We investigate the role played by both production and market risks on farmers’ decision to adopt long rotations considered as innovative cropping systems. We build a multi-period dynamic farm model which arbitrates each year between conventional and innovative rotations. With discrete stochastic programming, the production risk is accounted for as an intra-year risk, yearly farming operations being declined according to a decision tree where probabilities are assigned. The simulations for a sample of 13 farmers who are currently experimenting this innovation in south-western France, show that substantive sunk costs act as incentives to remain in the long rotation when the farmer is supported for his engagement. They also show that both a high risk aversion and a highly positive market trend tend to slow down the conversion towards innovative systems.