Confident leaders change their minds
I sometimes begin workshops by auctioning a one-dollar coin. The rules are simple. The highest bidder wins the coin and pays whatever they bid. The twist is that the second-highest bidder must also pay their final bid and receives nothing in return. At first the game is entertaining. Someone bids 20 cents for a dollar. Another bids 30. Then 50. Then 80. Before long the bidding reaches a dollar. Then it goes beyond a dollar. Rational people, many of them experienced executives, find themselves paying $1.20, $1.50 or even $2.00 for a one-dollar coin. The room usually erupts in laughter, followed by an uncomfortable realisation: everyone could see the trap, yet people still walked into it.
The game was devised by Yale game theorist Martin Shubik in 1971 as the “Dollar Auction” to illustrate escalation of commitment. Each individual decision appears rational. Once you have bid 95 cents, bidding $1.05 to avoid losing 95 cents seems perfectly sensible. The problem is that the same logic applies to your competitor. What begins as an attempt to make a profit becomes an attempt to minimise a loss. Organisations fall into the same trap every day. Failed technology projects receive another tranche of funding. Acquisitions are defended long after the strategic rationale has disappeared. Governments persist with policies that no longer achieve their objectives. The most expensive decisions are often not the ones that start badly, but the ones that continue long after people know they should stop.
As Dan Ariely argues in Predictably Irrational, irrational behaviour is rarely random. We systematically make the same mistakes because our decisions are shaped by predictable cognitive biases rather than perfect logic. The Dollar Auction is unsettling precisely because it reveals one of those biases in real time.
The real challenge in leadership is not deciding. It is revising. Bayesian statistics teaches that every new piece of evidence should cause us to update our beliefs. Yet organisations often do the opposite. The more money, time and reputation invested in a project, the harder it becomes to acknowledge that the underlying assumptions were wrong. The Dollar Auction captures this perfectly. Rational participants should become less confident as losses mount. Instead, they become more committed.
Leadership is often confused with certainty. We promote people who make bold calls, defend their decisions and appear unwavering under pressure. Once a strategy is announced, a project funded or an acquisition completed, the leader is expected to demonstrate conviction. Changing course can look weak. Admitting that new evidence has altered the case can be interpreted as failure.
The Dollar Auction exposes this flaw in miniature. The rational move is not to defend the previous bid, but to reassess the next one. Yet that is not how most organisations behave. They reward consistency over learning, confidence over curiosity and post-hoc justification over honest revision. A better model of leadership would treat decisions as hypotheses, not monuments. The question would shift from “How do I prove I was right?” to “What would I need to see to change my mind?”
Andy Grove once faced a similar challenge at Intel. The company had built its success on memory chips, but Japanese competitors were steadily winning the market. Grove later recalled asking co-founder Gordon Moore what a new CEO would do if they were both fired. Moore replied that the new CEO would exit the memory business. Grove’s response was simple: “Why shouldn’t you and I walk out the door, come back and do it ourselves?” Intel abandoned the business that had made it famous and focused on microprocessors instead. It was not an act of conviction. It was an act of revision.
Leaders are often taught, implicitly if not explicitly, that they are paid to know. They are expected to be decisive, to have answers, to project confidence and to reduce ambiguity for everyone else. But knowing is becoming the least distinctive part of leadership. In a connected AI era, facts, benchmarks, frameworks and expert summaries are available to everyone.
The real differentiation in leadership comes from combining unique information, judgement and a culture that welcomes challenge. Every organisation has knowledge that exists nowhere else but is only valuable if people are willing to surface it, especially when it challenges the current plan.
Most organisations already know more than their leaders can see. The weak signals are in customer complaints, frontline workarounds, failed pilots, quiet employee doubts and awkward data that does not fit the official narrative. In many cases the evidence that a strategy is failing exists months or years before the strategy is changed. The problem is not a lack of information. The problem is that people learn which information is welcome and which information is not. Once a leader becomes associated with a particular view, challenging that view can feel risky.
This is where the Dollar Auction becomes more than a game about sunk costs. It is a test of whether leaders can revise their approach. In a world of abundant information, the scarce capability is not knowing; it is updating. Truly confident leaders do not define strength as defending yesterday’s decision. They define it as changing course when the evidence demands it.