Spread of a multinomial count gap
Roll a weighted die times independently. Let count how many rolls show face (probability ) and count face (probability ). What is , the variance of the gap between the two counts?
Show hints (2)+
- Var(N_i - N_j) = Var(N_i) + Var(N_j) - 2 Cov(N_i, N_j), with the multinomial covariance Cov = -n p_i p_j.
- The covariance is negative, so -2 Cov adds a positive term: 1.6 + 2.1 + 1.2.
Answer
Reveal answer →Final answer
4.9 (± 0.01)
Want the full step-by-step worked solution? It's part of Premium — along with a worked solution for every question in the bank.
Asked at: Jane Street, Two Sigma