In dryland agricultural systems, efficient farm management requires a degree of flexibility according to variations in climate from year to year. Tactical adjustments to the mix of farm enterprises can capitalize on good growing conditions and minimise losses under poor growing conditions. In this paper, a discrete stochastic programming model of dryland wheat-sheep farms in Western Australia is used to identify optimal tactical adjustments to climate and to calculate the value of these tactical adjustments. The model, MUDAS, includes nine discrete season types with a wide range of options for tactical adjustments in each. In the standard model, optimal tactical responses increase expected net cash surplus by approximately 22% relative to a fixed or inflexible strategy. In most season types, changes to the long term farm strategy are made on less than 10% of the farm area, although in some seasons over 25% of the farm can require adjustments to the enterprise selected. The benefits of flexibility are not evenly distributed across different season types but occur predominantly in the best and worst seasons. The magnitude of benefits is affected differently by different commodity prices. Benefits of flexibility are due to capitalizing on knowledge about the greater volatility of profits from cropping than from livestock production. Deterministic models and even stochastic models which don't include activities for tactical adjustments miss this key feature of the system.