Mr. Black Swan

On Thursday I went to see a public lecture by the guy who wrote The Black Swan: The Impact of the Highly Improbable. The book was amazing, so I was hoping it would be a good lecture. The author has a pretty distinctive personality, so I was excited to see him in person. His name's Nassim Nicholas Taleb.

Taleb was pretty much as I imagined he would be, although he was more soft-spoken than I thought he would be - in the book, you get the impression that he frequently flies into a shouting rage. He was fairly restrained here. And he had some recent examples of randomness to underscore his points - for example, on Martin Luther King's day this year the stock markets in Europe crashed. The New York Times reported the markets crashed because of "recession fears," but Taleb pointed out there were recession fears the day before and the day after and it didn't crash. In fact, sometimes it rebounded heavily, amidst recession fears. No, the real cause of the market crash was something much more unpredictable, and much more random.

It was because a rogue trader at Society General, one of the biggest banks in all of Europe, had used 50 billion euros of the bank's money to make bets on the stock market - bad bets. The bank wanted to get rid of the stuff he had bought, because it was certainly going to drop in price eventually. Anyway, for some reason they decided to do it on Martin Luther King's day, as quickly as possible, perhaps not realizing that the USA stock market was closed for business that day. So this bank goes to the stock market and tries to sell 50 billion euros worth of stuff, but half the world's buyers, the Americans, are taking the day off. So not enough people wanted to buy the bank's stuff. And, in the stock market, when no one wants to buy something you're selling, you lower the price. So they had to drastically cut the prices of 50 billion euros worth of assets, and this started a panic and the markets crashed.

Not because of recession fears, but because one person at one bank in Europe committed mass fraud, and then the bank picked a stupid day to try and undo the damage. Random. And some people think that, maybe if the markets hadn't crashed that day, then maybe we would be thinking "whew! We dodged the recession bullet!"

His other good example used a chart about Hillary Clinton.

This is a graph from Intrade.com, a website were people bet on, well, anything. The idea is that, since you have tons of people with money at stake, this is a really accurate predictor. You get the wisdom of the crowd, and you know everyone is actually saying what they think is most likely to happen, because they've put their money down on it. This is a graph of the probability that Hillary Clinton will be the democratic nominee, based on the bets from this website. As you can see, everyone though she was sure to win, until she lost Iowa in early January. Then everyone thought she was sure to lose, until she won New Hampshire. Then everyone thought she was sure to win until she did worse than expected on Super Tuesday. Then, after she won Ohio and Texas, people think she has a better chance of winning. The point is, obviously these guys can't predict anything. They have money at stake, but they don't know any more than the rest of us. The point is, prediction is really hard in the real world, and we think we understand more than we do.

And one more cool thing he said was this.

He put up a picture of the earth and said, look, this is the mother of complex systems. If we can't even predict little things, how can we think we know what's going to happen if we mess with this thing. So, let's error on the side of caution.

So that was cool. But the lecture had a lot of problems.

First, the guy's microphone died four times. Seriously. He had to switch from a wireless clip on his jacket, to one at a podium, to a wireless handheld, to a corded handheld, to a second corded handheld. And there was all this feedback. He didn't get through his whole lecture. because of this, and I think he would have been a lot more shout-ey if he hadn't had to keep switching microphones.

Second, the seats in this auditorium were the smallest ones I've ever seen in my life. Far worse than any train or plane. If we had another four inches, my knees would still have been jammed right up against the back of the chair. I switched seats to an aisle row, and sat completely sideways the whole time.

Third, it was a public lecture so we had to contend with crazy people at question time. This was a major problem at ISU too. There are these nutcases with elaborate conspiracy theories and they always come to these lectures to ask the speaker some question about their own conspiracy theory. And it's always a huge, long, rambling question that no-one understands, especially not the speaker. This time we had two of these guys. The first guy was in the middle of his awful, horribly long speech about knowledge and human psychology, when from somewhere else in the auditorium some guy yells out, "YOU CAN'T PREDICT IT!!!! THAT'S THE WHOLE POINT!!!!!"

Later the guy had another chance to be a loud crazy person. Some college kid was asking a question about random things, like 9/11, and as soon as he said 9/11 this guy yells out, "SEPTEMBER 11TH, 2001!!! WE NEED LINGUISTIC ACCURACY HERE PEOPLE! WE'VE BEEN CORRUPTED BY THE MEDIA AND BRANDING OF THIS PHENOMENON!!!..." I don't even remember what else he said. It was just gibberish.

Then, the moderator says, "Please sir! Can you let the speaker answer the question!"

To which the guy answers, "YES, OF COURSE, IT'S JUST THAT HAS TO BE FLAGGED UP! IT'S VERY IMPORTANT AND WE CAN'T LET THAT LANGUAGE SLIP!"

And the moderator says, "No, you don't have to flag that up."

So the kid goes back to his question, careful to say, "September 11, 2001" and the guy yells out, "THANK YOU!"

Ay ay ay.