Abstract
A neural network trading model is developed for hard red winter wheat and Deutsche Mark futures markets. The inputs to the neural network are lagged prices. The results are generally unfavourable. The neural network does not produce statistically significant profits.