Sometimes the most basic tests provide the biggest insights and can later be used to refine other tests.
Idea: Buy or sell in the opposite direction of the day’s trade, i.e., if today was up from yesterday’s close, then go home short. If today was down, go home long. Exit at the close of the next day.
Results:
- 50% were profitable
- The average gain was +0.2 pts
- Longs were roughly twice as profitable as shorts
Refinement: What if the close was above/below the prior day’s high/low?
Results:
- 52% were profitable
- The average gain was +0.8 pts
- Longs were roughly 1.5 times as profitable as shorts
Refinement: What if the close was +/-10 above/below the prior day’s high/low?
- 55% were profitable
- The average gain was +1.7 pts
- Long and short expectation was roughly equal
- The results were statistically significant: z=2.7
Conclusion: We used a very simple idea, reversion, to gain an insight into which days are most likely to lead to a reversal day tomorrow. Simple tests are easier to run the complex tests and can be used as refinements in later tests adding further to their value.
Test data: SP futures, 1997-present