Color vision tests and multi-test protocols in current use often fail to detect small changes in red/green (RG) and yellow/blue (YB) color vision due to poor sensitivity. The tests also have low specificity. In this study, we examine how improved understanding of within- and inter-subject variability in RG and YB color vision and accurate assessment of the differences in color thresholds between the least-sensitive, age-matched normal trichromats, and the least-affected deutans and protans can be used to design an efficient color vision screener (CVS) test. To achieve this objective, we examined two extensive data sets from earlier studies and carried out new experiments to provide better estimates of within-subject variability in color thresholds and to validate the CVS test. The data sets provide essential information on inter-subject variability, the effects of normal aging on RG and YB thresholds, and the spread in RG color thresholds in deutan and protan subjects. A statistical model was developed to optimize the parameters of the CVS test and to predict the limits of what can be achieved in color assessment. The efficiency and repeatability of the CVS test were then assessed in 84 subjects. The results match model predictions and reveal close to 100% test efficiency. The test takes between 140 and 160 seconds to complete and has close to 100% repeatability. An efficient, "two-step" protocol based on the initial use of the CVS test followed by full color assessment in only those who fail the CVS test is also described.