Predicting the performance of a company’s stock for decision purposes is typically made using a scientifically rigorous method known as technical and fundamental analysis. In this paper, such techniques appear insufficient for potentially extreme decision making situations. For argumentation purposes a typical ‘random walk’ high volatility stock market scenario is reformulated using derivative instruments, as well as CFD’s (Contracts for Difference), as a way to control the interplay between results and risk. In the process attempts are made to transform an ‘ill’ structured decision situation into a manageable solution that is supported by an N*M factorial experimental design. The treatment consists of different types of Decision Support Systems (DSS) architectures that range from a simple calculator to an experimentally induced intelligent STOP and LIMIT mechanisms that control the critical entry and exit portfolio conditions. In the conclusion we discuss the results obtained in a laboratory experimentation as they appear “too good to be true” In particular, the results challenge the economic market efficiency principles with, it’s classical “no –arbitrage’ clause” and ‘portfolio diversification’ principle.