Emotions can be devastating to a traders bankroll as they can derail even the best laid plans in a blink of an eye. If you have trouble sticking to a position once you have put it on, or if you have problems pulling the trigger when a perfect setup lines up you could benefit from using a trading system. Here are four crucial steps to take.
Form a hypothesis
Trading systems are crucial for keeping the ‘monkey’ side of your brain at bay and many of the most successful traders use systems. However, the best systems are not created out of random patterns or ideas. Trading systems that have no theoretical grounding often perform well in back-tests but fail when put to work on live data.
It’s crucial therefore to create a system based on a hypothesis or pattern that you believe exists in the market for fundamental reasons.
Equip yourself with the right tools
Once you have a hypothesis in mind it’s time to get to work building and testing the system you want to trade. System traders do all their work in the designing and optimisation stages. The trading process actually become incredibly easy since all system traders really need to do is press the button on their strategies.
The real work is done out of the way of the market and it’s important to have the right tools from the start. You’ll need a strong and reliable back-testing program such as MetaTrader, TradeStation or Amibroker and you’ll need to learn enough about programming to code up your strategies and indicators.
You will also need the highest quality market data you can get your hands on. If you design a system with poor quality data there’s little chance that it will stand up to real trading.
You’ll need enough data too, otherwise you won’t be able to draw valid conclusions from your tests. For end of day systems think 5-10 years of historical data, for intraday systems think 18-24 months worth.
When you start running back tests on the historical data using the buy and sell arguments you have created you’ll find that your system may perform erratically over various settings.
There’s no time to manually go through every different combination of settings available so this is where optimization comes in.
By optimizing the system you can cycle through thousands of different settings combinations in order to find which ones work the best. It’s a good idea to start off with some very rough tests and then focus in on the settings which work the best.
Doing so, you’ll end up with a much higher performing strategy which can then be tested on out-of-sample data. Do this just once to see whether the system works on clean data.
There are many important parts to creating a robust and profitable trading system and there isn’t the space to go into all of them in detail.
The most crucial aspect of designing a trading system, however, is making sure to avoid the problem of curve-fitting.
Curve-fitting occurs when a system is too well optimised to past data that it is unlikely to work in the future. Statistical illusions can fool traders into thinking that a system is robust, but when it is used with real money it can fail spectacularly.
In order to avoid curve-fitting it’s important to remember a few guidelines.
Firstly, do not build your trading system with too many variables. Doing so, increases the chances of curve-fitting exponentially.
Second, always keep a good chunk of data out-of-sample so that you can test your system free from bias.
Thirdly, remember that no data is ever truly free from bias so make sure to paper trade the system on live data before putting it to work with real money.