The ideas are great, but not sure if they have practical application. I tried the Smoothed Adaptive Momentum Strategy and it really gives poor results, huge drawdowns for SPY, negative results with DIA, MDY, and SMH, even without transaction costs. If I optimize it for the training dataset, it performs badly on the test dataset. The Fisher Transform of normalized prices has definitely very sharp turning points, but backtesting shows the tendency to overtrade and has negative results for major indexes, without transaction costs.