Single-channel current-voltage (IV) curves of human large-conductance, voltage- and Ca(2+)-activated K(+) (BK) channels are quite linear in 150 mM KCl. In the presence of Ca(2+) and/or Mg(2+), they show a negative slope conductance at high positive potentials. This is generally explained by a Ca(2+)/Mg(2+) block as by Geng et al. (2013. J. Gen. Physiol. http://dx.doi.org/10.1085/jgp.201210955) in this issue. Here, we basically support this finding but add a refinement: the analysis of the open-channel noise by means of β distributions reveals what would be found if measurements were done with an amplifier of sufficient temporal resolution (10 MHz), namely that the block by 2.5 mM Ca(2+) and 2.5 mM Mg(2+) per se would only cause a saturating curve up to +160 mV. Further bending down requires the involvement of a second process related to flickering in the microsecond range. This flickering is hardly affected by the presence or absence of Ca(2+)/Mg(2+). In contrast to the experiments reported here, previous experiments in BK channels (Schroeder and Hansen. 2007. J. Gen. Physiol. http://dx.doi.org/10.1085/jgp.200709802) showed saturating IV curves already in the absence of Ca(2+)/Mg(2+). The reason for this discrepancy could not be identified so far. However, the flickering component was very similar in the old and new experiments, regardless of the occurrence of noncanonical IV curves.