Adrian Reid argues for a single-pass optimization approach to trading systems: brainstorm rules, code and test, optimize once, add filters, then run robustness checks. He says endless iteration loops waste time and capital. The goal is to reject bad ideas faster and protect your money.
How do I know when to stop optimizing my trading system? When is it enough? Most people will optimize and then change something and then optimize again and then change something else and add another filter and then optimize again. And there's this kind of looping kind of process that feels never ending and actually is never ending. It's a perfectionism that actually won't help your trading. My process for back testing and developing systems is a once through process. Brainstorm the rules, code them up, test them and optimize. Add filters or brainstorm ideas to improve that raw system. Then significant robustness out of sample testing just once through of certain stages that I go through. And optimization only happens one time. So I'm not looping and re optimizing, it's once. So when you have that linear process, you don't have to make decision about when to stop. You just do each step in the process. And if you don't have a back testing process then you know, we should talk because you need one. Because if you do it the same way every time it gets faster one you get better at it, you make better decisions, you'll be able to eliminate or reject the bad ideas more quickly so you can get onto the good ones. Having a solid backtesting process really accelerates your trading journey. That's one of the most powerful tools for a systematic trader. Once you know how to back test, get a process that is the same each time. It's like a pilot's checklist. When a pilot gets on the plane they don't just go oh, should I check the fuel gauge today? And look, the wings are nice and clean and oh, that engine hasn't started. They didn't just, they didn't do that. They have a pre flight checklist and they check everything. Why do they check everything? Because if they don't, people die. If you don't back test the same way with a checklist each time, guess what? People don't die. But your account's going to get hurt. So having that process will tell you when to stop. It'll shorten the process, allow you to get in more cycles, more testing idea cycles which will get you the good ideas more quickly. So hopefully that helps. It.
