Hey guys,

First off, thank you to anyone who helps out! I haven't been in statistics for a while and, quite honestly, learning the T-Test online hasn't been very clear, which is why I turned to people much smarter than myself.

Let's say we have a trading algorithm which generates buy/sell signals on a basket of 6 stocks. You continue to test this in the stock market and notice that it has been doing well, but are not sure whether the results are due to luck or actually mean something. Furthermore, you want to know how likely the results can be replicated in the future (predictability).

I was hoping you could help me understand which statistical test would best suit this situation and what its parameters should be. Also, how many trading days worth of data would be sufficient enough to form an adequate sample size?

Thank you again!