Problems

I found some problems with my model that reduce the return. While somewhat frustrating, the system still aids, more than enough, in forecasting price direction. In order to make it more robust I'll have to put in quite a bit more effort, which, at the moment, doesn't sound like something I want to do. But, I need to get to that point if I want to improve upon the system enough to make some serious money!

 Push, push, push.

On another note, I dabbled in some pairs trading research the other day. My expected returns originally stated something like 50%, which would be unbelievable. I still got excited. Until I realized I was looking on return based on change in the spread, without taking into consideration the amount of money needed to invest in the spread and the inevitable transaction costs.

Those two factors ultimately reduced the setup to practically break-even, probably with a slight return in the positive direction. If I could pin point a way to reduce transaction costs, ideally by at least half, the system would work well enough to implement. At least, that's how it looks.

Damn errors and transaction costs!