University crop variety trials are used to screen breeding lines and to obtain unbiased results to improve variety recommendations to crop producers. Dryland winter wheat yields are highly variable in the Great Plains which are prone to variable plant stands, spring freeze damage, drought, high temperatures at grain fill, and disease and insect infestation. Non-experimental error (NEE) and NEE interactions with varieties and environments are often ignored. Wheat producers need fast and reliable interpretation of results from multiple environments to make planting decisions before the next planting season. Recent variety trial analyses purport to improve predictability of variety performance by using spatial techniques to adjust data within trials, reduce variability, and increase the ratio of genetic to environmental variation. Several simple analytic methods are applied to data from 11 dryland winter wheat variety trials in 2007 and from 2005-2007 (three years) to help producers make better variety planting decisions. This study uses probability analysis to improve predictability of variety performance in future environments. Probability interpretations benefit by including all trial results and can be used to compare varieties for the probability of obtaining economically important, non-mean value, target goals for multiple traits; for example, comparing the probability of different varieties to yield above 2687 kg/ha and test weight above 772 g/L. The wisdom of applying spatial adjustment to small plot results is questioned for practical and theoretical reasons. Spatial methods confound experimental and non-experimental error, tend to produce non-repeatable results, and depend on specialized software requiring more advanced statistical skills. Variety planting decisions may be improved by interpretation of variety trial results with probabilities in highly variable environments.