How did you avoid “curve fitting” when you built the CCM Market Model?
There are two basic ways to build a model:
Method A: Start with the historical data. Use software to build a model that works well with the data. This is known as “curve fitting”, or building a model from a specific data set. This is not the way to build a robust model that can add value in the future. We purposely avoided Method A.
Method B: Build a model based on rational economic and common sense principles. For example, we all know intuitively that when investors are confident about future economic outcomes they would prefer to be in growth-oriented stocks, rather than defensive bonds. Therefore, we would expect that the ratio of stocks-to-bonds would rise during a bull market and fall during a bear market. The CCM Market Model was built using inputs that pass the “that makes sense” test. Once the model was built, historical data was used to backtest and refine the model. Method B is the proper way to build, backtest, and refine a model. We used Method B.