Over the years defining gradeGrade–recoveryRecovery curves numerically has often proved challenging with many a hand drawn curve being interpolated to generate gradeGrade and recoveryRecovery estimates for forecasting or budgeting. GradesGrade and recoveriesRecovery are the most basic method of measuring separation performancePerformance and are routinely calculated and reported. But the challenge becomes comparing gradesGrade and recoveriesRecovery under varying processingProcessing conditions to assess which conditions produce the best results. Often an improved recoveryRecovery comes at the expense of a loss of concentrateConcentrate gradeGrade, or vice versa. Sometimes both gradeGrade and recoveryRecovery are affected by changes in feed gradeGrade, and this takes no account of finding data for comparison where other influential factors are relatively constant and so can be concluded not to have affected the results. Obviously, the dilemma is deciding whether a genuine improvement in the gradeGrade–recoveryRecovery relationship has occurred, or has yield been traded for quality (or vice versa). A simple separation modelModel has been used to review plant operationOperation data from the Batu Hijau copperCopper/goldGold processingProcessing plant. The modelModel enables gradeGrade–recoveryRecovery relationships to be expressed numerically making performancePerformance at different operating conditionsOperating conditions easy to compare, and in many instances, also allows the differences in performancePerformance to be quantified. Over 20 years of plant data has been reviewed using the separation process modelModel. Comparison of gradeGrade–recoveryRecovery relationships has improved understanding of the impact upon plant performancePerformance when treating ore tarnished to varying degrees. The impact of varying feed gradeGrade upon the gradeGrade–recoveryRecovery relationship has also been quantified.