Hindsight and the Illusion of Inevitability

The idea in plain words

Ask people in 2026 whether the iPhone was obviously going to transform the phone industry, and most will say yes. Ask what the major phone makers should have done in 2007, and the answer seems easy. Yet in 2007 serious, informed people doubted that a phone without a keyboard, sold through a single carrier, at a high price, would matter much to a market dominated by Nokia and BlackBerry. The doubt was reasonable. It only looks foolish now, because we know the ending.

Hindsight bias is the tendency, after learning an outcome, to believe we would have predicted it, and to forget how uncertain things looked before. Baruch Fischhoff, the psychologist who first measured it, called it creeping determinism: the past slowly hardens in memory into something that could only have gone one way.

Where it comes from

Fischhoff’s 1975 experiments were simple. He gave people accounts of historical events, such as a nineteenth-century war between the British and the Gurkhas, with several possible outcomes. Some subjects were told which outcome had happened; others were not. Those who knew the outcome rated it as far more likely, and they also judged the details that pointed towards it as more relevant. Crucially, they believed they would have seen it without being told. In a companion study with Ruth Beyth, people asked to predict events before President Nixon’s trips to China and the Soviet Union in 1972 later misremembered their own predictions, shifting them towards what actually happened.

The philosophical background is older. Henri Bergson, in an essay of 1930, described what he called the retrograde movement of the true. When something new happens, we project it backwards, as if it had always been possible and was simply waiting to be realised. The possible, Bergson argued, does not come before the real; we manufacture it afterwards, out of the real. A finished outcome casts its shadow into the past and makes the past seem to have been heading towards it all along.

Kierkegaard put the human predicament in a sentence from his journal of 1843, usually rendered in English as: life can only be understood backwards, but it must be lived forwards. Decision-makers live forwards. Analysts, journalists, boards and historians understand backwards. Hindsight bias is what happens when we forget the difference and judge the first by the standards of the second.

What it does to the mind

The mechanism appears to be a kind of automatic updating. When we learn a fact, the mind reorganises what it knows around it, keeping the evidence that fits and quietly demoting what does not. The earlier state of knowledge is not stored separately; it is overwritten. That is why hindsight bias survives even when people are warned about it.

Organisations have a narrative version of the same thing. The sociologist Karl Weick described how groups make sense of events retrospectively, building a story that turns confusion into a sequence. The story is useful, since it lets people act, but it smooths away the ambiguity that the decision-makers actually faced. Post-mortems, case studies and business books are especially exposed to this, because they are written by people who know the ending, for readers who want a lesson.

Where it shows up

Almost every history on this site is vulnerable. Read the Internet Tidal Wave memo knowing how the browser war ended, and Gates looks prescient; read it as of May 1995, and it is a bet on a technology with a few million users, made by a company with a successful proprietary online service already in the works. Facebook’s purchase of Instagram for a billion dollars looked extravagant to many observers in 2012. The Roads Not Taken section exists partly to restore the uncertainty: technologies that lost were often the ones that looked most likely to win.

The Decide cases are an exercise against this bias. Each one hides the company’s name and the outcome, and gives only what was known at the time. The point is to feel, briefly, what the people involved felt: that the future was open.

What it asks of someone deciding

Before judging a past decision, reconstruct the information set. What did people know, what did the leading forecasts say, what did competitors do? If the people closest to the situation mostly expected the opposite outcome, the decision may have been reasonable even though it failed.

Keep a decision journal. Writing down predictions with dates and probabilities is the only way to check later what you actually believed, rather than what you now remember believing.

Treat the lessons of famous successes with suspicion. If a story makes the outcome seem inevitable, it has probably lost the most useful information: how close it came to going the other way.