Agricultural economists' research on price forecasting and marketing strategies has been used little by those in the real world. We argue that fresh approaches to research are needed. First, we argue that we need to adopt a new theoretical paradigm, noisy rational expectations. This paradigm suggests that gains from using price forecasting models with public data or from using a marketing strategy are not impossible, but any gains are likely to be small. We need to conduct falsification tests; to perform confirmation and replication; to adjust research to reflect structural changes, such as increased contracting; and always to conduct statistical tests. We also provide a modest agenda for changing our research and extension programs.