^ Ok, I don't want to hijack the beast thread anymore so you can PM me or post to my journal if you want to reply but I'll give one last point in this thread.
There are two points and I think you have to separate them to understand.
1) The first point is that you can't PROVE the someones training or ability to maintain weight is necessarily sub-optimal; it might be unorthodox but optimal for them. I give you this point but arguing about being able to prove something surely true is ridiculous and not in the domain of science. Like I said before, you have no problem asserting that a lesser athletes training methods are sub-optimal when you know you have strong evidence that their methods wouldn't help most humans. Now can you prove that someone who has used long distance marathoning to increase their vertical from 20'' to 24'' hasn't maximized their potential? No. You can't prove it. Maybe they are the training outlier and while unorthodox they have fully maximized their potential this way. But... You can be pretty sure what they are doing is suboptimal.
2) Your second point is that if someone is extremely good at something then it's much much more likely that their methods have been fruitful. For this reason the Maurice Clarrett example was actually quite compelling because in his case we can see the whole story in front of us. He was at one point the top running back in college football as a freshman. By any measure you would admit he was at the top of his craft. However, he had what looked to most like a detrimental training regiment. If we went back in time when he was at his best and told you of his training regiment would you say "Well, seems weird, but I can't say he isn't maximizing his potential because he's the best at what he does so it must be unorthodox?". I would say "He is a freak but I don't THINK he is maximizing his potential with what he is doing" (Again from pt 1, I don't know but I strongly think). Turns out in the future his poor approach to his craft cost him and I think now it's safe to say he didn't maximize his potential...
Maybe this isn't satisfying to you because Maurice Clarett didn't at any time maximize human potential like Usain Bolt. I understand that your intuition is since Bolt is the fastest human of all time it's highly unlikely that he isn't maximizing his potential. My main point is that this intuition is wrong. And we have the data to prove it. I used to work with a group that did modeling of earthquakes; the distribution of the outliers is what causes this to be false. While you might not be a statistician I am telling you that you arrived at the conclusion implicitly assuming a normal distribution among outliers as well as humans. Essentially the thought that one would use to assume Usain Bolt must be maximizing his potential is as follows:
A) Humans have a wide distribution natural sprinting ability [ For the hundred lets assume most males people are somewhere 10-20 second range in the absence of training ]
B) Humans can improve on that natural sprinting ability with training
C) Therefore, if you select a group of fast humans at random it's likely that they are fast because of a combination of A and B. Some may have more natural ability and train less optimally, some less ability but train better, but on average the group is made up of humans with better than average ability AND training.
D) Usain Bolt is that fastest human that ever lived. Therefore in his case it's extremely unlikely that he doesn't have a whole lot of A and a whole lot of B. It's unlikely that would be born with enough "A" to not need "B" to become the best.
*** I agree all the way up to part D. The statistics for why I don't believe in D are explained a bit here:
http://onlinelibrary.wiley.com/doi/10.1029/2008GL035967/abstract but it's REALLY dry and I think I can give an intuitive understanding of why we should not expect that the performance of Usain Bolt is proof that he is maximizing his potential and in fact take his otherwordly performance as evidence that's it's only as likely as it is for any amateur. Basically the proof is in the world records. Just think of when Usain Bolt ran his world record 100m. Assume the mean 100m time for an average trained male is about 13 seconds. Then a much smaller number of times would be run at 12 seconds. And a much smaller number at 11 seconds and so forth. And a much smaller number at 10 seconds. In fact with enough data each 1/100 of a second would go down in the graph. The number of men to achieve 9.98 would be greater than 9.97 which is greater than 9.96 and so forth. This is called monotonic decreasing and is a the hallmark of a normal distribution. However, then we notice that Bolt runs 9.58. Nobody runs 9.59, 9.60, 9.61, 9.62, 9.63.... All the way up to 9.69 which Bolt had run himself. At the time Tyson Gay's 9.71 was the next fastest. It doesn't make sense. 9.58 is MUCH MUCH more unlikely than 9.65. So one observation of 9.58 suggests multiple at 9.65. But we don't see that. This is how outliers work. The tallest man capable of standing up ever was 8' 11''. The second tallest man verified height is 8' 3''. Thats a 8 inch difference. One would assume that under normal conditions since reaching 8'11'' is so unlikely that everything must have lined up perfect to reach such a height. In fact the guy was still growing when he died of an infection unrelated to his height and thus clearly did not reach his potential....
That might be a bit rambling, but I'm just trying to get you to think that maybe the fact that someone is an extreme outlier actually allows us to assume less about the conditions that lead to the event. You have to think about the distribution like this: Most males are born somewhere on the normal curve. The further away from the mean the less likely you will have that level of natural ability. HOWEVER a small amount of people are outliers. Thus, they are born VERY far from the mean say 4 + standard deviations. The likelihood of this event occurring is very rare, but given that it occurs it's not much less likely that they are born 4.6 standard deviations away rather than 4.2. This is a consequence of statistics that I promise is true. It's an extreme value distribution. Given this information we can make the inference that in the event of the outlier the normal expectation on the conditions necessary for an event are not only not necessary; they are actually unexpected! Now apply this to Bolt. Imagine ability (A) and training (B) are both normally distributed. Some people train really hard and it allows them to improve at a fast rate. Consider these two cases:
1) Bolt is an outlier in "A" (natural ability) assume he is 4.2 standard deviations from the mean giving him the natural ability to run 10.3 seconds without much training. He is also highly separated from the mean (maybe not as much) but say maybe 2 standard deviations from the mean in training giving him the drive to train much harder than the average person resulting in the ability to lower his 10.3 second time to 9.58.
2) Bolt is an outlier in "A" and is 4.5 standard deviations from the mean giving him the natural ability to run 9.9 seconds without much training. He is about average as far as training goes thus lowering his 9.9 to 9.58.
******
Both of these are possible but #1 is much less likely because it includes one extremely rare event and one quite rare event. #2 assume only one rare event (which is only trivially less likely than the first even in #1 due to the extreme value distribution).
Hope that made sense. I do statistics for a living and I'm not necessarily the best at explaining it but hopefully you can consider the implications of the extreme value distribution. Also, I realize the models are contrived (as all models are) and that the real world is much more complicated (eg given that bolt is fantastically gifted elite training will be brought to him, etc), but it is worth considering that when faced with an extreme performance we should not use the correct intuition that we use for intermediate and even advanced performances. Thus, when Bolt admits he is lazy and his coach constantly admonishes him for his laziness and he comes into every season out of shape and running slow and admits he just doesn't try that hard cause he prefers playstation.... Maybe we should consider that maybe he could be even faster and not that it's impossible to be so great without everything lining up.... rather you just need one thing to line up in an extreme way.
