Jan 11, 2021 · 40m · 20vc
20VC: Annie Duke on Reversible vs Irreversible Decisions, How To Evaluate Risk, The Theory of Sunk Cost in Venture & How to Optimise Both the Discussion and Quality of Investment Decisions
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Former professional poker champion and cognitive decision strategist Annie Duke joins host Harry Stebbings on 20VC to discuss how mental models, decision reversibility, pre-mortems, and structured group deliberation can eliminate cognitive biases and optimize investment quality in venture capital.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 23.6% of the talking time here. How this is scored →
Harry as informed peer 3.0 Guest teaching 4.4 Guest disagreement 1.7 Harry pushing back 1.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 6:44 Rejection of Risk Frame
Annie directly rejects Harry's framing that switching from academia to poker was a high-risk gamble, pointing out that market informational asymmetry made it a high-probability win.
Hardest push from Harry ▶ 12:08 VC Market Reality PushbackHarry explicitly challenges Annie's framework for iterative risk testing by citing the high-speed, competitive reality of venture capital dealmaking.
Biggest teaching moment ▶ 18:46 Behavioral Science of Group DecisionsAnnie educates Harry on decision science research, explaining why classic investment committee discussions amplify bias and why nominal group elicitation before meetings is superior.
Harry holds his own ▶ 12:08 Domain Expertise on Deal PaceHarry demonstrates deep venture capital domain knowledge by challenging theoretical decision models with the realities of fast-moving capital deployment.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Annie Duke's Transition from Cognitive Science to Poker | 1 | 1 | 1 | 0 | Harry introduces Annie warmly as a fan and asks an open-ended background question about her transition from cognitive science to professional poker. Annie shares her personal story without any friction or technical correction. | |
| Evaluating Personal Risk and Decision Permanence | 2 | 4 | 2 | 1 | Harry asks how Annie evaluates personal risk, framing her career pivot as a high-risk gamble. Annie reframes the premise, explaining that informational advantages in 1990s poker meant the risk was much lower than perceived. | |
| Frameworks for Reversible vs Irreversible Decisions | 3 | 5 | 2 | 3 | Harry asks whether Annie maintains a reversible mindset when facing genuinely irreversible decisions like leaving a company. Annie breaks down how most decisions are actually reversible or carry a variable cost of reversal rather than being absolute binary choices. | |
| Managing High-Speed Decisions and Uncertainty in Venture Capital | 5 | 6 | 3 | 5 | Harry pushes back on Annie's iterative testing model by bringing up the reality of modern venture capital, where investors must move too fast to gather iterative data. Annie responds by applying portfolio theory and decision-making under uncertainty behind a veil of ignorance. | |
| Optimizing Collective Decision-Making in Partnerships | 4 | 7 | 2 | 2 | Harry asks how collective decision-making in VC partnerships differs from solo decision-making in poker. Annie delivers a detailed masterclass on Kahneman's mediating judgments and explains why gathering individual feedback before meetings reduces group bias. | |
| Addressing Sunk Cost Fallacy and Conducting Pre-Mortems | 4 | 6 | 2 | 4 | Harry asks how investors should handle the sunk cost fallacy when investments turn sour, and later challenges how pre-mortems can account for unexpected black swan failures. Annie outlines pre-mortem strategies and iterative re-forecasting techniques. | |
| Mitigating Past Bias and Exploring New Opportunities | 3 | 5 | 1 | 1 | Harry asks how to avoid letting past successes or failures bias future investment choices. Annie recommends tracking rejected deals against market consensus and maintaining dedicated exploratory lines for non-obvious opportunities. | |
| Fostering Critical Thinking and a Forecasting Mindset | 3 | 4 | 1 | 0 | Harry asks how emerging investors can foster critical thinking, actively agreeing with Annie's recommendation of Phil Tetlock's Superforecasting. Annie details how a forecasting mindset creates open, information-hungry thinking. | |
| Quickfire Round: Books, Strengths, Weaknesses, and Future Plans | 2 | 2 | 1 | 0 | Harry runs through a friendly quickfire round covering favorite books, risk misconceptions, strengths, and future plans. The interaction remains lightweight, collaborative, and conversational. |
Statements from this episode (17)
Duke: 1990s poker required less skill due to wider informational edge
“So the gap between the people who like understood the game and knew the game and everybody else was so wide that as a professional, you didn't need to be nearly as good then as you would need to be now in order to win.”
Duke: De-risk major life decisions by treating them as non-permanent
“I don't think about most decisions that I make as permanent. So that's a way to de-risk, right? Like, if you know, well, I'll see how this goes, and if it doesn't work out, then I can go do something else”
Duke: Most decisions assumed to be irreversible are actually reversible
“A lot of decisions that we think are irreversible are actually reversible.”
Duke: Move fast when decision reversal costs are low
“If the cost to change our mind is low, then you should just kind of move fast, because it's actually a really good way to build information and actually build those good models, right, is just to interact with the world.”
Stebbings: Capital glut forces VCs to invest without thorough diligence
“In venture, there's so much money today where you need to move so fast that by the time you make the investment, actually, you don't know them very well at all because we're all just throwing money so fast around because there's just too much money.”
Duke: Parallel portfolio bets reduce the penalty of individual inaccuracy
“Like if you can do a lot of bets in parallel, then the penalty for being less accurate on any single one of them is lower.”
Duke: Rating component parts before overall categories reduces decision bias
“When you break it down into component parts, and you rate those component parts first, before you rate the broader category, that this actually helps to discipline bias and noise”
Duke: Working independently yields better feedback than group discussions
“Nominal groups, which are groups of individuals working alone, tend to do better in terms of eliciting the, Feedback and opinions or even brainstorming than when people actually discuss things as a group.”
Duke: About 80% of meeting time is spent discussing existing consensus
“About 80% of meetings get spent on agreements.”
Duke: Decision outcomes depend solely on luck and information quality
“There's only two things that tell you how a decision is going to work out. One is luck. Can't do anything about it. The other is, is what's the information that went into the decision”
Duke: Sunk cost fallacy is the dark side of grit
“Grit is such a great topic, and, you know, sticking to things that are hard and not getting kind of stuck in local minima is a really, really important feature of people who are successful. But the question is, like, you know, there's a dark side to grit, righ…”
Duke: Research shows people escalate commitment when outcomes deteriorate
“There's really amazing work, some from Katie Milkman and Maurice Schweitzer, which shows this escalation of commitment, that when things are going poorly, we actually escalate our commitment.”
Duke: Perform a pre-mortem immediately after making an investment
“As soon as you make an investment, you need to do premortem, and you need to say, what are the ways that this investment can go wrong?”
Duke: Investors should track lost deals identically to won deals
“Right, if you win a deal or you don't win a deal, it's the same thing, because you wanted to win the deal, so you should treat those companies identically.”
Duke: Investors must take exploratory meetings on high-failure deals
“You always have to have some exploratory lines open, right? You have to say, I'm specifically going to take some meetings, like I'm going to do some deck reviews, I'm going to take some meetings with people that I think are going to be really bad because that'…”
Duke: Investors prioritize risk calculation over expected value
“I think the biggest misconception about risk is that, this is going to be kind of weird, is people spend a lot of time trying to calculate their risk, but they don't think enough about whether they actually... Have the expected value, right? Like, are they win…”
Duke: My stock portfolio is indexed and angel deals follow leads
“My stock portfolio is indexed because there's people who spend every single day trading in that market, and I hardly think that I'm smarter than them. And I think that's true of venture as well. So I have done angel investing, but it's always a long So I'm rea…”