It is generally acknowledged that a reduced ΔK, namely ΔKeff, is the best correlation between cyclic mechanical loading conditions and fatigue crack propagation rates. Yet, there are differing opinions among scientists as to the cause for such a reduction, and different concepts are proposed for the transfer of ΔK into ΔKeff. In this work, it is demonstrated that the highly constrained compressive deformation around the crack front during the unloading part of a fatigue cycle leads to a high hydrostatic stress field that needs to be overcome during the loading part of the following cycle before a crack can further extend. The stress intensity factor, K, describes the loading part of the fatigue cycle, while the unloading part of a cycle follows completely different rules with respect to stresses and strains in the plastic zone. A rigorous separation of these two parts of a fatigue cycle is conducted, leading to a new view at the problem. The crack growth increment of a cycle depends on Kmax and the crack propagation stress intensity factor, KPR, set by the previous cycle. KPR itself depends on Kmax and the unloading level, Kmin (or — σmin) and thus solely on the unloading part of the previous cycle. The crack propagation stress intensity factor, KPR, represents the load level where the hydrostatic stresses in the vicinity of the crack tip are released, thus dividing the loading part of a cycle into an effective and a noneffective part.