The answer is through understanding the performance statistics that the system is generating. When you run a backtest, you get information such as the number of trades, percentage of winners, percentage of losers, size of wins, and size of losses. You also get an equity curve that tells you how your account would’ve grown over time, and other information wherein you can monitor those statistics over time. If you see your win rate dropping, your average trade profit, profit per trade drop, your drawdown in your equity curve getting bigger over time, these are all warning signs that the system or the rules are no longer working as well as they did.
Therefore, once you’ve got a set of rules, you have to understand how do they perform and what are the statistics that you should expect from those rules. After that, you monitor those statistics and check for changes over time. I backtest them every day with the systems I’m running in the stock market. My systems are end-of-day systems, so I run them once a day. If I’ve got a new trade take, I’ll take it at the next open, whether it’s a buy or a sell. Depending on the system, it all happens then at the open or at the close. However, I run the backtest every day to generate my signals and look at what the system is doing. If the system is degrading, I watch it every day and see what’s happening. The easiest way to tell is by looking at the equity curve. If the daily changes in the equity curve and the slope of the equity curve over time are deteriorating, you revisit the system and investigate what’s happening.