The conduct of multi-environment yield trials (METs) has fostered significant improvement in wheat (Triticum aestivum L.) grain yield and yield stability throughout the U.S. hard winter wheat region. While information from such trials has been used to identify many superior wheat genotypes, yield data from these trials have not yet been used in any systematic fashion to improve trial design or demonstrate genetic progress over time. The objective of our study was to conduct a retrospective analysis of historical wheat grain yield trials in Colorado to: estimate variance components due to genotype, environment, and genotype x environment interaction effects; determine the optimum number of replications and trial locations to enhance genetic gain; assess genetic gain in Colorado resulting from local and non-local applied wheat breeding efforts; and characterize the pattern of yield response among trial locations. Nonirrigated, replicated grain yield trial data from 1990-2006 were used in mixed model analyses to estimate variance components due to main effects of genotypes and environments and interactions between genotypes, locations, years, and years x locations. Estimated variance components were then used to model the effect of the number of replications and trial locations on genotypic repeatability. Best linear unbiased predictors (BLUPs) from a genotype x location matrix were used in ordination and pattern analyses to assess the degree of similarity among trial locations in Colorado. Information from these analyses will be used to improve and refine trial design in future years and potentially increase our ability to make reliable cultivar recommendations to wheat producers.