I'm well aware how to analyze a dynamically growing array (each insert "pays" for two moves in the future), but I'm not sure what you're getting at. The amortized complexity of appending an element to the end is θ(1), worst-case non-amortized θ(n) when the growing needs to happen.