Nassim Taleb : How Bad We Are at Predicting | Young Nassim Nicholas Taleb on Black Swan Events

[00:00:00]One thing I discovered is that the beauty, the power of economics is that we have plenty of data. When you have data, you can do some real analysis on the data that's completely unbiased because the data is going to be there and you can throw numbers at uh as a computer.

[00:00:20]So I looked at history to see if anybody did something like that and I discovered that one of your future speakers uh Neil Ferguson, a brilliant man and he uh uh wrote a paper showing that although we believe that the first war was predictable bond the bond market there's something called war bonds in UK and the bond UK bond did not predict it.

[00:00:49]So it cannot be that predictable if the bond market nobody told you know the bond market. Okay. So we have we have prediction markets and stuff like so so we realize now when you run if you dig into history how bad we are at predicting how bad we are at seeing things how bad our predecessors were at predicting.

[00:01:14]Before the discovery of Australia, we had no reasons to believe that swans could be of any other color but white or people in the old world. And effectively there was a uh an expression in uh medieval England. you'd sooner see a black swan than say for example uh it was it was like saying when pigs fly or when uh when [snorts] uh when I don't know when George Bush uh does something intelligent or something exception so there was an expression until we saw Australia and effectively one the sighting of a single bird destroyed millennia of confirmation so It was a uh posed as a logical problem by showing that there's no reason you cannot rule out a black swan because you haven't seen any. Okay. So my problem is not a logical question. My black swan is an event. It's not a bird.

[00:02:21]And it's an event that has three properties. The first property it is hard to predict. very difficult to predict based on information before its uh occurrence, prior information based on historical information. You have here a sample of black swans. The most interesting one is the tie. Someone's going to forecast the future to have to forecast that human beings 2,000 years away would constrict their blood supply with this device, for example. Okay? And attend meetings. Okay? So that that's very difficult to to predict.

[00:03:01]The computer was a black swan. It changed the world and nobody thought the computer could do anything. You know, it was initially used for combinatorics. I mean, the Watson from IBM did not think that this tool could have any use. The rise of religions, black swans, totally unpredictable.

[00:03:22]Uh Harry Potter is a black swan.

[00:03:27]A lot of cultural phenomena are black swans. [clears throat] The to me the most significant black swan and the one I'm going to focus next few minutes is the first war.

[00:03:42]The first war we had after Napoleon, we thought for about 100 years that the world became civilized and that you know people became uh conscious of uh of uh the need for peace and you had this devastating war, the biggest war, something that destroyed uh and of course it came in two volumes. You had volume one and then you had a sequel.

[00:04:07]So, [clears throat] so here we have black swans events [snorts] of low predictability, high consequence.

[00:04:18]But the most vicious part is the following one is that before the fact they're extremely predictable, but after the fact, you know what? We saw them coming. So we have this is what I call the retrospective distortion is these events are prospectively unpredictable retrospectively predictable. Why we even have disciplines [snorts] to make us to give us the illusion of understanding the world. You see have disciplines that make us misunderstand the world by giving us this illusion of predictability. history for example, economics, other such things, astrology.

[00:05:04]Okay. So we have and what what's the mechanism by which [clears throat] we sort of like have this illusion of understanding the world. The first one is what I call silent evidence.

[00:05:17][clears throat] before people think that the first war was predictable particularly if you I don't know if you have been to school high school and they uh discuss the the first war it appears to result from tension between the UK on one hand Austria and Germany on the other okay so you think that there's tension that led to war if you see tension then you can predict war but you're not looking at episodes of tension that did not lead to And there were a lot of episodes of tension before that that did not lead to war.

[00:05:54]And these episodes usually led to uh uh parties in bod and boden you see with opera singers a lot of champagne and they get drunk. All right. Kings get drunk. Plus they know each other so they know how to do it. So you have to realize that after the war you think tension causes war. But if you're in a in a champagne business, you really think tension causes uh you know drunkenness by kings when they make up.

[00:06:19]Okay. So we have that problem. So let me give you an an illustration of this inability we have in looking at what I call the silent evidence, the pool of evidence that did not lead to the same result. But a comment made by publisher about the success of the black swan.

[00:06:39]And he was explaining it was as follows.

[00:06:41]Look, it has an animal and a color on the cover. That explains the success.

[00:06:48]Now, okay, this is So, I went on when I heard that, I said, "Okay, plain. I'm going to take care of this guy." I looked on Amazon for how many books have animals and colors on their on in their title and on the cover that ended up flopping. Okay. And you had plenty of them. I found 69 books with a black swan in their titles. Okay, that were flops.

[00:07:16]You don't hear if you don't hear about them is because they're flops. We're not looking at evidence. Plus, of course, you have permutation pink uh elephant uh uh different colors, different animals.

[00:07:26]Okay, so plenty plenty of book like that that flocked. So, this is what I call the cemetery of evidence. And for the those of you interested in probability, it's a big probability problem that that because we compute probability based on those who survived, probabilities of survival, not based on the pool of those who started it. It's very this is very endemic in the way we analyze the world, the way we understand information, the way we perceive information. Uh decision making under uncertainty is completely dominated. Okay. with uh this uh mistake of of taking a pool of information and excluding the rest. So it's what I call silent evidence what you have it on Wall Street. You look at the winners and say they have skills. You don't look at people who have the same sets of skills who end up losing because these people don't write biographies. You see they don't say how how I lost a million dollars. They tell you how they made a million dollars. So you have that. So, [laughter] and then but the problem is historians, you know, people on Wall Street, you can understand that they're not that smart, but historians did not know that, okay?

[00:08:38]Or did not deal with it empirically.

[00:08:40]Now, I happened to have spent 18 years as a trader and I hated it. Okay? But I stayed there because it was fun particularly because you had economists around and people who make forecasts and you could make fun of them. So there was some some uh some uh there was some advantages to it. But one thing I discovered is that the beauty the power of economics is that we have plenty of data. When you have data you can do some real analysis on the data that's completely unbiased because the data is going to be there and you can throw numbers at uh at a computer.

[00:09:19]So I looked at history to see if anybody did something like that and I discovered that one of your future speakers uh Neil Ferguson, a brilliant man and he uh uh wrote a paper showing that although we believe that the first war was predictable bond the bond market there's something called war bonds in UK and the bond the UK bond did not predict market. So it cannot be that predictable if the bond market nobody told you know the bond market. Okay. So we have we have prediction markets and stuff like so. So so we realize now when you run if you dig into history how bad we are at predicting how bad we are at seeing things how bad our predecessors were at predicting by mechanism of course is called overcausation. Also there's some psychological mechanisms involved that you make an actual prediction. You have the outcome and then this is uh PowerPoint this all right this is remembered predictions okay upside down what you remember typically that you you remember what you remember having predicted is more consistent with what you observed so you don't remember what you actually predicted but you revise your memory of what you actually predicted continuously to make it consistent into his current events.

[00:10:55]Not only you do that with your prediction, but you also do that with your intentions which is a big problem in for us in adjusting because you would realize that if we didn't have this effect, people would know that they're very bad at predicting.

[00:11:09]The economics departments would be empty or commit suicide or something like that. You would have no no social science uh uh to speak of. Okay? People would turn cab drivers or something. So you would realize it is a psychological uh bias.

[00:11:25]Now let me talk about the black swan problem in in history in philosophy and I'm going somewhere with this particularly with my next uh next work. The first gentleman up there on the left the one who's a little horizontally challenged.

[00:11:43]His name is Hume. Okay. It's called Hume's problem but it's not his problem.

[00:11:49]It's because he wrote in English. It's a great idea to write in English. Then it could be remembered by people who write in English that it's a Hume effectively he took it from someone else. But Hume is worthy discussing because he was completely um annoyed with the black swan problem.

[00:12:09]Completely annoyed with that problem of induction. It was not called black swan at the time. It was called the problem of generalizing from finite sample or problem of induction.

[00:12:18]>> [clears throat] >> And what he did with it is very simple.

[00:12:22]He said, "You know what? I leave it for the philosophical cabinet."

[00:12:26]And in real life, I can't deal with it.

[00:12:28]He was a party animal. And his reaction to the black swan problem is to become even more horizontal. You see, [laughter] which he gained a lot of weight and then he died and he had a happy life in Edinburgh in Paris and Edinburgh. Okay.

[00:12:40]So let's forget about Hume because he could be of no help except [snorts] also to illustrate one thing that happened in philosophy is that increasingly philosophers became what I call domain dependent is they're good at talking about a problem in a classroom and then they forget about it when they leave the classroom and it's a bias statisticians for example don't understand statistics in real life they're good in front of the blackboard we know that from a lot of experiments the way I I discuss it in Black Swan is I [clears throat] talk about that that domain of dependence about the Reebok Club in New York where people go get in their gym clothes and then take the the the elevator to the to the stair masters and then get on the stair masters and log their 112 floors stories and then go and then stop and then take a log of it. Okay, so we have domain dependence not recognizing something in texture of real life. So, let's forget about this guy. Next to him, there's a French guy, but the French don't know about him or at least forgot about him, wanted to forget about him. He is a bishop called Ui.

[00:13:50]And he dealt with the black swan problem by becoming extremely religious.

[00:13:55]He he not liking science. And of course, we had the enlightenment that became pro-science and uh with all the tragedies that we have coming from it.

[00:14:05]And uh uh so he is forgotten. He became very religious. The gentleman here is Al Gazelle. He was the Arabic the Arabic language uh I mean the sorry the Arabs call him an Arab. The the Persians call him a Persian. So uh he's a Arabic language philosopher who uh attacked the classical philosophers uh by writing a treatis called uh the on the incoherence of philosophy very famous and he created Sufi Islam out of his thing. So the black point problem led these two gentlemen uh to become extremely religious.

[00:14:44]Now the one on the right is my hero or you know I think we don't really know if he existed or if you know but what he represents uh would make him my hero is sexist empiricus he's not in philosophy books not very common philosophy books he was he had two things he was a skeptical gentleman who phrased the problem induction just the way Hume later on repeated it 2 century uh BC uh uh AD And his second attribute he was a doctor.

[00:15:18]So there was a school of medicine of decision maker under makers under uncertainty called the empirical doctors who were damn good doctors.

[00:15:28]They did not like theories. Did not like to generalize. Did not like to extend into unobservables. Okay. did not like to make an you know go from from what they know to what they don't know extremely careful they call themselves empiricist they did not like to generalize and these people were extremely successful unfortunately uh they were uh completely you know medicine became intellectual rationalists that we thought we understand human body so these people were out of business for about uh 15 centuries before medicine came back via the barbers. Okay, their ideas came and if you guys are alive today, it's because these guys or their ideas or because barbers, not because of intellectual doctors, uh the contribution of intellectual doctors. Finally, there's a gentleman.

[00:16:27]I'm sure you recognize him particularly if you live in Berkeley. All right, so this is Karl Marx. All right. So, uh Karl Marx had this idea of wanting he wanted you know in his uh thesis on um uh fearba he wanted to say he said the philosophy [clears throat] you know was just talk let's do something with it unfortunately but so his his idea was to turn knowledge into action.

[00:16:57]So my idea is the exact opposite.