Skip to main content
SearchLoginLogin or Signup

Disclosing the TVR Prison Algorithm

Published onFeb 25, 2024
Disclosing the TVR Prison Algorithm


In this paper we disclose an algorithm of the Spanish CJS that assesses risk and classifies the inmate within seven categories ranging from low risk to maximal risk. This, so-called TVR, is a Risk Assessment Instrument (RAI) that serves as part of the decision-making process for granting prison leaves in one of the three prison systems of Spain. Although the factors and their levels involved in the TVR formula are known, we disclose here their associated weights and the complete procedure to determine the associated TVR risk level. Although this algorithm has been in use for thirty years and forms part of the prison system policy, it has remained undisclosed for the general public. Hence, it has not been subject of rigorous impartial assessment. In our research, we managed to find it within an unpublished source and therefore, we have been able to develop a quantitative assessment of it. Our research shows the TVR algorithm leads to a category of high, very high, extreme, or maximal risk in 99.3% of cases. Furthermore, out of the 6145 possible outcomes of the TVR algorithm, only 9 outcomes lead to a category of low or very low risk, what entails that there is evidence of bias towards false positives. This research provide light on some untransparent and unfair practices which affect the inmate reintegration processes of a current European prison system. By disclosing and detailing the procedure of the TVR algorithm, we encourage further assessment on the fairness and accuracy of this instrument.

No comments here
Why not start the discussion?