Even though backtesting is only of limited use, you should try your algorithms on fresh (you haven't previously used them) time periods and see what types of re

Even though backtesting is only of limited use, you should try your algorithms on fresh (you haven't previously used them) time periods and see what types of returns you can get. When I did this with my algos I was playing with, even though it seemed positive, when I accounted for the cost of the trades and the cost of the data (near-real time market data is damn expensive), I wouldn't come out much ahead of a simple buy-hold index fund approach. Invest the minimum amount you can in this project to learn that it isn't a great idea.