Fluorescent proteins are used extensively for biological imaging applications; photoactivatable and photoconvertible fluorescent proteins (PAFPs) are used widely in superresolution localization microscopy methods such as fluorescence photoactivation localization microscopy and photoactivated localization microscopy. However, their optimal use depends on knowledge of not only their bulk fluorescence properties, but also their photophysical properties at the single molecule level. We have used fluorescence correlation spectroscopy and cross-correlation spectroscopy to quantify the diffusion, photobleaching, fluorescence intermittency, and photoconversion dynamics of Dendra2, a well-known PAFP used in localization microscopy. Numerous dark states of Dendra2 are observed both in inactive (green fluorescent) and active (orange fluorescent) forms; the interconversion rates are both light-and pH-dependent, as observed for other PAFPs. The dark states limit the detected count rate per molecule, which is a crucial parameter for localization microscopy. We then developed, to our knowledge, a new mathematical estimate for the resolution in localization microscopy as a function of the measured photophysical parameters of the probe such as photobleaching quantum yield, count rate per molecule, and intensity of saturation. The model was used to predict the dependence of resolution on acquisition parameters such as illumination intensity and time per frame, demonstrating an optimal set of acquisition parameters for a given probe for a variety of measures of resolution. The best possible resolution was then compared for Dendra2 and other widely used probes, including Alexa dyes and quantum dots. This work establishes a framework for determination of the best possible resolution using a localization microscope to image a particular fluorophore, and suggests that development of probes for use in superresolution localization microscopy must consider the count rate per molecule, the saturation intensity, the photobleaching yield, and, crucially, management of bright/dark state transitions, to optimize image resolution.