6 ms·
"Number of Olympic medals" is a very poor analogous concept here, because it's not a directly measurable and variable trait, it's a complex tail-end derivative
by ianferrel 4y ago
"Number of Olympic medals" is a very poor analogous concept here, because it's not a directly measurable and variable trait, it's a complex tail-end derivative of several factors.
If you looked at "swimming speed", for example, which is the simple directly measurable thing, you would find that while Phelps is faster than other swimmers, , he is only a tiny bit faster than the next fastest swimmer, not 2.5 times faster than the next fastest swimmer.
- masklinn 4y ago> "Number of Olympic medals" is a very poor analogous concept here, because it's not a directly measurable and variable trait, it's a complex tail-end derivative of several factors. As if extreme longevity is not a complex tail-end derivative of several factors?
- grandinj 4y agoOne is rank data, the other a continuous variable. Two very different sub-fields of statistics apply to the two kinds of data.
- ianferrel 4y agoWell, so is swimming speed. I may not have the statistical terminology correct, but the difference is that "age" and "swimming speed" are directly measurable continuous aspects of reality, while "# of olympic medals is some kind of discretized derivative of others. Imagine that instead of measuring "age" we measured "number of days person was the oldest person in the world". You'd get wildly divergent results for the latter that would be more like the olympic medal count. Or if we determined how wet or dry a climate was not by measuring "annual rainfall" but, like "number of minutes per year in which more rain was falling here than other places".
- chongli 4y agoYeah, plus swimming is anomalous in and of itself due to the large number of very similar events. I’m sure we’d see sprinters like Usain Bolt win a lot more medals if there were events like “100m in sandals, 100m on grass, 100m barefoot” to go along with the usual 100m race.
- omnimus 4y agoBolt being Jamaican would indeed be strong favorite in "100m on grass" event.
- joshuahedlund 4y agoA better analogy would be Usain Bolt's 0.11 gap in the 100m world dash record[0] [0]https://en.wikipedia.org/wiki/100_metres#All-time_top_25_men https://en.wikipedia.org/wiki/100_metres#All-time_top_25_men
- saalweachter 4y agoIt depends on whether "living longer" is a continuous trait, I suppose. I could imagine there being 20 or 30 genetic factors that affect your longevity, like whether you are susceptible to lung cancer from smoking. Having a particular trait is a binary, which introduces the possibility for steps in the distribution; each slice of the population with a particular number of longevity factors has a certain distribution of actual lifespans that add together to form the observed life span, but the distribution of how many people have each distribution could decrease sharply; perhaps millions of people have 24 factors and produce centenarians, thousands have 25 factors and produce 110+, and dozens have 26 factors and produce 120+. Such a model is entirely made up, pulled from my ass, but it is also quite compatible with our discrete genetics, and would be perfectly compatible with large gaps at the end of the distribution.
- toast0 4y agoAge isn't directly measurable either. Or we would just measure the age of the person and have our answer. Measuring age depends on records and tying records to people. Both of which are easily failable proceses.
- ianferrel 4y agoI take your point that many measurements of age are faulty, but age is directly measurable the same way that "time to complete a swimming race" is directly measurable. It just takes a lot more patience to measure the former. You can't test a swimmer after a race to see how fast they are either, but it's still a directly measurable quality.