This article suggests that how to include the information generated by risk scales is not straightforward. Assertions about the validity of actuarial methods hinges on numerous studies showing that rating scales clearly do better than chance at sorting individuals into categories of relatively lower and higher risk for re-offending. For practicing clinicians, the article gives reason to hesitate before interpreting an evaluee’s actuarial score as a numerical probability or recidivism. Even though it is reasonable to believe that the ranking abilities of a proven risk assessment scale will generalize to new populations in which the scale has not been tested, one should not translate a scale’s scores into probabilities of recidivism without knowing how the scale functions in those new populations and those populations’ base rates of recidivism. Over the past decade, investigators working in a broad variety of contexts have established that simple instruments for evaluating levels of risk, often called “actuarial tools” or “actuarial measures” can accurately rank sex offenders’ long-term risk of recidivism, such as the Rapid Risk Assessment for Sex Offender Recidivism (RRASOR) and the STATIC-99. In 2004, Doren suggested that evaluators might use the recidivism percentages published in original studies of the RRASOR and the STATIC-99 without regard to differences in populations or base rates. This article describes how simply comparing percentages across studies can mislead detection properties of risk assessment instruments alone, independent of the population or setting-specific base rate. Tables, figures, and references