I identify a simple pattern in evaluated isotope half-life trends as a function of neutron number that I call the 2-neutron stability shoulder. For each element, I define a stable band as the interval in mass number spanned by the element’s lightest and heaviest observationally stable isotopes. On the neutron-poor and neutron-rich sides of this band, I select a representative wing isotope by taking the longest-lived radioactive nuclide on that side, breaking ties in favor of the nuclide closest to the band edge. For each wing, I then test whether moving toward the stable band by neutrons produces a nuclide with a longer half-life, or reaches stability. Using evaluated ground-state half-lives from NUBASE2020, I benchmark for elements with . Under this definition, is the dominant step size on both wings. For the 90 elements for which both wings are defined and the required stepped-to nuclides exist in the dataset, yields zero “neither” cases: every element has at least one wing where the step increases half-life (or reaches stability), and 56/90 elements improve on both wings simultaneously. I compare against and discuss interpretations consistent with odd–even (pairing) systematics and the geometry of the valley of stability.