An assessment of the severity of illness of a patient is one of the most important components of their early management. It guides decisions regarding the most appropriate site of care and the selection of empiric antibiotic therapy. In recent years, prediction tools, known as severity scores, have been promoted to assist early assessments of the severity of illness for patients with community-acquired pneumonia. Several different severity scores now exist and these have been modified over time. Each tool has particular strengths and weaknesses. This article reviews the evolution of severity scores for patients with community-acquired pneumonia and compares their performance in different patient cohorts for different outcomes of interest, as described in the published literature to date. It also discusses how these tools could be evaluated more comprehensively so that their place in patient management can be better appreciated.