https://www.chessprogramming.org/Piece-Lists

I am a beginner to chess programming (and programming in general). I've already made a functioning move generator in Java that passes perft tests, but it's really slow and is bloated with unnecessary OO design so I want to completely restart. My new engine will use a hybrid approach of 0x88 array and a piece list for the board representation (I know bitboards are better but I want to try something more intuitive first, that might be my next step). The above link mentions linked lists as an alternative representation for piece lists, and I'm wondering how that could be more efficient than the array-based one that is explained in the article, especially since it does not require any shifting of array elements. As far as I know there is some overhead in traversing linked lists, compared to arrays. What exactly is the advantage of using a linked list instead of arrays for the piece-lists?