Modeling techniques forecast future behavior based upon available data. Traditional models use linear or differential equations to create deterministic models that use the available information to estimate parameters in the model and then current information is used as input to the model to forecast subsequent evolution. The idea of fractal or chaotic forecasting is to use historical data that closely matches the pattern of the current situation, and only utilize the most relevant historical data for the construction of the forecast. The technique is fairly simple and can be effective, especially if the phenomenon contains subtly hidden, deterministic, nonlinear behavior. For more discussion of fractal forecasting, see [1,2]; earlier versions of some of these experiments and some variants of the implementation appear in [3]