Following the schedule of our reading calendar
This is another devastating chapter on “experts”. Kahneman is not debunking the whole idea of expertise, there can be experts in making violins or bridges, let’s say, but mostly on the idea of experts in what he calls “low-validity environments” that is environments in which there is a relevant degree of uncertainty, so predictions rapidly turn pointless: economics, policy, law, and almost anything relevant to daily human lives are low-validity environments.
More specifically, we are introduced to a shocking idea: very simple algorithms are a lot better to predict future outcomes than experts. Simple combinations of two or three variables and a simple mathematical formula makes better predictions than any expert, that tries to be clever and impressive and, therefore, makes a lot of mistakes. Even, when the experts are given the formula, they make worse predictions than the formula. The reason, of course, is that experts think that they really know a lot more about the subject than the common man or woman, so they try to outsmart the formula.
We are not talking here to elegant algorithms based on elaborated regression formulae, but very very simple linear models (a*x +b). That sounds preposterous to social scientists, and consider that such formulae are inhuman, unreal, artificial, while expert text pert knowledge is holistic, rich, subtle and so on
Expert text pert choking smokers,
Don’t you think the joker laughs at you?
Another scientific result to make your day miserable: Humans are inconsistent when evaluating information. If you give an expert the same information twice in different days he’ll probably give you a different answer each day
The chapter ends with a short story in which Kahneman described how he developed a simple algorithm based in just six dimensions, scored from 1 to 5 in order to decide recruits for the Israeli army. We are invited to do a similar method in our daily lives, when having to take a complex decision, like hiring someone for a job or not, to consider not more than six dimensions relevant for the problem and score them on a scale of 1 to 5 (1=very weak, 5= very strong).
I’d say lately in the book the connection between the experimental results and differences between system 1 and system 2 are not that relevant. When Kahneman talks about them, it is a sort of general background, and not anymore as the mechanism to explain why we make mistakes. I have the impression that this section is no longer about a rational system that works well (system 2) that is hijacked by a more intuitive system (system 1), but a general claim of irrationality of humans.