Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Greater?
Posted:
Nov 25, 2013 11:14 AM


Le lundi 25 novembre 2013 16:44:48 UTC+1, Roland Franzius a écrit : > Am 25.11.2013 16:10, schrieb luttgma@gmail.com: > > > Le lundi 25 novembre 2013 14:37:17 UTC+1, Roland Franzius a ï¿½crit : > > >> Am 25.11.2013 12:29, schrieb luttgma@gmail.com: > > >> > > >>> What is the probability that 14.5 +/ 0.8 is greater than 13.8 +/ 0.04 ? > > >>
> > >> > > >> > > >> For two random production processes X, Y, normally and stochastic > > >> > > >> independently distributed with CDF NormalDistribution(m_k,s_k), the > > >> > > >> difference is normally distibuted with > > >> > > >> mean = difference of the means and variance = sum of variances: > > >> > > >> > > >> > > >> F(z)=Pr[XY < z ] > > >> > > >> = N( m1m2, sqrt(s1^2+s2^2) ) = N(0.7,sqrt(0.064+0.0016) ) > > >> > > >> So the answer may be probably > > >> > > >> > > >> CDF[NormalDistribution[0.7, Sqrt[0.8^2 + 0.04^2]]][0] > 0.191084 > > > > >> > > >> Roland Franzius > > > > > > 14.5 +/ 0.8 billion years is the estimated age of the socalled "Methuselah star," cataloged as HD 140283. > > > 13.8 +/ 0.04 billion years is the estimated age of the Universe. > > > > > > According to your aanalysis, the probability that HD 140283 is older than the Universe is rather small. > > > > > > No the other way around, swap the variables. > > > > But, by the way, the statistical independency of birth of parents and > > childs is currently not yet widespreadly assumed and the error marges in > > astrophysics don't describe distributions of times but statistical > > errors in measurement procedures eg by repeated mesurements of the same > > variable. > > > > In the preparatoy laboratory courses the question is discussed thorouhly > > if one can estimate the the distance between the main train station and > > BunsenSt. only by counting steps and taking the mean of 10000 walks and > > their variance to get a precision of 3 inches. > > > > > > So, even if by radio carbon methods the age of bones of parents and > > childs in an egyptian tomb dates the children before the parents with no > > overlap, its not worth taking notice. > > > > > > > >  > > > > Roland Franzius
Thank you again!
If I understand correctly, it cannot statistically be assumed that HD 140283 is older than the Universe, in spite of a probability of about 0.8
Marcel



