Added good box conjecture for combinatorial wall-crossing

Let two Young diagrams be related by adding one box to the first row, with the corresponding parameter in the interval $(0,1/n)$. Added good box conjecture. At every step of the above composition of wall-crossings, the two resulting diagrams differ by adding one good box to some row. The claim was checked using a computer program for Young diagrams of size at least $40$, but the source does not report a proof or disproof.