4 ms·

You Are the Best at Something

1 points by alexmolas 3y ago

coldtea 3y ago
>A person is defined by N different traits. From my experience it’s fair to assume that is very big, let’s set for N=500 the moment. We can assign to each skill a value between 0 and 1, and for the sake of simplicity assume also that the distribution of these values is uniform (probably it’s better to assume that’s Gaussian or a fat-tailed distribution, but let’s keep it simple). Skills values are independent. When you’re born you get assigned a value at random for each different trait. We say that you are very good at some skill if you have a value bigger than the 90% of the population for that trait. With these assumptions, the question we ask is what is the probability of being the best at something? Most people have nowhere near 100% probability of being the "best in the world at something". They might have the potential, but so do countless others not being it. Let's see the assumptions to get to that "100%": "Since each trait is independent we have a probability of p = 1 - 0.99 = 0.01 of being very good at that dimension. Since we have N dimensions the probability of being very good in at least one of the features is given by 1 - (1-p)^N = 0.9933, which means that you’re probably one of the best in the world on at least one thing with a probability of almost 1, which is pretty cool". First, traits are anything than independent. If your IQ is low, most of the intellectual traits for example lose the chance to be good, much less "better than the general population". If you're overweight or lack exercize you're not going to do well on any athletic or physically demanding dimensions (well, except Sumo). If you lack an education, you ain't going to be the best doctor, physicist, lawyer, scholar, and so on. If you have low EQ, the related trait scopes also plummet altogether. Second, the probability to being good in any dimension is related to genetics, upbringing, opportunity (not many astronauts born in Lagos), and training, none of which is remotely close to random. In fact, given the assumptions (independece of traits, random probability of 0 to 1 score in any of them), and how percentiles work, you don't even have to have anything to do with the trait in actual life to be in the "top 99 percentile". E.g. given a world population of ~ 8 billion, about 80 million people would be in the 99% percentile of being good at a trait, e.g. at "being a doctor", even though there are only about 14 million actual doctors in the world. Which is mathematically sound, but meaningless, an artifact of assuming the trait skill score "random".

1 points by alexmolas 3y ago

>A person is defined by N different traits. From my experience it’s fair to assume that is very big, let’s set for N=500 the moment. We can assign to each skill a value between 0 and 1, and for the sake of simplicity assume also that the distribution of these values is uniform (probably it’s better to assume that’s Gaussian or a fat-tailed distribution, but let’s keep it simple). Skills values are independent. When you’re born you get assigned a value at random for each different trait. We say that you are very good at some skill if you have a value bigger than the 90% of the population for that trait. With these assumptions, the question we ask is what is the probability of being the best at something? Most people have nowhere near 100% probability of being the "best in the world at something". They might have the potential, but so do countless others not being it. Let's see the assumptions to get to that "100%": "Since each trait is independent we have a probability of p = 1 - 0.99 = 0.01 of being very good at that dimension. Since we have N dimensions the probability of being very good in at least one of the features is given by 1 - (1-p)^N = 0.9933, which means that you’re probably one of the best in the world on at least one thing with a probability of almost 1, which is pretty cool". First, traits are anything than independent. If your IQ is low, most of the intellectual traits for example lose the chance to be good, much less "better than the general population". If you're overweight or lack exercize you're not going to do well on any athletic or physically demanding dimensions (well, except Sumo). If you lack an education, you ain't going to be the best doctor, physicist, lawyer, scholar, and so on. If you have low EQ, the related trait scopes also plummet altogether. Second, the probability to being good in any dimension is related to genetics, upbringing, opportunity (not many astronauts born in Lagos), and training, none of which is remotely close to random. In fact, given the assumptions (independece of traits, random probability of 0 to 1 score in any of them), and how percentiles work, you don't even have to have anything to do with the trait in actual life to be in the "top 99 percentile". E.g. given a world population of ~ 8 billion, about 80 million people would be in the 99% percentile of being good at a trait, e.g. at "being a doctor", even though there are only about 14 million actual doctors in the world. Which is mathematically sound, but meaningless, an artifact of assuming the trait skill score "random".