A new method for solving a class of ballot problems (Q1066902)

The author gives a novel method, using a formula for derivatives of a determinant, to solve the k-candidate generalization of the well-known ballot problem in the following form. The final votes for the candidates are \(a_ i\), such that \(a_ i<a_{i-1}+t_ i\) where the \(t_ i\) are positive integers \((i=1,2,...,k)\); determine the probability that \(A_ i(m)<A_{i-1}(m)-t_ i\) for \(m=1,...,n-1\), where \(A_ i(m)\) denotes the number of votes gained by candidate i when exactly m votes have been cast.