(This article was originally published at R – Xi'an's Og, and syndicated at StatsBlogs.)
A straightforward but probabilistic riddle this week in the Riddler, which is to find the expected order of integer i when the sequence {1,2,…,n} is partitioned at random into two sets, A and B, each of which is then sorted before both sets are merged. For instance, if {1,2,3,4} is divided in A={1,4} and B={2,3}, the order of 2 in {1,4,2,3} is 3. An R rendering of the experiment is
m=rbinom(1,n,.5)
if (m*(n-m)>0){
fist=sort(sample(1:n,m))
return(order(c(fist,sort((1:n)[-fist])))[i])
}else{
return(i)}
[\sourcecode]
It is rather easy to find that the probability that the order of i takes the value j is
if j<i (i is in A) and
if $j>i$ (i is in B), the case i=j being the addition of both cases, but the mean can be found (almost) immediately by considering that, when i is in A, its average value is (i+1)/2 and when it is in B, its average value is (n+i)/2 [by symmetry]. Hence a global mean of (2i+n+1)/4….
Filed under: Books, Kids, R, Statistics Tagged: order statistics, permutation tests, R, sorting, The Riddler
Please comment on the article here: R – Xi'an's Og
