This paper examines how sound the evidence of risk of recidivism made with structured risk assessment tools can be, when the law requires to take into account a recidivism risk assessment in making decisions that restrict the rights of the accused or convicted person. We use the well-known statistical aphorism that "all models are false, but some are useful" as a guiding thread, to examine whether quantitative recidivism or violence risk assessments, which are always "false" (inaccurate) to some extent, are however good enough to be used in criminal law, a question that, in turn, depends on what objectives the criminal law is intended to achieve.... [Leer más...]