Book: Black Swan. Page 1

The perversity of all state government plans are obvious: they can't create the possible margin of error (tolerance for error).

Like many biological variables, life expectancy - the value of Srednekanskaya, that is subject to the ordinary accidents. It is not scalable because the older we get, the less chance we have to move on. In a developed country tables insurance companies predict the newborn girl's death in 79 years. When it will celebrate the 79th birthday, the expected duration of her life in the typical case will be 10 years. At the age of 90 years, she can expect to 4.7 years.
In 100 years - 2.5 years. If she miraculously live to 119 years old, it will be about 9 months. As it crosses another threshold, the number of additional years decreases. This is an illustration of the main properties of random variables, described by a "Gaussian curve". The older a person is, the less additional years he has in reserve.
With human plans and projects, the situation is different. They are often scalable, as I said in Chapter 3. And in the case of scalable, i.e. crinastanciu, variables, you will get exactly the opposite effect. For example, it is assumed that the project will be completed in 79 days (take the same number as in the example with the age of the woman). If on the 79th day the project is not completed, you will need to take him for another 25 days. 90-day - about 58. 100-th - 89. On 119 -149. If the project is not completed on the day number 600, you will need 1590 days. As you can see, the longer you wait, the longer you have to wait.

Here has a classic mental mechanism - the so-called anchor effect. You tame your fear of uncertainty, thinking and clinging to it as an island in the middle of nowhere. This effect was discovered by the fathers of the psychology of uncertainty Daniel Kahneman and Amos Tversky when they were just starting to explore misconceptions in the use of heuristics. The effect works as follows. Kahneman and Tversky offered volunteers to rotate the roulette wheel. Volunteers looked at the number, knowing it by accident. They were then asked to mention the number of members of the United Nations of African countries. The response correlated with the number drawn.
Ask someone to call you the last four digits of your social security number, and then the approximate number of dentists in Manhattan. You will see that by activating the consciousness of the interlocutor a four-digit number, you force him to relate his assumption with this number.

In the not so distant past, say the pre-computer era, the plans remained vague and qualitative (non-quantitative sense), had to make a mental effort to keep them in mind, and to do the layout for the distant future was a real pain. To plan, required pencils, erasers, desks, littered with papers, and a huge trash can. Accounting add love to a tedious and monotonous work. In short, planning was time-consuming and unpleasant activity related to the same with constant doubt.
But the introduction of spreadsheets has produced a real revolution. Using ekselevskoy table computer literate person can effortlessly renew the "sales plan" to infinity. On paper, on a computer screen or, more trenchant, in powerpointonly format the plan takes on a life of its own, loses its vagueness and abstractness, "beastlies", as the philosophers say, is filled with concrete, starts a new life as a material object.
My friend Brian Hinchcliffe, while we're together sweat in the local gym, put forward the following idea. Perhaps it is the ease with which you can design a future, multiplying the cells in these programmes-tables, has led to the emergence of an army of soothsayers who confidently issue long-term forecasts (without leaving a "tunnel" of their own assumptions).

Plans fail because of the so-called "tunneling", the lack of attention to areas of uncertainty that are beyond the scope of the project.

To survive, the institution must convince yourself and others that it has the "vision".

Scholarship without erudition and natural curiosity leads to narrow thinking and the fragmentation of disciplines.

When an economist fails to predict the next crisis, he often writes off all on an earthquake or a revolution, stating that he is not an expert on geodesy, meteorology and political science, is to integrate these areas in their research and to recognize that his region cannot exist in isolation from the other.

The problem is that we focus on those rare cases where our method works, and almost never - numerous examples of failure.

The most interesting analysis of how academic methods work in real life, was carried out by Spyros Makridakis. He arranged the competition forecasters, using the so-called econometrics - the "scientific method" that connects economic theory with statistical measurements. Simply put, Makridakis people were made to predict in real life, and then evaluated the accuracy of their predictions. With Michelle Hibon he spent a few "M-events"; the third and last of them, M-3 - was completed in 1999. Makridakis and Hibon came to a sad conclusion: "the Newest and most complex statistical methods do not have achieved more accurate results than the simple".

We humans are victims of an asymmetry in the perception of random events. We attribute our successes to their mastery, and failures - external events beyond our control. Namely, accidents. We take responsibility for the good, but not bad. This allows us to think that we are better than others - no matter what we are doing. For example, 94 percent of Swedes believe that are included in 50 percent of Swedish drivers; 84 percent of French believe their sexual abilities provide them a place in the top half of the rating of French lovers.
Another consequence of this asymmetry is that we feel to a certain extent unique from anyone else (because they are, in our view, this asymmetry does not apply).

Tetlock studied the problem of "experts" in politics and Economics. He asked experts from different fields to estimate the probability that during a given time period (about five years) there will be certain political, economic and military events. At the exit he was about twenty-seven thousand predictions of almost three hundred professionals. Economists accounted for about a quarter of the sample. The study found that the experts went far beyond the limits of their tolerances for error. It was discovered and the expert problem: between the results of the doctors and students there was no difference. Professors with extensive list of publications, coped no better than the journalists. The only pattern that is found Tetlock was the inverse dependence of the forecast from the reputation: the owners of the big names predicted worse than those who have not got.
But Tetlock set ourselves the task not so much to show what really are the experts (although he did a great job), but find out why before they did not realize that they do not do their job, in other words, how they manage to "RUB points".
<...>
You say you played another game. Let's say you failed to predict the sudden collapse of the Soviet Union (didn't see this coming no sociologist). It is easy to say that you perfectly understood the essence of political processes in the Soviet Union, but the Russian, being overly Russian, artfully concealed significant economic components. If you possess these economic data, you would have certainly failed to predict the fall of the Soviet regime. We must not blame your qualifications. You can also do if you have predicted a decisive victory of al Gore over George W. Bush. You didn't know that the economy is in such dire Straits; indeed, this fact seems to have been a mystery to all. You're not an economist, but a struggle, it turns out, was economic the ring.
You just blame it on unforeseen circumstances. Happened something extraordinary, outside of your field of science. Since it was impossible to predict, you are not to blame. It was a Black Swan, and you are not required to predict Black swans.

Bull showed me an amazing research work, has just finished his trainee and already accepted for publication; it is considered in detail by two thousand predictions made by securities analysts. The work clearly showed that these broker-analysts did not predict anything: the naive forecast of a person, simply transferring the numbers from the past period to the next would be slightly worse. But analysts have information about orders, forthcoming contracts, and planned expenditures and due to this valuable knowledge seems to predict much better than our naive predictor that does not hold any information except past data. Worse, the gap between predicted and actual values was more substantial than between different predictions and, therefore, analysts led the herd instinct. Otherwise, the predictions have defended from each other as they are separated from reality.

Several researchers studied the work and the self-esteem of analysts of securities, got amazing results, especially in relation to epistemic arrogance of these figures. Comparing them with the meteorologists predicting the weather, Tyszka Tadeusz and Piotr Zielonka concluded that analysts predict the worse, but have more faith in their own ability. In any case, they lack the aplomb not to reduce the margin of error after blatant failures.