4 ms·

Looking at <100 years of data, fitting a logistic equation to it and concluding that the limit is close to the current plateau is ... unsatisfying. Maybe that'

Looking at <100 years of data, fitting a logistic equation to it and concluding that the limit is close to the current plateau is ... unsatisfying. Maybe that's an unfair summary, there is more statistic rigor than that in the paper. But nonetheless the observed time period seems way too small to support any conclusion

Aachen 2y ago
I was similarly wondering whether the author has tried taking a subset of the data and seeing if the result still comes out the same. It seems mildly suspect that, in each case, it just so happens that we've just about reached the limit. By excluding recent data, there can't be a bias to recent values
hansvm 2y ago
It depends a bit on your model of the observed performance improvements. Do we have better training and nutrition (a logistic model seems simplistic if so)? Do we simply perform better when we have other high-performing peers to compete against (in which case the logistic model isn't perfect, but it's not terrible -- you'd expect an exponential increase in population to have enough additional brackets to get every constant, additive gain, and assuming logistic growth for the population a logistic curve-fitting roughly inverts that function)? At best though, it's a model of what _will_ be achieved, as opposed to the hypothetical bounds on those achievements if we were going to invest resources into hitting them.
PaulDavisThe1st 2y ago
What it incredibly frustrating to me that is on the one hand, we might somewhat glibly say that we know we can ignore the data for the preceding 5000 years because we know that all the fastest performances have been in the last century. But merely saying that then leads to the question: do we actually know that? And what's frustrating is that I don't there is any way to be confident about any answer to that question.
- scott_w 2y ago
  I think we can. We have archeological records showing us what even the fittest people looked like and what clothes they wore. Modern equipment is so much better than what ancient people had access to, it’s just not even close. Better nutrition means people can get bigger/stronger more easily. Easier access to training equipment amplifies this. The access to information we have means even more people can train more optimally than people in the past possibly could. The sheer volume of people gives us statistical probability that genetic freaks will appear more frequently. All of these gives us more confidence that current records are “true” records.
AlbertCory 2y ago
Right. A data-based approach is unsatisfying. Here is a related question: How fast can a baseball pitcher throw a ball? Will we ever see a 125 mph fastball? There's no equipment involved (running shoes, etc.) I saw an physics-based analysis of this once, based on the tendons in the arm and shoulder. Your muscles might get stronger and your mechanics might be optimized, but the tendons do not improve. At what speed will they just snap? This video is only a little bit about that, when the physiologist mentions the ligaments. It's hugely entertaining, though: https://www.youtube.com/watch?v=8udNOTFiqUs https://www.youtube.com/watch?v=8udNOTFiqUs

Aachen 2y ago

I was similarly wondering whether the author has tried taking a subset of the data and seeing if the result still comes out the same. It seems mildly suspect that, in each case, it just so happens that we've just about reached the limit. By excluding recent data, there can't be a bias to recent values

hansvm 2y ago

It depends a bit on your model of the observed performance improvements. Do we have better training and nutrition (a logistic model seems simplistic if so)? Do we simply perform better when we have other high-performing peers to compete against (in which case the logistic model isn't perfect, but it's not terrible -- you'd expect an exponential increase in population to have enough additional brackets to get every constant, additive gain, and assuming logistic growth for the population a logistic curve-fitting roughly inverts that function)? At best though, it's a model of what _will_ be achieved, as opposed to the hypothetical bounds on those achievements if we were going to invest resources into hitting them.

PaulDavisThe1st 2y ago

What it incredibly frustrating to me that is on the one hand, we might somewhat glibly say that we know we can ignore the data for the preceding 5000 years because we know that all the fastest performances have been in the last century. But merely saying that then leads to the question: do we actually know that? And what's frustrating is that I don't there is any way to be confident about any answer to that question.

scott_w 2y ago
I think we can. We have archeological records showing us what even the fittest people looked like and what clothes they wore. Modern equipment is so much better than what ancient people had access to, it’s just not even close. Better nutrition means people can get bigger/stronger more easily. Easier access to training equipment amplifies this. The access to information we have means even more people can train more optimally than people in the past possibly could. The sheer volume of people gives us statistical probability that genetic freaks will appear more frequently. All of these gives us more confidence that current records are “true” records.

AlbertCory 2y ago

Right. A data-based approach is unsatisfying. Here is a related question: How fast can a baseball pitcher throw a ball? Will we ever see a 125 mph fastball? There's no equipment involved (running shoes, etc.) I saw an physics-based analysis of this once, based on the tendons in the arm and shoulder. Your muscles might get stronger and your mechanics might be optimized, but the tendons do not improve. At what speed will they just snap? This video is only a little bit about that, when the physiologist mentions the ligaments. It's hugely entertaining, though: https://www.youtube.com/watch?v=8udNOTFiqUs https://www.youtube.com/watch?v=8udNOTFiqUs