Split time vs metric accuracy. The x-axis is a split time parameter added to the Human demographic model indicating the point when two populations start diverging. The y-axis has two labels, the first, Ancestry Accuracy, indicates how accurate the model parameters correctly cluster the two populations, where 50% accuracy is a random assignment. The second y-axis label indicates the % accuracy of AdmixKJump or cross-validation to correctly identify K
∗=2 or two clusters. I am reporting population sample sizes of 10 (blue), 30 (red), and 50 (purple).