Figure 1
From: How to quantify information loss due to phase ambiguity in haplotype case-control studies

Forward stepwise selection of the informative individuals based on D-optimality in 100 cases by maximizing | I tot |. (1) The group identification denotes the genotypes at SNP: '1' and '2' represent homozygotes 1/1 and 2/2, 'H' a heterozygote 1/2. (2) The y-label is ordered by A-optimality (the highest 'HHH' group for the first selection, the 'H1H', 'HH1', etc), the red points by D-optimality. So the first individuals to be selected are 'H1H' group, not 'HHH', and hence it shows discrepancy using two different measures. The jumps between groups indicate the correlation between parameters.