Farmers evaluate the weather and soil conditions when choosing the right seeds to maximize success of their crops. They look to the past and do a lot of guesswork about the future. What if machine learning could do it for them? Two Washington University in St. Louis researchers devised a computational model so farmers and seed-makers could take the guesswork out of which variety of seeds to plant each year. Farmers could receive recommendations based on simulated success rates of the five best seed types to grow- given the average yields, weather conditions and soil composition of his or her region. The optimization comes to agriculture through a web tool called SimSoy, or SimSeed.