
Sign up to save your podcasts
Or
https://astralcodexten.substack.com/p/crowds-are-wise-and-ones-a-crowd
The long road to MoscowThe “wisdom of crowds” hypothesis claims that the average of many guesses is better than a single guess. Ask one person to guess how much a cow weighs, and they’ll be off by some amount. Ask a hundred people and take the average of their answers, and you’ll be off by less.
I was intrigued by a claim in this book review that:
You can play “wisdom of crowds” in single-player mode. Say you want to know the weight of a cow. Then take a guess. Now throw your guess out of the window, and take another guess. Finally, compute the average of your two guesses. The claim is that this average is better than your individual guesses.
This is spooky. We talk a lot about how to make accurate predictions here - and you can improve your accuracy on anything just by guessing twice and averaging, no additional knowledge required? It’s like God has handed us a creepy cow-weight oracle.
I wanted to test this myself, so I included some relevant questions in last year’s ACX Survey:
4.8
123123 ratings
https://astralcodexten.substack.com/p/crowds-are-wise-and-ones-a-crowd
The long road to MoscowThe “wisdom of crowds” hypothesis claims that the average of many guesses is better than a single guess. Ask one person to guess how much a cow weighs, and they’ll be off by some amount. Ask a hundred people and take the average of their answers, and you’ll be off by less.
I was intrigued by a claim in this book review that:
You can play “wisdom of crowds” in single-player mode. Say you want to know the weight of a cow. Then take a guess. Now throw your guess out of the window, and take another guess. Finally, compute the average of your two guesses. The claim is that this average is better than your individual guesses.
This is spooky. We talk a lot about how to make accurate predictions here - and you can improve your accuracy on anything just by guessing twice and averaging, no additional knowledge required? It’s like God has handed us a creepy cow-weight oracle.
I wanted to test this myself, so I included some relevant questions in last year’s ACX Survey:
4,221 Listeners
13,357 Listeners
26,400 Listeners
2,388 Listeners
87 Listeners
3,755 Listeners
86 Listeners
388 Listeners
128 Listeners
199 Listeners
47 Listeners
90 Listeners
75 Listeners
145 Listeners
114 Listeners