Sample Size Theory

[Professor1215]

A apologize if this is going to make me look like a rookie, but I figured it's worth asking anyway. If you back test something enough times for it to be both profitable and "statistically significant", does it matter what you are back testing? By theory, could it be ANYTHING, as long as it was deemed "statistically significant"?

[thehoorse]

A apologize if this is going to make me look like a rookie, but I figured it's worth asking anyway. If you back test something enough times for it to be both profitable and "statistically significant", does it matter what you are back testing? By theory, could it be ANYTHING, as long as it was deemed "statistically significant"?

Depends on what you are testing. In sports nothing is precisely constant. Some things stray farther from the mean while some don't (as much). Another thing you need to take into account when back testing is you will not get 100% accuracy due to line changes. You should know the volatility on line changes of what you are testing, and how it relates to end results and future expectancy.