Comparisons of measured total radiated x-ray power from annular wire-array z-pinches with a variety of models as a function of wire number, array mass, and load radius are reviewed. The data, which are comprehensive, have provided important insights into the features of wire-array dynamics that are critical for high x-ray power generation. Collectively, the comparisons of the data with the model calculations suggest that a number of underlying dynamical mechanisms involving cylindrical asymmetries and plasma instabilities contribute to the measured characteristics. For example, under the general assumption that the measured risetime of the total-radiated-power pulse is related to the thickness of the plasma shell formed on axis, the Heuristic Model [IEEE Trans. Plasma Sci. 26, 1275 (1998)] agrees with the measured risetime under a number of specific assumptions about the way the breakdown of the wires, the wire-plasma expansion, and the Rayleigh–Taylor instability in the r–z plane, develop. Likewise, in the high wire-number regime (where the wires are calculated to form a plasma shell prior to significant radial motion of the shell) the comparisons show that the variation in the power of the radiation generated as a function of load mass and array radius can be simulated by the two-dimensional Eulerian-radiation- magnetohydrodynamics code (E-RMHC) [Phys. Plasmas 3, 368 (1996)], using a single random-density perturbation that seeds the Rayleigh–Taylor instability in the r–z plane. For a given pulse-power generator, the comparisons suggest that (1) the smallest interwire gaps compatible with practical load construction and (2) the minimum implosion time consistent with the optimum required energy coupling of the generator to the load should produce the highest total-radiated-power levels.