Probabilistic risk assessment (PRA) has been used in various technological fields to assist regulatory agencies, managerial decision makers, and systems designers in assessing and mitigating the risks inherent in these complex arrangements. Has PRA delivered on its promise? How do we gage PRA performance? Are our expectations about value of PRA realistic? Are there disparities between what we get and what we think we are getting form PRA and its various derivatives? Do current PRAs reflect the knowledge gained from actual events? How do we address potential gaps? These are some of the questions that have been raised over the years since the inception of the field more than forty years ago. This paper offers a brief assessment of PRA as a technical discipline in theory and practice, its key strengths and weaknesses, and suggestions on ways to address real and perceived shortcomings.