What the data actually shows
There is a documented tendency to judge a decision by its result rather than its quality, known as outcome bias. Baron and Hershey's 1988 experiments found that people rated the very same decision — made with the same information — as better or worse depending only on whether it happened to turn out well or badly. The decision did not change; only the known outcome did, and that alone shifted people's judgments.
Annie Duke, in Thinking in Bets, draws on this to describe 'resulting': the habit of working backward from a good or bad outcome to conclude the decision must have been good or bad. Because the world is uncertain, she argues, a high-quality decision can still lose and a low-quality one can still win — so outcome alone is an unreliable guide to whether you decided well.
A closely related effect is hindsight bias, documented in Fischhoff's work: once people know how something turned out, they tend to believe the outcome was more predictable and obvious all along than it actually was. This 'I-knew-it-all-along' tilt makes past decisions look more clear-cut in retrospect, which feeds harsher and less accurate judgments of choices that broke badly.
Why this feels different from how it actually is
Outcomes feel like the truth about a decision because they are concrete and visible, while the odds and uncertainty that existed beforehand are abstract and easy to forget once the result is in. You can see exactly what happened; you cannot see the other ways it could have gone, so the single outcome takes up all the space in your judgment.
Hindsight bias makes this worse by quietly rewriting your memory of what you knew. After a bad result, it genuinely feels as though the warning signs were obvious and you should have seen them coming — but research suggests that sense of obviousness is largely manufactured after the fact, not an accurate record of your foresight at the time.
And regret tends to attach to outcomes rather than processes, so a choice that turned out badly invites self-blame even when the reasoning was sound. The felt intensity of 'I made the wrong call' often reflects how the dice landed more than the actual quality of the call — which is exactly the distinction resulting collapses.
A high-quality decision can still lose and a low-quality one can still win — so outcome alone is an unreliable guide to whether you decided well.
What the research says to do about it
The core corrective from this research is to evaluate past decisions on what you knew and could reasonably foresee at the time, separating the quality of the decision from the quality of the outcome. Asking 'given what I actually knew then, was this a reasonable bet?' is a fairer and more accurate question than 'did it work out?' — and it tends to be far kinder to yourself.
Thinking in terms of probabilities rather than certainties is the broader habit Duke recommends: treating decisions as bets with odds means a bad outcome is expected some fraction of the time even from good decisions, so a single bad result is not by itself proof of a bad choice. This reframing makes it easier to learn the real lesson rather than the one the outcome falsely advertises.
Where outcome-driven self-blame has hardened into persistent rumination, low mood, or interferes with daily life, the broader evidence supports working with a qualified professional. This page is educational and not a substitute for that. As a thinking tool, separating decision from outcome is offered here for self-compassion and clearer judgment, not as therapy.
What the research says does not help
Judging your past choice solely by how it turned out — resulting — does not help, because it conflates two different things and reliably misleads. A good outcome can flatter a bad decision and a bad outcome can condemn a good one, so outcome-only evaluation teaches the wrong lessons and fuels unwarranted regret.
Telling yourself you 'should have known' after a bad result is usually hindsight bias talking, not an accurate account of your foresight. Research suggests the sense that the outcome was obvious all along is largely constructed after the fact, so treating it as proof of negligence tends to manufacture self-blame rather than reveal a real mistake.
Resolving to simply 'make better decisions' so that nothing ever goes wrong is not achievable, because uncertainty means even excellent decisions will sometimes lose. Aiming to eliminate bad outcomes entirely sets an impossible standard; the realistic aim is a better process over many decisions, accepting that any single one can still break badly.
A single bad outcome is weak evidence about the quality of the decision behind it, especially when real uncertainty was involved.
What this looks like in real life
The sound call that lost
Someone takes a well-researched job at a stable company, then the firm folds a year later in a downturn nobody saw coming. Looking back, it feels like an obvious mistake — but on the information available at the time, it was a reasonable bet. The decision didn't change; only the outcome did, which is exactly what outcome bias latches onto.
The 'I should have known' that wasn't
After a bad result, the warning signs suddenly look glaring: of course it was going to go wrong. But research on hindsight bias suggests that sense of obviousness is largely manufactured after the fact. At the time, the signals were genuinely ambiguous — the clarity only arrived once you knew the ending.
Real numbers in context
The evidence here is experimental rather than statistical, and the key result is qualitative: in Baron and Hershey's studies, identical decisions were rated differently based only on their outcomes, demonstrating outcome bias under controlled conditions. There is no clean population statistic for how often good decisions turn out badly — by their nature, that depends entirely on the odds of each individual decision.
What the research robustly establishes is the principle, not a number: decision quality and outcome quality are separable, people systematically conflate them, and hindsight makes past outcomes look more predictable than they were. The practical takeaway is directional — a single bad outcome is weak evidence about the quality of the decision behind it, especially when real uncertainty was involved.