Some stock screeners offer backtesting. They may sell everything and buying 10 top stocks passing their criteria once a month.
There can be many different rule sets, all with stellar returns over different periods of backtesting.
I suspect it's just too simple, otherwise everyone would be doing it. Is it? Why?
You are right about one thing, backtesting is a way of testing the performance of a set of rules for picking stocks over a period of time in the past. You can do the same by paper trading the set of rules into the future.
Not all sets of rules will give you positive results and even some sets of rules that may give you positive results over one period of time may give negative returns over a different period of time.
The benefits of backtesting is that you can, in a short period of time, check the performance of different criteria over different time frames to see which criteria or set of rules work best over various time frames and various market conditions.
A robust system with a set of criteria that performs reasonably well over different market conditions is better than a system that has been tweaked to perform very well over one period of time and one market condition only.
