What Makes an Automated Forex Trading System a Winner

In this article I would like to present to you the best principles one should focus on in filtering what makes an automated Forex trading system a consistent winner throughout the years, after its creation during real live trading. There are certain metrics and concepts that are available and can show us which strategy is likely to perform well when real money is put on the table.

They are a total of 5 and every one of them is important and one cannot make up for another one. Let`s start!

Long-term backtesting

The best auto systems have performed well during a wide variety of market occasions. As we all know from experience markets are always changing from quiet to volatile, from trending to sideways and corrective, from uptrend to downtrend etc. When a trader backtests a system during all mentioned market cycles it gives more confidence that it would gain again in the future. The longer the backtest period, the better.

It is imperative to check as many different market scenarios beforehand as possible, which can be done only using a large time span for backtesting. There is no minimum threshold for the number of years, but at least 5 years must be covered regardless of the trading strategy.

Simple trading logic

Many automated strategies fail because they are just curve fitted to the past data. Fooling yourselves with a great looking equity curve on backtesting is not very lucrative and can be avoided. We should bear in mind that we are examining a past period in order to exploit the fact that the markets tend to be similar over the years. Similar, not exactly the same!

Different market patterns could be tracked back many, many years ago – like double tops and bottoms, head and shoulders, etc. Since the markets act similarly, we should leave them a breathing space to do so and not to expect a strong bull trend 5 years ago to be repeated exactly the same next year. One of the ways for giving the leeway is using simple trading logic without having a lot of inputs. From 2 to 3 major settings are more than enough if we want to remain within the simple area.

If an auto Forex trading system has 7 major inputs and each of them can be set up in 10 different levels, then we can end up with 10x10x10x10x10x10x10 or 10 000 000 possible sets. We can agree that out of so many different results at least few of them would be very profitable. The important question here is what is the likelihood of these sets to remain winning in the foreseeable future. The less degree of freedom (inputs) the lesser the chances of overfitting the past data (bigger the leeway).

Robustness test

After its creation every Forex system should be tested for its robustness. There are many ways that could be applied and I would like to share with you the one that is most difficult to pass and thus most secure.

Changing the chosen inputs by a lot would give us the needed confirmation whether we have curve fitted or not. Since the markets are similar we should expect them to behave a little bit different compared to the backtested period. By changing the settings we are checking if our automated Forex strategy would perform good if the market conditions are different.

If a small change in inputs results in a big difference in the end results, it signals to us very loudly that the system will be profitable only if the market conditions repeat 100% from the backtested period, which for sure no one is expecting.

It shouldn’t be expected that every single set of inputs will be profitable either. The greater the change of settings and subsequent good looking equity curve, the more robust the Forex strategy. For example if the entry inputs are changed by 220% and the system remains profitable – this is the kind of behavior we would like to see, to call a system robust, not curvefitted.

Not broker/spread and slippage dependency

When an auto Forex trading strategy is a high-frequent trader and thus has a very low average trade measured in pips, then the real-time performance is very dependent on the commissions we pay to the broker, the spreads we are working with and the slippage.

The latter could be severe when it comes to publishing important news like non-farm payrolls and interest rates updates. Since the backtest environment cannot simulate 100% the real trading conditions because of using only fixed spread, no slippage, etc. one would need a long time of live trading in order to see whether this particular system is profitable or not. It is a kind of a validation test after the backtest.

Most scalpers and arbitrage systems fall into this category.

You can avoid deterioration of live results compared to backtest by focusing on systems with high average trade in pips. Like 4+ pips as a bare minimum and choosing a realistic or even bigger spread during the backtest. These precautions will put you in a safe zone and you will get similar to backtested results in your live trading.

Thus even a big slippage and occasional wide spread will eat only a small part of your profits. If you subtract 0.5 pips slippage from your 4 pip average trade, there would be enough remaining for you. If you are working with 0.5 pips average trade, then it can lose all profits in spreads and slippage.

Avoiding dangerous approaches

Long-term profitable automated systems usually don’t use any of the approaches I will include in this section. By avoiding them, one will save a lot of money. They may seem compelling sometimes but most of them just increase the risk in order to create more profits and Return/Risk ration won’t go up.

– Martingale money management: you are increasing your risk when a losing trade occurs.

– Averaging up/down: you are adding to a losing trade.

– Tight profit target: your take profit is set to 1-2 pips.

– No Stop Loss: you don’t limit your losses.

– Tight Stop Loss: you don’t allow the market to fluctuate and often you are stopped out.


By applying all the above mentioned approaches the odds of successful real-time trading are much, much greater. If a trader backtests a simple, not dangerous system for a long period of time and it passes the robustness test, then the odds are in his favor. It is a very conservative approach and most of the automated Forex systems will fail to pass the test and this is the way it should be – only a very small proposition of trading strategies is making money long-term and our task is to focus only on them by filtering the rest.


Deixe uma resposta

O seu endereço de email não será publicado. Campos obrigatórios marcados com *