty. Scientists already do this through p-values and confidence intervals; the rest of us should too.

Self-serving bias is universal. Phil Hellmuth, the biggest winner in World Series of Poker history, said on ESPN: "If it weren't for luck, I'd win every one." In auto-accident reports, 37% of single-vehicle drivers still blamed someone else. Chris Christie, in a single debate exchange, deflected blame for Bridgegate while claiming credit for New Jersey's job growth — a textbook pivot.

For others' outcomes, we flip the script: their failures are their fault, their successes are luck. Steve Bartman received death threats from Cubs fans despite dozens of spectators reaching for the same foul ball. Phil Ivey offers the antidote. After winning a half-million-dollar tournament, he spent dinner analyzing his mistakes rather than celebrating. Actively crediting luck in your wins and seeking lessons in your losses is the fastest path to learning.

Two words change everything. Imagine confidently declaring "Citizen Kane obviously won best picture" — then hearing "Wanna bet?" Suddenly you're auditing your evidence: Where did I hear that? How reliable is the source? What are the alternatives? The challenge triggers the vetting step we routinely skip in belief formation.

You don't need actual wagers or poker-playing friends. The point is internalizing the frame. When you catch yourself expressing certainty, mentally ask: Would I bet real money on this? The questions that follow are powerful:
1. How do I know this?
2. What is the quality of my sources?
3. What am I missing?
4. What are the plausible alternatives?
5. How confident am I really, on a scale of zero to ten?
This habit alone can interrupt motivated reasoning before it calcifies into false confidence.

Echo chambers are the default. Without an explicit charter, groups drift toward confirmation — nodding along and swapping hard-luck stories. Duke's poker mentor Erik Seidel set different rules: "I don't want to hear it. If you have a question about a hand, you can ask me about strategy all day long." That three-sentence charter established accuracy as the goal, accountability as the norm, and diverse thinking as the currency.

Research supports this. Cass Sunstein studied 6,000+ federal appeals and found that a single judge from the opposing party had "a large disciplining effect" on panel decisions. Homogeneous panels went to ideological extremes. A truthseeking pod needs just three elements: focus on accuracy, accountability with advance notice, and genuine diversity of viewpoints. Even three members — two to disagree and one to referee — can produce robust reasoning.

Odysseus tied himself to the mast before hearing the Sirens' song — binding his future self to better behavior. This Ulysses contract strategy applies everywhere. Arrange a ride-share before going to a bar. Set automatic retirement contributions. Agree on a maximum home price before attending the auction. Duke maintained a poker loss limit enforced by her group — if she lost $600, she left the game, knowing her in-the-moment self couldn't be trusted.

The science confirms this. In a Stanford study, subjects who saw an age-progressed image of themselves allocated $178 to retirement versus $74 for those seeing their current face. Temporal discounting — our tendency to massively favor present rewards — can be countered by making our future self vivid and present. Suzy Welch's 10-10-10 technique helps: ask how you'll feel about this decision in ten minutes, ten months, and ten years.

Positive visualization backfires. NYU psychologist Gabriele Oettingen found that women who fantasized most positively about slimming down lost twenty-four fewer pounds than those who pictured obstacles. Dreaming about success actually impeded it across multiple studies — job seekers, students, and surgery patients showed the same pattern.

The solution combines two techniques. Backcasting starts from an imagined success ("We doubled market share!") and works backward to identify what went right. A premortem starts from imagined failure ("We missed our goal") and identifies what went wrong. Research shows that prospective hindsight — imagining an event has already occurred — increases the ability to correctly identify reasons for outcomes by 30%. Duke helped the nonprofit After-School All-Stars use this approach for grant planning, transforming their budgeting from guesswork into probability-weighted scenario planning.