Deciduous fruit and nut trees fall dormant during the cold season, and they usually require exposure to cool temperatures, known as ‘chill’, to break out of this state. Global climate change will almost certainly lead to warming winters in most growing regions, giving rise to concerns that the productivity of orchards will be compromised. Accumulating evidence of chill-related production problems in places that rarely experienced such challenges in the past highlights an urgent need for adaptation strategies. Yet adaptation requires anticipation of the challenges that lie ahead. To adapt to climate change, we need to understand the magnitude of upcoming changes (exposure) and a reliable mechanism to translate this understanding into predictions of system performance (sensitivity). Through climate scenario ensembles, we can reasonably capture uncertainty regarding climate change exposure. But do we have reliable tools to predict the effects of projected changes on deciduous trees? Chill modeling in particular has hardly progressed over the past 30 years. Models differ in their sensitivity to temperature change, they predict substantially different climatic requirements for the same cultivars in different locations, and they often leave a sizeable share of bloom date variation unexplained. More importantly, none of the common models are based on sound and up-to-date understanding of the biology of dormancy-breaking processes. These deficiencies can have serious implications, because using an inaccurate model to produce precise (but possibly wrong) predictions cannot only be useless – it can lead to wrong impact expectations and, as a consequence, maladaptation. To adapt deciduous orchards to climate change, we need renewed efforts to produce better models that synthesize knowledge on dormancy processes across scientific disciplines. Such tools should be capable of projecting changes that can be anticipated, while recognizing uncertainties resulting from our imperfect understanding of the complex biological organisms we work with.