Counts of males displaying on breeding grounds are the primary management tool used to assess population trends in lekking grouse species. Despite the importance of male lek attendance (i.e., proportion of males on leks available for detection) influencing lek counts, patterns of within season and between season variability in attendance rates are not well understood. We used high‐frequency global positioning system (GPS) telemetry data from male greater sage‐grouse (Centrocercus urophasianus; n = 67) over five lekking seasons (2013–2017) at eight study sites in Nevada to estimate lek attendance rates. Specifically, we recorded daily locations of sage‐grouse in relation to mapped lek boundaries and used generalized additive models to assess temporal variation in attendance rates by age class (subadult vs. adult). Average timing of peak attendance occurred on 16 April but varied from March 16, 2014 to April 21 , 2016. Overall, adult males attended leks at higher rates (0.683 at peak) and earlier in the season (19 March) than subadults (0.421 at peak on April 19). Peak attendance probability was positively related to cumulative winter precipitation. Daily probabilities of lek switching differed between adults (0.019 at peak on March 3) and subadults (0.046 at peak on March 22), and lek switching was negatively related to distance to nearest lek. Our results indicate variable patterns in lek attendance through time, and that lek switching may occur at higher rates than previously thought. We demonstrate the use of generalizable daily attendance curves to date‐correct lek counts and derive estimates of male abundance, although such an approach will likely require the incorporation of information on age structure to produce robust results that are useful for population monitoring.