Adrian Reid explains why traders fail when they optimize for the single best historical parameters. Instead, test many combinations and pick strategies that stay profitable across broad parameter ranges. Systems that work only on one perfect setting will crash in live trading.
Have you ever wondered why, despite all of the cool tools, the sophisticated analytics and the huge amounts of data, why most private traders still can't make money in the market and don't have systems which are consistently profitable? Well, the reason came to me in a flash when I was having a conversation with a student a little while ago. They programmed Metatrader and they set up the optimization through brute force test hundreds of thousands of companies, combinations of trading rules and what the intention was. Their aim was to choose that optimization, to choose the set of parameters that worked best. Now the problem is this. When you look for the best set of parameters on your historical data, it's exactly that historical data. It's not forward looking, it's looking at the past. And so even if you choose the best set of parameters for last year for your trading system, that's absolutely guaranteed to not be the best set of parameters for the future. Now the problem is most modern trading software is set up to make optimization really, really easy. But the trouble is it's all set up for the wrong reason. What you need to do as a private trader, as any trader for that matter, is not look for the best combination. What you need to do is look for a strategy, a set of rules, a system which performs well no matter what set of parameters you use. You're not choosing a system and optimizing it to find the best possible parameters for you to trade with. What trading system development is about is formulating a strategy and a system and optimizing or testing all different parameter values to make sure that that system is stable and works no matter what parameters you use, or maybe not quite no matter what parameters you use. But you certainly want a system that to perform well and be profitable over a very wide range of parameters. That's totally opposite to what this student was doing. They were optimizing 2, 300,000 combinations of variables to find the one best one. They wondered why when they did a walk forward test, it wasn't working. It's because they were missing the point. So after a brief conversation, what the redirect was that I gave this student was don't look for the best combination. You're looking for a strategy that works across as many combinations of parameters as possible. That way it doesn't matter what happens in the market, you know your system is going to be pretty stable. Choosing last year or last decade's best parameters is absolutely not going to work in the future as well as it did in that time, because there are hundreds of thousands of possible parameters that you could have chosen. And by looking at the past, you've chosen just one that you considered one. Optimal optimization is not about getting the optimal solution. It's about finding a set of rules that are stable and perform well over a very wide range of parameter values. So that way, no matter what happens as the market shifts and change moving into the future, you know that your parameters will probably work, because most of the parameters you test in your system have worked. And once you've got that confidence, it doesn't matter quite as much precisely what parameters you choose, because you know that the system will perform well in the future, no matter what the future may bring. That's the true purpose of optimization, to get a stable system, not to get the best set of parameters and the best possible historical performance. That's what most system vendors and most people selling trading systems and trading advisory services and all of this sort of thing miss, because they talk about how accurate their system was, how great their back test was, how awesome the historical equity curve looks like. None of that is relevant in the future. What's relevant in the future is how stable the system is, how likely it is to keep on performing under different sets of market conditions. So that's the purpose of optimization of your trading system. And that's the reason why most people who develop trading systems nowadays fail, because they're using the software the wrong way. And it's not your fault, it's not the trader's fault, because the software is set up to push you down this path of using these cool tools and fancy analytics and finding the best solution. But what you need, what you need to do in order to develop a trading system that works is not optimize to find the best answer. It's not doing a brute force optimization, testing hundreds of thousands of combinations. If you want to test hundreds of thousands of combinations, that's great. But the purpose is not to find the one best solution, the one best parameter set. The purpose is to make sure that if there's 200,000 combinations of parameters that you're going to test, you want to make sure that 150,000 or more of them are, are profitable and tradable. Because if you get results like that out of your optimization, you know you're onto a winner. Because you know that no matter what setting is optimal in the future, your system is probably going to make money. So when you're looking at a set of rules, don't look for how profitable it is in isolation. Of course it has to be profitable. But what you got to do is look for how many combinations of those rules, how many parameters combinations that you use are actually profitable in your backtest over different market conditions. And when you've got a system that, when you optimize it performs over a huge range of parameter values, you're onto a winner. And that's where you're going to make money trading in the Real World in the Future if you're interested in learning more about how to develop your own stock trading system that has a good chance of actually working in the future in the real world and not just in your historical backtest, then enter your name and email below to find out more about the Enlightened Stock Trader Development Program. Because in that program I take you step by step through the process of developing your own stock trading system that's robust and stable and will keep working well into the future. I've got trading systems that are profitable that have been working now for over a decade and keep on bringing in the money. And if that's what you want, then enter your name and email in the box below and I'll see you in the Enlightened stock to our development program. Bye for now.
