"I Knew It All Along"
Photo by Toa Heftiba on Unsplash
Hindsight bias is the tendency, after learning the outcome of an event, to believe that we had predicted or known that outcome before it occurred. Once we know how things turned out, the outcome seems obvious — even inevitable. We reconstruct our memories of our prior beliefs to be more consistent with what actually happened.
The foundational research was conducted by psychologist Baruch Fischhoff, then at the Hebrew University of Jerusalem, in a series of studies beginning in 1975. Fischhoff coined the term "creeping determinism" to describe the way known outcomes retrospectively seem to have been inevitable. In his experiments, participants who were told the outcome of a historical event consistently rated that outcome as having been more probable before the fact than participants who had not been told the outcome — even when instructed to ignore the outcome when judging prior probability.
Why Hindsight Bias Is Dangerous for Learning
Photo by Vitaly Gariev on Unsplash
The most serious consequence of hindsight bias is that it disrupts learning from experience. If, after the fact, an outcome feels like it was obvious or predictable, there is nothing to learn — no surprise, no puzzle to resolve, no reason to search for what went wrong in your decision process.
In financial markets, every major crash looks inevitable in hindsight. The 2000 dot-com collapse, the 2008 financial crisis, the 2020 pandemic sell-off: read the post-event analysis and they all appear to have been clearly signalled and predictable. But the pre-event record shows that most sophisticated market participants did not predict them — including many of those writing the post-event analysis. Hindsight bias edits out the genuine uncertainty that existed before the outcome was known.
This matters practically because investors who believe they "should have seen it coming" often blame themselves for poor judgment rather than asking whether their process was sound. A good process can produce bad outcomes; hindsight bias makes it hard to distinguish process quality from outcome quality.
The Memory Revision Process
Photo by Kelly Sikkema on Unsplash
Fischhoff's research showed that hindsight bias is not merely about saying the wrong thing — people genuinely misremember their prior beliefs. The memory is revised, not just the report. This makes it particularly insidious: you cannot simply decide not to do it. The revision happens automatically, below conscious awareness.
Subsequent researchers, including Ulrich Hoffrage and Rüdiger Pohl who edited the comprehensive review volume Hindsight Bias (2004), confirmed that the bias is remarkably robust across cultures, domains, and levels of expertise. Experts are not immune — and in some domains, expertise may increase hindsight bias by giving people more material to construct compelling retrospective narratives.
Protecting Decision Quality
Photo by Kelly Sikkema on Unsplash
The most effective defence against hindsight bias is to record beliefs and reasoning before outcomes are known. Decision journals — written records of what you believed, what you expected, and why you made a choice — create evidence that cannot be revised by hindsight. Reviewing those records after outcomes are known provides an accurate picture of your prior reasoning, making genuine learning possible.
Pre-mortems (imagining a future failure in advance) and explicit probability estimates before events also serve as anchors that can partially resist hindsight revision.