Establishing the prognosis for patients with myelodysplastic syndromes (MDS) is a key element of their care. It helps patients understand the severity of their disease and set expectations for their future. For physicians, an accurate estimate of prognosis drives decisions about the timing and choice of therapeutic options to consider. The International Prognostic Scoring System (IPSS) has been the standard tool for MDS risk stratification since it was released in 1997. It has been used to describe patients in pivotal clinical trials and is a key element of practice guidelines. Subsequent changes to the classification scheme for MDS and an underestimation of risk in some patients from the low and intermediate-1 categories have led to the development of several newer prognostic models. The most recent is the revised IPSS (IPSS-R), which addresses several of the perceived deficiencies of its predecessor. Despite their utility, none of the available prognostic systems incorporates disease-related molecular abnormalities such as somatic mutations. These lesions are present in the nearly all cases and many have been shown to improve upon existing prognostic models. However, the interpretation of somatic mutations can be challenging and it is not yet clear how best to combine them with clinical predictors of outcome. Here I review several prognostic scoring systems developed after the IPSS and describe the emerging use of molecular markers to refine risk stratification in the MDS patient population.