Atomic Knowledge #18: Survivorship Bias
How ignoring failures distorts reality and harms decision-making
⏱️ Reading Time ≈ 2 min
Survivorship Bias is a cognitive distortion that occurs when analyses and conclusions focus solely on individuals or elements that “survive” a selection process, neglecting those that fail. This leads us to treat the successful subgroup as though it were the entire group. One key cause is that successes tend to be more visible than failures—or may be the only samples left after adversity strikes. Another factor is confirmation bias, which amplifies our tendency to notice data that supports what we already believe, making it even easier to overlook missing failures. There's also the common error of confusing correlation with causation: if every survivor exhibits a specific trait, it doesn’t automatically mean that trait caused their success—others may have had the same characteristic but never made it. Some of the main consequences include making decisions that aren’t fully rational, planning poorly, underestimating risk, overestimating one’s chances of success, and—perhaps most critically—losing the opportunity to learn from failure. Indeed, it’s often the unsuccessful experiences of others that prove most instructive to those who follow. To avoid or lessen survivorship bias, question what might be excluded from your analysis, actively investigate known failures, and verify data completeness. Knowing how to recognize and counter survivorship bias is essential for more accurate decision-making and preventing distorted conclusions. Remember: if you only look at the survivors, you lose sight of the bigger picture. Survivors might just be extremely lucky exceptions indeed.
Example: one of the most famous survivorship bias examples comes from World War II. Allied analysts were studying bomber planes that returned from missions, mapping out where the bullet holes were on these aircraft. The surviving planes showed heavy damage to wings and tail, but little to the engines and cockpit. The initial conclusion was to reinforce the areas that showed the most bullet holes (wings and tail). This was survivorship bias in action – they were only looking at planes that survived their bullets. Statistician Abraham Wald recognized the error: the reason no returning planes had engine hits is that any plane hit in the engine didn’t make it back. Those lost planes’ data were missing. Wald famously advised to armor the engines and cockpit, the areas where surviving planes had no damage, because those were the fatal hit areas for planes that were shot down. This insight saved countless lives and aircraft by correcting the bias. It highlights how considering the “invisible” failures (downed planes) completely reversed the conclusion – a perfect demonstration that focusing only on survivors (returned planes) led to a dangerously wrong plan.
👋🏼 Make the most of it! Until next time, S.
Deepen Your Knowledge
What Every Founder Needs to Know About Survivorship Bias (article)
Understanding Survivorship Bias: Implications for Research and Decision-Making (literature review)
The Impact of Survivorship Bias on Innovation Policy (article)
Survivorship Bias (YouTube video)
Why It's So Hard To Succeed - The Survivorship Bias (animated) (YouTube video)
Previously on Atomic Knowledge
Whenever I encounter an interesting concept—whether it’s a theory, speculative idea, formula, or law—I strive to deeply understand it and see how it connects within my knowledge network. Once I’ve grasped its essence, I distill it into a concise, no-frills note: simple, atomic, and memorable. To keep things sharp and focused, I stick to a “lazy” limit of 1,500 characters, give or take. These atomic notes, often described as mental models, have revolutionized how I understand and link ideas, fostering a more organic and interconnected expertise. Believing in the power of sharing, I’ve decided to make them public. Think of these notes as tools: mental models to keep in your back pocket for quick use or prompts to deepen your understanding when inspiration strikes. Subscribe if you’re curious—you won’t want to miss them.