The empirical basis of sentencing guidelines is discussed, and some of the data used to construct the guidelines are analyzed. The extent of judicial compliance with sentencing guidelines is examined, and the effect of guidelines on the overall severity of sentences is considered. Further, the role of sentencing guidelines within the larger system of discretionary decisionmaking, including the prosecution and the defense, is explored. Additionally, the hypothesis that when judicial sentencing discretion is constrained so that sentences become more predictable or predetermined, control over sentences shifts to the attorneys via charge bargaining is tested. Finally, the ramifications of plea bargaining on the structuring of judicial sentencing discretion are examined. The empirical approach to the formulation of sentencing guidelines is found to be fundamentally flawed. The guidelines are shown to have had no measurable impact on judicial sentencing behavior. Indeed, compliance rates were lower than had been expected, and sentencing disparity was not reduced. In addition, the guidelines did not affect plea bargaining practices. On the contrary, the failure to account for the existing plea bargaining practices is shown to be part of the reason for the guidelines' inefficacy. A new approach to the problem of structuring judicial sentencing discretion is called for. Included are 148 tables, 21 figures, and 5 appendixes.