flagler

I suppose if you FFT the time domain price data and identified enough dominant frequencies you should in theory be able to reconstruct the time domain curve and extrapolate into the FUTURE! Of course, things will not be that straightforward, as linear and maybe much more sophisticated curve fitting algorithms are needed as a pre-process to remove as much useless extreme low frequencies as possible, and other filtering techniques could help the accuracy. It could be worth a try. Thanks for sharing your exciting experience.