For magnetic energy loss measurements on electric steel sheets, the Epstein tester (ET) is applied for 100 years. However, it suffers from unsolved problems, such as inhomogeneity of flux in the corners and even more of impact of sample cutting. As a matter of fact, a detailed study of flux distributions—as a pre-condition for improvements—has not been performed so far. In this work, we present the results of an advanced numerical 3D-magnetic anisotropic circuit calculation modeling. It considers most relevant impacts such as anisotropy, non-linearity, method of staggering, and in particular, consequences of sample cutting. According to the literature, the latter may yield material deterioration for edge zones of high width WDET. Here, we set WDET to 2.5 mm as a “coarse-cut” convention for practice, and we compare it with perfect “fine-cut” (W = 0). As a main result, deteriorated edge zones of standard ET strips of low width W = 30 mm yield a distinct concentration of flux along inner axial strip regions as a reason for strong local flux inhomogeneity and distortion. Consistent numerical comparisons predict that higher grades of magnetic homogeneity can be expected from wider sample strips, e.g., from a “Triple ET” with W = 90 mm. As a global conclusion, the Epstein frame represents a system of extreme 3D inhomogeneity, especially in the corners. The grade of inhomogeneity reaches its maximum in instants of high permeability. It may be strongly increased through coarse cutting. On the other hand, consequences of cutting can be decreased by increasing the strip width. This would allow for loss measurements using tangential field coils arranged in the central limb regions of TET.