Solution:
Let us solve a more generalized version of the problem: Let S be a set with 2n+1 elements, and partition S into sets A0,A1,…,An such that ∣A0∣=1 and ∣A1∣=∣A2∣=⋯=∣An∣=2. (In this problem, we have A0={0} and Ak={k,−k} for k=1,2,…,100.) Let T be a randomly chosen m-element subset of S. What is the expected number of Ak's that have a representative in T?
For k=0,1,…,n, let wk=1 if T∩Ak=∅ and 0 otherwise, so that the number of Ak's that have a representative in T is equal to ∑k=0nwk. It follows that the expected number of Ak's that have a representative in T is equal to
E[w0+w1+⋯+wn]=E[w0]+E[w1]+⋯+E[wn]=E[w0]+nE[w1]
since E[w1]=E[w2]=⋯=E[wn] by symmetry.
Now E[w0] is equal to the probability that T∩A0=∅, that is, the probability that the single element of A0 is in T, which is m/(2n+1). Similarly, E[w1] is the probability that T∩A1=∅, that is, the probability that at least one of the two elements of A1 is in T. Since there are (m2n−1) m-element subsets of S that exclude both elements of A1, and there are (m2n+1) m-element subsets of S in total, we have that
E[w1]=1−(m2n+1)(m2n−1)=1−2n(2n+1)(2n−m)(2n−m+1)
Putting this together, we find that the expected number of Ak's that have a representative in T is
2n+1m+n−2(2n+1)(2n−m+1)(2n−m)
In this particular problem, we have n=100 and m=50, so substituting these values gives our answer of 2018825.