4 ms·

I don't think that answer helps. How are you assigning "too much" weight on outliers, and the process behind deciding the right amount of weight? Can you think

• azakai 10y ago
  Obviously "too much" depends on the subject matter. But the point is that, 1. STD gives a lot more weight to outliers than MAD. 2. People constantly hear STD and think it means MAD, for all the reasons the article mentions. The argument isn't that STD always gives too much. It is that it gives a lot more than people expect. The extreme example given in the article is that a statistical process can have infinite STD, but finite MAD. In other cases, say income, the STD might be double the MAD. That's bad if you think STD means MAD. Anyhow, this could be solved by educating people on what STD actually means, or by just using MAD. The article apparently thinks the latter is more practical, especially since the benefits of STD have decreased over time.
• reso 10y ago
  In this case what Taleb is concerned with is decision making. The right amount of weight is what allows human beings to make good decisions. He believes that MAD is much more intuitive to humans and therefore leads to better decisions.
  • ori_b 10y ago
    In that case, I'm looking for a concrete example.