LSSBB
Devilish Dictionarian
So, there are many hurdles cleared, but there are some pending. Thing is, I don't think I should be the one doing this, because I haven't made the claim. My view on that is similar to how proving works in mathematics: if I find a mistake in a theorem, a step that doesn't follow from the previous step, the theorem is not proved, and it doesn't correspond to me to correct the proof. Ozeco seems to think that "the error is small and the theorem is roughly proved". There is no such thing as a "roughly proved" theorem. It is proved or it isn't, yet he tries to throw the burden of proof on me to provide a proof showing that the magnitude of some error matters. femr2, on the other hand, has repeatedly asked me to show my proof that the theorem is false, when I haven't claimed it's false (if I did, I would indeed have the burden to present a proof). Now, is there a series of steps that can fix the broken link in the deduction chain and complete the proof? That's what we're working on.
I don't think the proper way to go about analyzing any claim involving measurements is from the point of view of a mathematical theorem. Real world measurements always carry some measure of uncertainty. There are sources of variation that will not be completely characterized in nearly any continuous measurement. The analysis should then proceed in a method similar to "roughly proved". In other words, there is a certain confidence interval, say 95%, that you operate from, and given that you can either accept or reject the null hypothesis on whether the data sufficiently differs from some set. That of course does not prevent you from searching for as many sources of variation as you can, which is as much as you can do to "fix the broken link".

n
