Optimizing Community Supervision Practices with the Elo-rating System: A Proof-of-Concept

Abstract

Abstract Record numbers of offenders are being released to community supervision. This poses a challenge to agencies and officers in charge of providing supervision because there are far more clients in need of supervision than officers can reasonably attend to. This challenge represents an opportunity for criminologists to work with agencies to find innovative ways to allocate their limited resources. Standardized risk assessments are the most commonly used tool for determining which clients receive the most resources (e.g., officer attention). But risk assessment is a static instrument that is incapable of making cross-client predictions. We offer a novel, dynamic approach by integrating the Elo-rating system into the risk assessment paradigm. Our study is meant to present a potential innovative solution and to demonstrate that it may have merit when applied in practice. We test our idea against data from the Pathways to Desistance Study and show that the Elo-rating system is a flexible tool that can be used to adjust to changes in risk more rapidly than standard risk assessment tools. We also show that the Elo-rating system, when combined with risk assessment, can increase the accuracy of prospective predictions of recidivism risk. We discuss how Elo-rating system could be used by community supervision agencies seeking to optimize decision-making.