What is a Trading System?

What is a system?

A trading system is a set of rules that unequivocally define the way to trade one or several financial instruments in order to obtain benefits. Experts say that a trading system only exists if it is written, so that there should not be any ambiguity in the investment methodology that is intended to be applied. The rules must be coherent with each other and can not be contradicted.

Advantages and disadvantages of trading systems

When you get into the world of investment in the financial markets, you realize that there are many ways and methods for trading. A first classification of these methods that can be done is in terms of trading with a set of fixed rules (a system) or not, through individual decisions of the trader or operator, who decides when and what to buy or sell.

Depending on our personality, we may be interested in trading with a 100% automatic system as defined previously in this site. This should bring us the following benefits:

It is not emotional, our results will not depend on our mood or ability at the time of executing the transaction.
It is 100% objective and is not subject to interpretation. It is impossible to deceive oneself. A buy signal is always a buy signal, it does not matter if we know that tomorrow a relevant piece of information of the market will be published (assuming that this is not included in the strategy) or that we think that this time the signal will not give good result.
It has few rules. With a limited set of rules (less than 10 and sometimes less than 5) you can create a pretty good trading system. The investor does not have the feeling that there is something unknown that can influence the transactions. Everything is included in that set of rules. If with this limited number of rules the system works then there is no need to complicate it further.
It is quantitative or “measurable”, that means that in advance we will have an approximate idea of the evolution of our results. This is a great advantage for two reasons. The first is that it allows us to be psychologically prepared for difficult periods. The second is that we have some simulation data that will allow to implement money management strategies to improve results.

In all the previous points it is assumed that the investor uses the system without deviating from it. In addition, the trader takes each and every one of the trading signals that the system generate. That must always be the case because if one does not follow his system to the letter, why have you tried it with the historical data? In that case the results will never resemble what you have simulated. If a trader does not follow a system faithfully then he has no system.

Not all are advantages when one decides to use a trading system. Some disadvantages are:

Trading systems do not take everything into account. Any unexpected event in the markets usually increases their volatility. An investor who follows a system should ignore this new information because it is assumed that traders should not deviate from their system and that this should already be included in the data on which it has simulated. Obviously this is not always true and often when the market reacts strongly many systems (especially those designed for a low volatile environment) generate losses.
Often the investor has the feeling that he could do better without a system. When a system enters a period of losses or drawdown, the investor begins to think that he would not take this or other signals, that in its place he would trade in another way. Depending on the investor, trading without a system could bring more benefits or not, but it is assumed that the trader that has decided to trade with a system is because he believes that this system will give better results than trading without it.
The systems do not work for life. A good example is the “Turtles” trading system (a system based on buying and selling new maximums/minimums in the market). A prolonged lateral market made this trading system stop working. Who can trade with a system that produces losses for two years or more? Who will take all trading signals in the third year? Whoever uses a system and realizes that it has stopped working should continue to investigate. Designing trading systems is a task that never ends.
A trading system requires taking all the signals. For good and for bad. Let’s suppose that we have a trend following system with 35% of winning trades (quite typical). That means that 65% of our transactions will result in losses. A minority of transactions will generate sufficient profitability to offset most losses. If we go on vacation or we can not enter an order, and it turns out to be a winner trade then we will have ruined the system only for failing to execute an order, since the results will be altered and until we have a new operation belonging to the 35 % of profits all will be losses.

In short, using a trading system is not easy and does not imply success per se, however their advantages may make systems the most advisable option to achieve profitability on a regular basis in the vast majority of investors.

Requirements of trading systems

Obviously the first requirement is that a system is profitable with its historical data. But that does not guarantee that it will be profitable in the future, and here we have a common reasoning error that is as follows:

A profitable system must have a positive backtesting (with simulation gains), but a positive backtesting does not imply that we have a profitable system.

The scientific method helps us with this common error of reasoning which is to think that with one or more profitable simulations we are facing a system with predictive power that generates profits. The scientific method tells us that we must raise the question backwards and assume that our system does not generate benefits. If we are able to falsify this statement with a certain degree of confidence then we will be facing a profitable system. For this it is assumed that the system has a zero average gain and if the tests indicate that it is not true, that at a certain degree of confidence (commonly 95%) the average gain is greater than zero then we can affirm mathematically (statistics ) that we have a winning system.

This requirement to have a statistically reliable system is very relevant, since an intelligent trader that has to choose between a system with small average gain but statistically reliable, and a system with great gain but unreliable will always choose the first because it only has to increase the size of the position in the market to achieve a system with high gain and high reliability. Trading systems that are traded with real money must have an average gain greater than zero with at least 95% confidence. On the other hand, and as we have seen previously, another important requirement for trading systems is that they should consist of a few simple and well-defined rules. If it were not so, it would be very difficult to supervise its proper functioning.

Trading systems types

100% automatic (non-discretionary) systems can be classified in terms of the way they operate or the particularities of their strategy to generate signals. In this case we would have:

Trend following systems. These systems seek to exploit a trend in the time interval of choice. Until they find the trend they make several attempts so they usually have a fairly low percentage of winning trades, usually 20% to 40%, however when these systems works they win much more than they lose (commonly the triple or 3:1 ratio).
Counter-trend systems. They operate contrary to the trend, so they buy when the market has fallen and open short positions after market rises. Usually these systems have a high percentage of winning trades, but at the cost of a very low ratio between winning and losing trades. A percentage of winning trades typical of a counter trend system would be between 60% and 80%.
Breakout systems. These are very common trading systems which buy when the market exceeds a certain price level. These systems have a little percentage of winning trades but they never lose a tendency due to the way they operate. They are an alternative to trend following systems with the advantage that they guarantee an entry into any prolonged trend. We have already seen an example of this type of system when we talked about the “Turtles” system. In some publications these systems can be found within the trend following category, however, a distinction has been preferred here.
Volatility systems. They are based on the intraday overcoming of a certain level that ensures as far as possible that a trend has just begun. These systems perform many trades but are generally very profitable. They have a percentage of winning trades that can be between 50%-60%.
Systems based on cycles. These systems exploit the existence of market cycles, so that the transactions are generated based on certain dates. Here we have the systems for agricultural futures (soy, corn, cotton, etc.) that are influenced by the seasonality of the crops. There are also systems based on lunar cycles and even planetary cycles.
Prediction systems. They use the most advanced techniques to anticipate the evolution of the price. A good example is the systems based on neural networks that learn and train with the historical data until obtaining a prediction as reliable as possible of the next evolution of the market. There are also mathematical prediction techniques with regressive models that, without using neural networks, make predictions that are adjusted to minimize error, such as auto-regression models with moving averages (ARIMA).

More information on trading systems types in the following article: The Main Types of Trading Systems

Performance evaluation of trading systems

There is a series of statistical performance indicators that are used to evaluate trading systems. These indicators can be used to determine not only whether a system is profitable, but also whether it is viable in the long term. Some examples are:

You can find more information about the main ratios used to evaluate systems in the following section: Ratios to evaluate performance and risk of trading systems