Ongoing challenges to understanding how hazard exposure and disaster experiences influence perceived risk lead us to ask: Is seeing believing? We approach risk perception by attending to two components of overall risk perception: perceived probability of an event occurring and perceived consequences if an event occurs. Using a two-period longitudinal data set collected from a survey of homeowners living in a fire-prone area of Colorado, we find that study participants' initial high levels of perceived probability and consequences of a wildfire did not change substantially after extreme wildfire events in the intervening years. More specifically, perceived probability of a wildfire changed very little, whereas the perceived consequences of a wildfire went up a bit. In addition, models of risk perceptions show that the two components of overall risk perception are correlated with somewhat different factors, and experience is not found to be one of the strongest correlates with perceived risk. These results reflect the importance of distinguishing the components of overall risk and modeling them separately to facilitate additional insights into the complexities of risk perceptions, factors related to perceived risk, and change in risk perceptions over time.