Korva is a scroll of the internet’s best ideas.
Learn more from successes than failures
Most things don't work, and there are a lot of reasons why, so it is hard to find the correct cause of any one failure. The usual takeaway is something generic about grit and determination, and a thing not to do. It echoes the opening line of Anna Karenina: all happy families are the same, and all unhappy families are unhappy in their own way.
Understanding what parts of Y Combinator really worked, or what parts of OpenAI really worked, and applying those lessons going forward has been much more helpful than applying the anti-lessons of what didn't work. The lessons from failure were either fairly generic and already known, or they got in the way of something else.
Every data point is still worth learning from, failures included. The lessons from success are the ones that deserved more application, and that holds for a lot of people.
Why is the cause of failure so hard to find?
Failures leave behind fewer clues than successes, and the clues point in many directions. Survivorship bias adds to the problem: people study the cases that remain visible and miss the ones that vanished. During World War II, the statistician Abraham Wald analyzed bullet holes on bombers that returned from missions. He concluded that armor belonged where the returning planes showed no damage, because planes hit there had not come back.
How does the Anna Karenina principle apply beyond families?
The principle holds that success requires many things to go right, while failure needs only one thing to go wrong. Jared Diamond used it in his 1997 book Guns, Germs, and Steel to explain why so few wild animals were ever domesticated. A candidate species needed the right diet, growth rate, temperament, and breeding habits all at once. Of the large mammals Diamond surveyed, only a small fraction met every requirement.
Why do postmortems of failure mislead?
Once an outcome is known, people tend to believe it was predictable all along. Baruch Fischhoff demonstrated this hindsight bias in a 1975 study in which participants told the result of an event rated it as far more likely than those who were not told. That makes a failure's story seem tidier than the events were. The explanation that feels obvious afterward is often built from the outcome itself.
More from David Senra
