Science only recently began predicting parolee behavior after release

Categorization

corroborated: 3 independent passages support this claim
XREF: connects to: Statistical prediction of parole success outperforms subjective judicial decisions; ProPublica tested COMPAS recidivism predictions against actual reoffense outcomes for bias; Parole outcomes vary predictably by measurable background factors — Brian Christian, The Alignment Problem, 244

Parole outcomes vary predictably by measurable background factors (similarity 0.87)
Statistical prediction of parole success outperforms subjective judicial decisions (similarity 0.83)
Parole prediction is possible only as aggregate statistics, not certainty (similarity 0.82)
ProPublica tested COMPAS recidivism predictions against actual reoffense outcomes for bias (similarity 0.81)
ProPublica used 2014 recidivism outcomes to test the risk model’s accuracy and bias (similarity 0.79)