Survivorship Bias: Why We Only Learn From the Winners

SPEAKER_1 (0:00)

Picture this, you are standing in ancient Greece, like right in front of a grand temple, and you're hanging out with the philosopher Diagoras of Melos.

SPEAKER_2 (0:09)

Oh wow, okay, setting the scene.

SPEAKER_1 (0:10)

Yeah, exactly. So someone walks up to Diagoras, and this guy is clearly feeling a huge sense of philosophical superiority.

He points to these magnificent paintings hang on the wall, and they depict these sailors caught in terrifying violent shipwrecks.

SPEAKER_2 (0:27)

Like the classic divine intervention trope.

SPEAKER_1 (0:29)

Exactly. The story goes that these sailors prayed to the gods during the storms and miraculously survived. So the man points at the artwork and essentially says to Diagoras like, look at this, you who think the gods don't care about us, how do you explain all these people who are saved by divine favor?

SPEAKER_2 (0:46)

Which is, I mean, it's essentially the antiquity equivalent of a mic drop, right? This guy really thinks he's just provided absolute empirical proof of the gods.

SPEAKER_1 (0:55)

Completely. But Diagoras, he doesn't even blink. He just looks at these beautiful paintings of those survivors, turns back to the guy and just asks, where are the pictures of those who prayed but drowned?

SPEAKER_2 (1:05)

Ouch. Yeah, that is such a profound way to flip the argument.

SPEAKER_1 (1:10)

Right. Diagoras was pointing out this massive mathematical illusion that has tricked humans for, well, for millennia.

So welcome to this deep dive. Today, we are opening up a huge stack of sources to explore this reality-distorting statistical error known as survivorship bias.

SPEAKER_2 (1:27)

It's a fascinating topic because it's a cognitive blind spot that is just deeply, deeply embedded in human psychology. We are hardwired to do this.

SPEAKER_1 (1:36)

We really are. And our mission today is to figure out how this exact same invisible flaw, you know, looking only at the winners, is currently manipulating your stock portfolio, your medical data and even your dating life.

SPEAKER_2 (1:48)

Yeah, because we instinctively look at the successes, right? We study the things that made it through whatever filters in place, and we just completely forget to consider the day that they got erased by failure.

SPEAKER_1 (1:58)

Okay, let's unpack this, because to truly understand the mechanics of this blind spot, we need to jump from ancient shipwrecks to a very modern literal life or death puzzle. Our sources trace the formalization of this concept right back to World War II.

SPEAKER_2 (2:11)

Specifically, the Statistical Research Group, the SRG, which was stationed over at Columbia University.

SPEAKER_1 (2:16)

Right. And the SRG was just this extraordinary brain trust. The military pulled together some of the greatest mathematicians and statisticians of the time and essentially said solve our tactical problems.

SPEAKER_2 (2:28)

Yeah. I mean, these guys weren't on the battlefield with rifles. They were fighting the war with probability and statistics.

SPEAKER_1 (2:35)

Exactly. And one of the most pressing logistical nightmares the military was dealing with was the sheer number of bombers getting shot down over Europe. They obviously needed to put heavier armor on the planes to protect the crews, but they were fighting basic physics.

SPEAKER_2 (2:51)

Right. Because heavy armor grounds a plane, you can't just turn it into a flying tank. You can only armor a really small fraction of the aircraft before it literally becomes too heavy to take off.

SPEAKER_1 (3:00)

Right. So they had to be incredibly precise about where those armor plates actually went.

So the military brass, they conducted this rigorous data collection process. When the bombers returned from combat missions, they meticulously inspected them and mapped out every single bullet hole.

SPEAKER_2 (3:16)

And the data they got back was remarkably clear, right? It showed this massive pattern, the fuselage, the outer edges of the wings, the tail section. They were just absolutely riddled with bullet holes.

SPEAKER_1 (3:28)

Which leads to a conclusion that, honestly, it seems entirely rational on the surface. The generals looked at the maps and assumed, okay, the areas absorbing the most fire are obviously the areas that need the most protection.

SPEAKER_2 (3:40)

Yeah, I mean, that is the intuitive leap almost anyone would make. Hey, there's damage here. Let's put armor over the damage.

SPEAKER_1 (3:45)

So, their mandate was to put the heavy armor over the fuselage, wings and tail. But this is where the statistician, Abraham Wald, steps in. And he essentially tells the military, you are looking at a completely inverted reality here.

SPEAKER_2 (4:00)

Yeah, Wald's insight is just brilliant. He examined the exact same planes, looked at the exact same data, and told the generals to put the armor in the places where there are absolutely zero bullet holes. He said, you know, armor the engines in the cockpit.