ABSTRACT
As someone said, prediction is always uncertain, and especially so when it concerns the future! In situations of self-contained universes where repetitive behavior occurs over long periods of time (e.g., the stock market), we can study the past to see if it gives us a basis for predicting the future. From this study we may construct a method for dealing with the behavior of this universe as we record similar patterns of behavior recurring with necessary and sufficient frequency to give us some statistical reliability in predictions. We then back-test our method on fresh data to determine its performance on recorded history. Finally, as in the case of the market, we may paper-trade the method to see how well it holds up in real time. In all this we are not looking for perfection but for an acceptable level of efficiency.
