A Markov chain- gambler's ruin problem in the case that the probabilities of winning/ losing a particular game depend on the amount of the current fortune with ties allowed is considered. Closed-form formulas for the solution of the MC-GR problem and the expected duration of the game in the presence of partially absorbing- partially segregating- partially reflecting boundaries are deduced, and some very simple closed forms which immediately lead to exact and explicit formulas for some special cases are given.