“Predictions are hard, especially about the future”

The Sunday Fail
March 2nd, 2014

Your Daily Quote: “Predictions are hard, especially about the future” ~ Niels Bohr

Your Daily Read: “A day after U.S. intelligence said there would be no Russian invasion of Ukraine, Putin’s troops started coming over the border.”

I’ve made some good predictions in my time.  I noticed the housing bubble in 2002.  In December 2012, almost a year before the federal exchange launched, I said that given how late they were making decisions, I thought there was “approximately a 0% chance” that it would be ready on time.  I was a Groupon skeptic from the first moment the business model was explained to me.

But every time I’m tempted to break my arm patting myself on the back, I remember that I, like our intelligence experts, have botched more than a few important predictions.  How about Google stock, which I thought was overvalued at the IPO?  Or my confidence that we would find weapons of mass destruction in Iraq?  How about my risibly low estimates of what that war would cost?  Or my belief that Democrats would never pass Obamacare because doing so was so clearly political suicide?

(And it was political suicide; a lot of Democrats lost their seats because of that vote.  But the legislators who made the vote for some reason couldn’t see the danger.)

Luckily, one of the comforting things I discovered as I wrote my book is that I’m not alone. Phillip Tetlock’s ground-breaking work on expert prediction shows that even bona fide experts with sterling credentials get it wrong an astonishing percentage of the time.  They generally do better than random chance.  But even the best predictors don’t do that much better.

Yet if you think about it, this is not surprising. A realistic model of the universe is . . .  the universe.  Since we can’t build that in a box and watch it run, we need some simplifying assumptions.  Which means we are at risk of simplifying important details right out of the picture.

It’s tempting when experts make mistakes like this to point and laugh.  Especially since we tend to remember our own good predictions and forget our bad ones, making it seem as if we must be much smarter than the people around us, and those distant idiots in the US government.  But that is false vanity–dangerous false vanity, because it makes us think our own judgment better than it is.  This inability to correctly predict the future is not a problem of special idiots who couldn’t see the obvious; it’s a problem of the human condition.

Your Daily Smile: These animals will make you love your imperfections.

In case you missed it:  here’s the video of me talking about my book at Cato with Arnold Kling as respondent.

Want to get this by email?  Sign up for my email newsletter, which includes the Daily Fail post, plus updates on my other writing.

2 thoughts on ““Predictions are hard, especially about the future”

  1. I’d be very interested to hear your call on Bitcoin ( not the currency as such, but the underlying protocol of the blockchain )

    Many of us believe that this is the start of a completely different kind of internet activity. Digital currency is essential if ideas like ‘intelligent machines’ are able to function independently ( as they will in the future ) and a decentralised currency ( outside anyone’s’ influence and control ) is key to this happening to give this currency – payment exchange – world wide acceptance. that is why bitcoin is such a game changer for the future

    your call ?

  2. I used to think that electronic currency would enable what you say–decentralized money. I no longer do. Rich world governments are in the process of killing tax havens, the US killed online poker, and they will kill this if it gets big.

    But how can they kill a decentralized currency, you ask? Easy: they can make it illegal. Who cares, you ask? Well, most businesses will not accept an illegal currency, there will be no legal exchange to convert it, and it can’t join the legitimate electronic payment streams via the banking system. At which point, you have a very illiquid asset that is only good for buying illicit materials, something like the prison money systems, which are vastly inferior to normal money and only used where normal money is banned or so hyperinflated as to be actually worthless. Neither issue obtains in any rich-world country.

    Moreover, as I understand it, the blockchain turns out to have some actual disadvantages over cash: if they can attach your transactions to a person, they have a record of every single thing you ever paid for, and who you paid it to. So if they manage to catch you buying drugs on the internet, for example, they don’t just have you on simple possession, they way they would if you’d paid cash.

    Could bitcoin become something like a better Western Union–an accounting system for doing (highly regulated) legal transactions in local currency? Quite possibly, though I think that there may some drawbacks–again, as I understand it, the processing load needed to keep track of the blockchain scales as a quadratic function of the number of nodes, so that the more common the currency becomes, the more expensive it becomes in computing power. Right now, I am under the impression that this is being subsidized by bitcoin mining, but that will eventually stop when the last bitcoin is mined. At that point, it will no longer be so cheap to run this network, especially as you layer on top the regulatory and insurance functions that make current networks quite expensive.

    But if bitcoin becomes a serious rival to normal money that operates outside of government control, the governments will shut it down. The internet simply isn’t nearly the libertarian tool for state evasion that was hoped 20 years ago.

Comments are closed.