Cost, time, linkage drag, and genetic drift work against the incorporation of potentially valuable alleles from exotic or non-adapted germplasm into elite crop plants, particularly for quantitative traits. We present a model, motivated by narrow-leafed lupin (Lupinus angustifolius), for efficient incorporation of new alleles from exotic or non-adapted donors into elite gene pools during two phases of breeding. In Phase 1, probability functions from the binomial distribution provide at least 95% confidence that a potentially valuable donor allele (A′) will survive two cycles of backcrossing to elite lines and is fixed in BC₂-derived lines. During backcrossing, up to 6 major domestication or adaptation genes from the elite parents are reselected and made homozygous in BC₂S_0 : 1 family rows. Each plant in the BC₂S_0 : 1 contains on average 12.5% donor alleles, with >95% probability that a particular donor allele is homozygous in at least one fully domesticated plant in the BC₂S_0 : 1 population. Plants in these rows or subsequent field trials are selected for valuable quantitative traits, and crossed into elite germplasm to commence Phase 2. Phase 1 is rapid and relatively low cost, and provides a continuous flow of novel genetic diversity into the elite breeding pool.