It sucks
It sucks
You could start with formally stating your profit hypothesis and ensuring that your algorithm adheres to your intended logic.
Once you've proven that your algo matches your intent, you would span outward with regards to either security selection and expansion (trade more than one instrument), [limited] parameter tuning, application of trailing stops.
You could perform a rudimentary edge measurement by using a period stop, exit after X periods. If your strategy has any edge then, across 50+ securities, over 4+ years, it should show a profit greater than, oh the common 9% return of the broad market returns over the last few decades.
• Form a hypothesis.
• Build an algorithm to test the hypothesis.
• Test the hypothesis.
• Trade hypotheses that reject the null.
• Repeat as needed until rich.
Greater potential than this chart implies. Some ideas here, and let me know your thoughts.