Book: Black Swan. Page 4

Greg Barron and IDO Erev in a series of experiments found that people tend to underestimate low probability, if they themselves have to calculate, that is, if it is not indicated in the figures. Imagine that you are pulling balls from an urn, in which very little red and a lot of black balls, and you need to guess the color of the ball you'll get in and you don't know the exact ratio of red and black. Most likely, in your estimation, the probability to pull a red ball will appear lower than it really is. But if you say, for example, that the red ball is 3 percent, you, on the contrary, will make mistakes, saying "red" more often than necessary.

Kahneman and Tversky noticed that people are able to think seriously about the seemingly unlikely events, if you engage them in discussion of such phenomena and to give the feel that these events are not so unreal. For example, if you ask a person, what is the probability of death in a plane crash, he is likely to inflate the figure.

According to the psychologist Paul Slavich and his colleagues, people imagine are more willing to pay for insurance against terrorist attacks than plane crashes (although the number of recent terrorist acts includes as a special case). <...>
At the same time, according to Slovic, the insurance people usually neglect events with low probability. This can be formulated as follows: people prefer insurance against possible small losses, while less probable but much more severe losses are not taken into account.

Those Black swans we imagine, discuss, and fear isn't really threatening us Black swans. Soon you will become clear that we do not fear what should be feared.

Which of these situations, in your opinion, more likely?
1. Joey, apparently, was happily married. He killed his wife.
2. Joey, apparently, was happily married. He killed his wife, to take possession of her inheritance.
Certainly, obeying the first impression, you say, more likely the second option, although this is pure logical fallacy, because the first statement is broader and involves more than one possible cause, but many: joy killed his wife because he is mad, or because she cheated on him with the postman and the ski instructor, or because in a state of confusion, he took her for a financial forecasting.

Psychologists Kahneman and Tversky <...> made the experiment: the subjects (and these were people whose profession is associated with the preparation of the forecasts) were asked to rate the probability of two events:
a) in America will happen a huge flood, which will die more than a thousand people;
b) an earthquake in California will cause severe flooding, which will killed more than a thousand people.
According to the respondents, the first variant is less probable than the second: the earthquake in California is the reason that easy to make and that makes the flood situation more representable, thereby increasing its probability in our eyes.
Similarly, if I ask: "what do you think, how many of your fellow citizens sick with lung cancer?" - you call any number (let's say half a million). However, if I ask, how many people in your country did develop lung cancer as a result of Smoking, the number you called, will be much (perhaps, as in two or three times) more. If the cause of the phenomenon, this phenomenon seems much more plausible and much more likely. Cancer from Smoking is easier to believe than cancer without a cause (if the cause is unclear, it's like, what it is not).

Proversion something full of facts, but powerwiki intellect, alas, no.

Journalists prefer infinite precision erroneous conclusions "approximate" truth-teller.

Experienced chess players-grandmasters, as you know, really focusing on the weakness of the potential stroke. But do not have to play chess to practice skepticism. Scientists believe that the ability to dig into their own weaknesses makes them good chess players, not chess turns them into skeptics. In the same way stock market player George Soros, before you make a bid, collects data that would refute his original theory.

If I saw a man commits murder, I can hardly doubt that he is a criminal. But if I saw a man commits murder, I can't be sure of his innocence. The same applies to cancer detection: the identification of a single malignant tumor proves that you have cancer, but finding it not allow you with absolute certainty that the cancer you have.
Us closer to the truth of the negative and not confirming the examples! Wrong to deduce a General rule from observed facts.

In 1960-e years, at the time of the highest conceit of science, doctors, confident in their ability to recreate in the laboratory the mother's milk, in a penny not put, not realizing that breast milk may contain valuable components that took refuge from their enlightened attention; the lack of evidence about the benefits of breast milk was taken as evidence of lack of use thereof (another example of Platonism: "it makes no sense" to breast-feed, if you can bottle feed). Many suffered from this naive logic: those who as infants were not breast-fed, appeared to be more susceptible to a number of diseases, including certain types of cancer - probably in the breast milk there are some essential protective nutrients that we do not know anything definite. Moreover, the beneficial effects that breastfeeding has on mothers, such as reducing the risk of breast cancer, were not taken into account.
Same story with the tonsils: a tonsillectomy increases the risk of throat cancer, but for decades doctors never suspected that this "useless" tissue can be for anything, unknown to them, is necessary. The same with dietary fiber in fruits and vegetables: doctors in the 1960-ies considered it useless, because it lacked the evidence it needed, and as a result we got the wrong fed generation.
Fiber, as it turns out, slows the absorption of sugar in the blood "and cleans out the intestinal tract of precancerous cells. General medicine in its history caused a lot of harm, and the reason for that - these simple flaws Changeling.

In 1971, the psychologists Danny Kahneman and Amos Tversky decided to torture professors of statistics questions, formulated as statistical issues. One was approximately as follows (for clarity, I changed example): imagine that you live in a city where there are two hospitals - one large, the other small. On a certain day in one of these two hospitals is born with 60% of boys. What hospital this might happen? A Professor made a mistake (during normal conversation), calling a large hospital, while the essence of statistics is that large samples are more stable and have less deviations from the long-term average (in this case 50 percent of each sex) than a small sample. These professors would have failed the exams that they themselves accept. Still working as a quantum engineer, I identified hundreds of such serious mistakes made by statisticians who forgot that they statistics.

Erudition is important to me. It reflects a sincere intellectual curiosity.

Question by will help You to remain calm.