Agricultural Commodities are a large portion of the New Zealand Economy, and a better understanding of commodity markets is useful for New Zealand market participants, especially in the dairy sector. This thesis studies two of the main functions of commodity markets, price discovery and hedging. The first chapter reports the effectiveness of hedging as a function of spot market design. The second chapter illustrates the reduced explanatory power of the Theory of Storage in recent time as contrasted with the early 1960’s. The third chapter demonstrates that cutting edge machine learning techniques are promising alternatives for learning features of commodity markets.