Hacker News Reader

Comment by rajacombinator | Hacker News Reader

“How I predicted x”: flipped a coin, forgot about the times it was wrong!

I understand what you're saying about selection bias, but he has given a reasoned argument, providing a model that matched reality in multiple places. You're response of "it doesn't always work" is without value unless it provides an example of where it doesn't work.

His prediction would carry a lot more weight if he had written about it prior to the recession. A lot of people can say "I predicted that" after it happened, but those who have proof they can point to have a lot more credibility. All he has is that he "won a bet" after a "conversation with a friend."

Unless he was wrote about it ahead of time, we have no way of knowing whether he also made a similar bet with another friend in a different conversation to say that housing prices were going to continue towards the moon. If he had also structured the bet in which he got his school payed for and seemingly no downside if he lost, then he could have guaranteed free school, and looked like Nostradamus either way.

So it's quite possible that this is selection bias at work, and he "flips coins" on a regular basis and only brags about the ones that come up heads.

Exactly. And even if you have one million people making only a single prediction. Some people are going to be right. Especially because a downturn is like a 1 in 20 chance if you go by years. So out of a million people having one guess at the next downturn, about 200.000 will be right and can brag about it, without having any higher understanding about what happened (even if they will staunchly believe they understood something, because they were right).

What I observe though, is not large numbers of people making random guesses and exponentially fewer being validated through random events.

The people who become notable are those who predicted an economic crisis, for decades before circa 2008, and once they got notoriety for that, have continued to predict doom for the subsequent decade. Eventually they either die or are right again.

I think the article is not about being right about something but more about the concept that s/he learned about it in a virtual (extremely limited) environment.

IF s/he was confident about the prognosis, making really money in the marketplace would have given leverage x-fold higher than a bet between friends would allow.

Reply on news.ycombinator.com