I don’t really get this. It seems like both types of prediction matter quite a bit.
The only way I can interpret it that makes sense to me is something like:
Thinking really hard about the infinity of things that might happen this week is an unproductive way to generate predictions, because the hypothesis space is too large and you’re just going to excessively privilege some salient hypothesis.
Is he giving advice about making correct predictions given that you just randomly feel like predicting stuff? Or is he giving advice about how to predict things you actually care about?
Is he giving advice about making correct predictions given that you just randomly feel like predicting stuff? Or is he giving advice about how to predict things you actually care about?
The latter. Specifically predicting high impact events.
The parent is a concrete example of selection (or survivor) bias. Picking post factum one case which turned out to be right (and ignoring unknown but possibly large number of cases which turned out to be wrong and faded into the dark pit of obscurity) does not help you predict anything.
Consider a forecast: the stock market will crash. No idea when, but at some point it will. It is a safe prediction to make? Yes, it is. Is it a useful prediction? No, it is not.
Taleb’s advice is good for burnishing one’s reputation as a psychic. It’s not so good for making actionable forecasts.
To recall a well-known remark by Paul Samuelson,
To prove that Wall Street is an early omen of movements still to come in GNP, commentators quote economic studies alleging that market downturns predicted four out of the last five recessions. That is an understatement. Wall Street indexes predicted nine out of the last five recessions!
ETA: So the guy sold his Washington DC condo in 2004? That looks to have been a pretty poor decision.
Of course it’s a useful bloody prediction! It means you shouldn’t put yourself in a position where any stock market crash will kill you or drastically lower your standard of living.
LOL. In that case I have a lot of useful predictions to make:
Inflation will wipe out the savings held in cash
There will be riots in major cities
Public transportation will have a horrible accident with many people killed
Some food in a supermarket will turn out to be tainted
There will be a serial killer on the loose
You will die
...I can easily continue...
P.S. Maybe you should mention your advice to all the financial gurus on LW who insist that the only place to put your money into is an equities index fund X-D
Did you just link to the change in the housing market over the past year? Washington Post:
D.C.’s median sale price soared to $460,000 from $405,000 in March 2012, an increase of 13.6 percent year over year. (sic)
My link:
Sold: $445,000 in May 2004
He would prefer to own if it made financial sense. “I expect we’ll probably end up buying again, but only when prices adjust,” Baker says.
Let’s subtract the $1000 he paid for the best argument against the existence of a housing bubble. On the face of it, you appear to be arguing with a man who made $39,000 cash by betting on the obvious—though the actual number may of course be less.
Your general argument seems to deny the usefulness of hedging.
Did you just link to the change in the housing market over the past year?
There is a multi-year graph of real estate prices on that web page, if you click the “Max” button you will get a plot of prices from August 2004 till today.
Your general argument seems to deny the usefulness of hedging
No, my general argument denies the usefulness of forecasts which don’t provide time estimates other than “at some point in the future”.
Let me offer you three more examples of such forecasts:
The website does work when I enable cookies, and it says he sold his apartment for much more than the median price. I think it also supports the claim that after buying a house, he had a profit left of roughly 10 percent of that house’s value (the amount of equity he supposedly said he wouldn’t mind losing post-purchase).
Your general argument seems to misrepresent Taleb. Again, we have here a case of someone doing pretty well by focusing on the predictions you can make. (His profit was likely sub-optimal, but that sounds like an example of a prediction you can’t make.) And hedging can indeed protect you against the events you keep weirdly suggesting are useless to think about.
I don’t really get this. It seems like both types of prediction matter quite a bit.
The only way I can interpret it that makes sense to me is something like:
Is he giving advice about making correct predictions given that you just randomly feel like predicting stuff? Or is he giving advice about how to predict things you actually care about?
The latter. Specifically predicting high impact events.
People predicted the housing bubble collapse using Taleb’s reasoning.
Did someone other that Eugene Nier downvote this? If so, how is the parent not a concrete example of “how to predict things you actually care about?”
The parent is a concrete example of selection (or survivor) bias. Picking post factum one case which turned out to be right (and ignoring unknown but possibly large number of cases which turned out to be wrong and faded into the dark pit of obscurity) does not help you predict anything.
Consider a forecast: the stock market will crash. No idea when, but at some point it will. It is a safe prediction to make? Yes, it is. Is it a useful prediction? No, it is not.
Taleb’s advice is good for burnishing one’s reputation as a psychic. It’s not so good for making actionable forecasts.
To recall a well-known remark by Paul Samuelson,
ETA: So the guy sold his Washington DC condo in 2004? That looks to have been a pretty poor decision.
Of course it’s a useful bloody prediction! It means you shouldn’t put yourself in a position where any stock market crash will kill you or drastically lower your standard of living.
LOL. In that case I have a lot of useful predictions to make:
Inflation will wipe out the savings held in cash
There will be riots in major cities
Public transportation will have a horrible accident with many people killed
Some food in a supermarket will turn out to be tainted
There will be a serial killer on the loose
You will die
...I can easily continue...
P.S. Maybe you should mention your advice to all the financial gurus on LW who insist that the only place to put your money into is an equities index fund X-D
Did you just link to the change in the housing market over the past year? Washington Post:
My link:
Let’s subtract the $1000 he paid for the best argument against the existence of a housing bubble. On the face of it, you appear to be arguing with a man who made $39,000 cash by betting on the obvious—though the actual number may of course be less.
Your general argument seems to deny the usefulness of hedging.
There is a multi-year graph of real estate prices on that web page, if you click the “Max” button you will get a plot of prices from August 2004 till today.
No, my general argument denies the usefulness of forecasts which don’t provide time estimates other than “at some point in the future”.
Let me offer you three more examples of such forecasts:
The stock market will go up 20%
The stock market will go down 20%
The stock market will stay flat for a while
The website does work when I enable cookies, and it says he sold his apartment for much more than the median price. I think it also supports the claim that after buying a house, he had a profit left of roughly 10 percent of that house’s value (the amount of equity he supposedly said he wouldn’t mind losing post-purchase).
Your general argument seems to misrepresent Taleb. Again, we have here a case of someone doing pretty well by focusing on the predictions you can make. (His profit was likely sub-optimal, but that sounds like an example of a prediction you can’t make.) And hedging can indeed protect you against the events you keep weirdly suggesting are useless to think about.
If I may point you to the first paragraph of this post..?
I don’t believe I said anything at all about what’s useful or useless to think about.
It didn’t really bear on whether one should try to predict events of uncertain occurence.