I have a strategy that shows profitability in backtesting (including slippage and with Calc On Bar Close = true).
Basically, it checks slope of a 50-period SMA and, when that is above (for uptrends) a certain value, it looks for a shorter Linear Regression Slope to cross above zero, suggesting there'd been a pullback and that a continuation was starting.
The issue is that, parameters that are profitable for each 2-month period are different from other 2-month periods. My optimizer does not come up with parameters that are profitable over a year, for example.
Can anyone suggest ways to exploit the profitability that seems inherent in the strategy?