Plant density is one of several factors that can affect crop yield. In conventional variety trials, it is generally assumed that the plant density is optimal and similar across all varieties, but in practice, field emergence and plant density may vary among varieties. In this study, we want to investigate an alternative experimental design that may help to reduce the part of the genotype x environment interactions caused by heterogeneity in plant density among varieties. We investigate the use of a non-replicated quantitative design adapted towards regression of yield on plant density by intentionally inducing variation in plant density on the cost of repetitions, and compare it to a conventional ANOVA-focused block design. Based on two field experiments and a simulation study, we show that the estimated yield accuracy is comparable between the two designs under the condition that plant density is similar across varieties. If plant density varies, the quantitative design provides the opportunity for crop yield adjustment. Furthermore, the quantitative design allows estimation of the yield potential, the optimal plant density, and the sensitivity to low plant densities for each variety. Across experiments, varieties, and years, the yield potential was 2.18 t/ha - 7.82 t/ha, the economically optimal net yield was 2.75 t/ha - 6.62 t/ha, and the economically optimal plant density was 39.8–67.7 plants m⁻². The study demonstrates that the quantitative design provides added value compared to ordinary qualitative designs, by offering additional information such as the yield potential, optimal plant density,flowering and sensitivity measures.