# Survivorship Bias
1943. A cramped war room.
Abraham Wald is studying bomber damage reports. The air smells like sweat and cigarette smoke.
The military brass is confident: armor the wings... armor the fuselage. That's where the bullet holes are. Makes sense, right? Protect the spots that get hit.
Except Wald doesn't nod along.
He's looking at the planes that made it back, sure. But in his mind? He's also cataloging an invisible fleet. The ones that didn't come home.
The bullet holes, he realizes... are a trap.
These planes made it back *despite* the damage in those spots. The ones that got hit in the engines? The cockpit? They're at the bottom of the Channel.
The bullet holes are a map of survival. Not mortality.
So Wald, the statistician, tells them: armor the places with *no* bullet holes. Protect the negative space.
This flips the whole room.
The military's plan wasn't just off. It was backwards.
And in that moment, Wald carved out the textbook definition of survivorship bias. Once you see it? You can't unsee it.
Survivorship bias is what happens when you study the winners... but ignore the losers.
You're trying to figure out success, but you're using a data set that's rigged. The failures? Deleted.
And that's usually where the truth lives.
Wald got this because he thought in populations. Not individuals.
The bombers in the reports weren't "the bombers." They were *some* bombers. A subset. A biased sample of survivors.
And survivors lie... just by existing. Their existence makes you think they're the rule. Not the exception.
Think about evolution.
Darwin's masterpiece gives us survival of the fittest. So we look at giraffes with their skyscraper necks. Hawks with their sniper vision. Cheetahs that sprint faster than cars in school zones.
These traits are so optimized. So *perfect.*
But that's because the species with bad necks? Blurry eyes? Slow legs? They didn't make it.
We're looking at life through the lens of survivors. Reverse engineering the rules from what's left.
And sure, it mostly works. But it means we're always staring through a filter of the winners.
Here's where it gets expensive: finance.
In the nineteen eighties, mutual funds seemed like a solid bet. Respectable returns. Decent track records.
But in 1992, researchers did something sneaky. They added back the funds that had *closed.* The ones that bombed so hard they vanished.
Suddenly? The numbers shifted.
Average performance dropped by more than one percent annually. That's the kind of percentage point that compounds... into a retirement plan going off the rails.
The surviving funds weren't lying. They just weren't the full story.
The failures had been erased. And everyone's spreadsheets were wrong.
And this isn't just about war stories or Wall Street.
Survivorship bias is shaping your life right now.
Like those startup success stories. The college dropout who built an empire. The garage tinkerers who rewired the planet. Apple, Google, Facebook. The mythology is everywhere.
But here's the gut punch: researcher Scott Shane dug into the numbers. Only *one in ten* startups makes it past a few years.
One. In. Ten.
Meanwhile, the headlines make it look like success is normal... and failure's the weird outlier.
Picture a tech conference today. Someone on stage, TED Talk energy, talking about their scrappy rise from idea to IPO. They've got the crowd buzzing. Everyone's thinking: *that could be me.*
But for every person on that stage? Nine others are at home, staring at their laptops, applying for jobs to pay off the debt from their own failed startup.
They're not at the conference. They're not in the story.
So the story you're hearing? It's not reality.
It's survivorship bias in a two hundred dollar hoodie.
This bias isn't just about numbers. It's baked into how our brains work.
In 1973, Kahneman and Tversky coined something called the availability heuristic. We over-weight information that's vivid. Easy to access. Easy to remember.
Success stories are vivid.
Failure is invisible.
So we build our mental models out of what survived the journey to our attention.
"What you see is all there is."
That's Kahneman's line. Haunting, isn't it?
Your brain doesn't know what it doesn't know. It just grabs the data in front of it... no matter how skewed.
Survivorship bias is one big, systematic way the world misleads you.
Medicine's got this problem too.
Clinical trials follow patients, measure outcomes, publish the results. But here's what happens: some patients drop out. Side effects too rough. Treatment too intense. They just vanish from the data.
The final analysis? It's based on the survivors of the trial process.
And the results? They'll look rosier than they would if you counted everyone who started.
Nobody's lying. It's just the math of who's left in the room.
Sports analytics? Same problem.
We study the LeBrons and Serenas to figure out what makes them great. Training regimens, psychology, childhoods.
But thousands of kids trained just as hard. Had similar talents. And didn't make it.
Maybe the thing we think is the secret sauce? Is actually just common. We don't notice it in the failures because we're not looking at them.
The data set only includes the winners.
History itself is a survivorship bias machine.
The civilizations we study? They're the ones that left records. Built pyramids. Wrote epic poetry.
Failed societies? Quiet exits from the historical record.
Same with famous artists. We celebrate the ones whose work survived. But how many equally brilliant painters, sculptors, writers... lost their work to fire, floods, indifference?
We're studying history through the lens of what didn't disappear.
And that's not the same as what actually happened.
So what do we do?
How do you fight bias that's baked into the structure of the world?
First: look for what's missing.
When you see a success, ask about the failures. When you see a pattern, ask about the exceptions.
This isn't natural. Your brain doesn't want to do it. But you can train yourself to look at the negative space.
Second: be suspicious of advice from successful people.
Not because they're lying. But because they don't know which parts of their story mattered. They felt like all ten steps were necessary... but maybe only three were.
Success feels logical from the inside. But their experience is just one data point in a biased sample.
And third: watch out for survivorship bias in the machines we're building.
AI systems are trained on data. And if that data only includes successes? The algorithms learn the same skewed story.
It's already happening.
Hiring models that only see candidates from elite schools. Loan algorithms trained on previously approved applicants. Medical AI that under-diagnoses certain populations because the training data didn't include their failures.
The biases of the past are getting baked into the code of the future.
Wald's insight was poetic.
He solved a problem by thinking about what wasn't there. The bombers with engine damage didn't make it back... so there were no holes in the engines to count.
The solution was in the negative space.
That's the skill.
Learning to see the gaps. The missing data. The failures that got erased.
The planes that didn't come home.
So here's your move this week.
Pick something you're trying to get better at. Doesn't matter what.
Now find stories of people who *failed* at that thing.
Not to demoralize yourself. But to get the full picture. Post-mortems. Retrospectives. The "here's what I wish I'd known" essays.
You're hunting for the data that survivorship bias usually hides.
Because the failures? They've got something to teach you that the success stories never will.
They'll show you the real contours of risk. The truth... not the highlight reel.
The bullet holes tell you what can survive.
But the empty spaces?
That's where the danger actually lives.