Quick solution is expected return from strategy = 0.Possible outcomes are +9 and -3 if game is played until one is hit.So +9 x P -3 x (1-P) =0 where P is probability of hitting 9 before -3.Solution is P=1/4.

If the basket option has best of/worse of features, the risk will probably condense onto a small number of constituents as the trade progresses.If priced using an average correlation you risk that the correlation is "wrong" when you condense into a smaller number of constituents.