Many discretionary traders (manual style of trading when you open and close positions solely on the basis of your own opinions and analyses, without assistance of a programmed code) demonise and execrate Automated trading systems (ATS). The main argument of discretionary traders is that a programmed code can never accurately identify certain nuances in markets which can only be detected by the human eye and processed by the human mind. I believed this argument for a very long time, perhaps simply because I wanted to believe it. I never programmed anything by myself, I studied rather statistical processes and programming gave me a wide berth. Therefore I thought that programming was definitely beyond my abilities and that I could not handle it because I’m not an exceptionally intelligent person. But to my own detriment I was actually robbing myself of precious time. At the beginning I blindly believed in the discretionary access (of course, it can be very profitable if applied properly) and I missed the huge potential of ATS. Honestly, I do not remember when the breaking point came, but after this moment everything happened very quickly.
During my personal meetings with foreign traders and reading foreign discussions I always found out that advanced traders used the analytical platform TradeStation. So I decided to study this platform in detail. I found a promo offer on the Internet within which you could download a 60-day trial version of TradeStation. After this my life of a trader completely changed. I began to thoroughly study all aspects of the software. Its basis is the EasyLanguage programming language so I started to intensively study this language by the use of various tutorials, articles, and books. I was surprised by the speed of my progress. Today, I am able to program most of my ideas that I get through observation of markets. The knowledge of the programming code brought me new possibilities of my systems´ robustness verification, such as Walk Forward Cluster analysis, optimisation tests, sophisticated performance evaluation of trading systems, and so on. A completely new world of unprecedented possibilities opened to me and I would like to share these possibilities with you. Discretionary trading is based on combination of various visual and technical tools (such as trend lines, supports and resistance, double bottoms and double tops, crossings of different moving averages, various indicators, and oversold or overbought areas) for evaluation of trading opportunities according to fulfilment of predefined entry conditions. Of course, also exiting from markets works on the same principle here. Positions are opened and closed by manual submissions of orders. In other words, if you want to execute an order you must by physically present at the computer. I. e. the trader decides about executions of his orders in real time. This technical analysis is usually complemented with the trader´s subjective view when the trader evaluates the overall market situation – whether the market trends or stagnates (is in a “chop”), whether it grows (bull market) or declines (bear market), and so on. A disadvantage of this approach is a very slow and limited backtesting of our trading strategies. However, it is necessary to say that this approach can be very profitable.
Automated (also called systematic or algorithmic) approach to trading, on the other hand, is based on creation of a programmed code defining entry and exit conditions that are then applied on a particular market or group of markets. This approach allows fully automatic executions of trades, i.e. trading without the physical presence of the trader and the necessity to manually enter orders. An advantage against the discretionary approach is the possibility of a very fast and accurate backtesting and the possibility of carrying out a large amount of arithmetic operations that ensure quality of optimisation and robustness tests (a robust trading system is a system that is likely to be profitable in the future, i.e. during the yet unknown price development). A disadvantage of this approach may be a distorted or inaccurate setting of entry and exit conditions which may result in failure of the programmed strategy. And it is the correct definition of conditions for executions of individual trades that determines profitability of each automated system. We obtain these conditions by long-term observations of various nuances in individual markets and by extensive backtests and optimisations. I believe that if you are new to this area you are still not clear about many things. For this reason I have prepared for you a demonstration of several differences in building automated and discretionary trading systems. In the following table you can see a typical procedure of building, testing, and implementation of a trading system.
|1. Observation and market analysis||This step is common to both approaches. The point is to define the basic ideas for achieving edge = a set of rules the application of which will result in achieving a long-term and stable statistical advantage in the market. Example: At the beginning of the day’s trading session we see a sharp increase in the number of traded contracts (volume) in comparison with a defined number of the previous days. At the same time the market significantly moves to the Long side and it has a tendency to grow until the end of the trading session.|
|2. Specification of rules for the subsequent backtest||It is also common to both approaches. After we have identified a certain tendency in the market we have to give our assumptions a concrete form. Example: If during the first 30 minutes after the session opening the highest number of contracts (volume) over the last X days is traded and at the same time the price increases during these 30 minutes, we enter into a long positions by the MARKET BUY order and we close the position at the end of the trading session by the SELL MARKET order.|
|3. Backtest – testing of the trading system||At this stage traders using ATS must define the algorithm in the programming language used by his software platform (e.g. EasyLanguage in TradeStation). After a successful writing of a fully functional code a trader using ATS gains a huge advantage over a discretionary trader. It is because the code of a trading system can be applied to many markets and it is possible to immediately verify whether the system can be profitable in a long-term and under different conditions of individual markets.||A discretionary trader must manually go through price charts in his software platform. This process is time-consuming because it is necessary to manually scroll bar by bar. The trader must be very concentrated and do not make any mistakes in order to secure statistical relevance of the results. “Manually” means that all trades that meet the selected criteria must be recorded in the trading diary (e.g. in Excel). He has to record entry and exit prices and the resulting profit or loss for each trade.|
|4. Optimisation– searching for entry and exit rules´ optimal values||In our particular example it is the highest number of contracts (Volume) over the last X days. No problem for a trader using programmable algorithms. He just defines the range of the entry parameter values in his software platform. In our case, it may be for instance a value from 1 to 100 (the number of previous days). The software platform then performs the calculations and sorts the parameters by a certain performance indicator (from best to worst results).||In our particular example it is the highest number of contracts (Volume) over the last X days. A discretionary trader would have to manually go through charts over and over again in order to find the values of individual parameters. It is a time-consuming and laborious process that would take years, maybe longer. In other words, it is virtually impossible to optimise parameters in discretionary trading. On the other hand, we must note that many discretionary traders achieve great results and do not need any parameter optimisation.|
|5. Trading system performance and robustness evaluation||In programs using the EasyLanguage programming language you can very easily display performance results for multiple markets, even if you set different ranges of entry parameters in each of them. Furthermore, professional platforms include various verification tests for assessing the statistical robustness (in other words, sets of many In-Sample and Out-of-Sample tests). The aim of these robustness tests is to verify whether the strategy is likely to generate profits even in real trading, i.e. in the future. This topic is closely related to the WALK FORWARD ANALYSIS.||If a discretionary trader conscientiously records each transaction in an Excel-type spreadsheet, it is not a problem to define functions displaying performance evaluation. But it requires an advanced knowledge of working with spreadsheets.|
|6. Real-time application of simulated and live trading strategies||The trader practically does not have to sit at his computer. Execution of both the entry and exit orders is performed automatically by the predefined code. There are traders who deliberately do not watch their trading strategies during trading sessions. It may be for several reasons – they may prefer to spend their time in a different way than by sitting at their computer, or they avoid the tendency to interfere in trades under emotional pressure.||
Time-demandingness of trading depends on its style. If you keep your positions open for longer periods, i.e. for several days or weeks, you are “positional traders”. You check and analyse markets usually once a day and, if needed, you adjust your positions. The second style may be the intraday trading where you open and close your positions within one day. In this style of trading your physical presence is required throughout the trading session. It is therefore a very time-consuming approach which, however, may bring accordingly higher profits.
Table 1: The classic approach to building, testing, and real implementation of a trading system.
As we have already said, there are two groups of traders the approaches of which are completely different. Those who believe that discretionary trading is the right way to a stable and long-term profits. These traders usually do not trust algorithmic trading because want to have everything under their control. And then there is the second group which maintains that trading via ATS has many substantial benefits in terms of backtesting speed, optimisation, and statistical robustness tests which can be never achieved by a discretionary trader. But of course we must mention traders who trade both by using discretionary approach and ATS.
Novice traders usually fully believe in the discretionary approach and it is the prevailing style of individual traders. Various authors (often purposely) strengthen this belief. Moreover not many of us studied programming. I also hear many arguments like: “Traders who try to define algorithms for their ATS cannot program delicate nuances and the overall context of the market. It brings a huge advantage to discretionary traders because they, unlike traders using ATS, “get a feel” for the markets through numerous observations“.
And what should we imagine under the “feel for the markets”? In essence, it is the concept of understanding the market´s structure, i.e. understanding the principles of price charts gained through long-term observations and evaluations. A discretionary trader with such a feel should be able to predict the probability of the further price development and thus gain a strong edge in the extremely competitive environment of exchange markets. Of course, discretionary traders should be also able to convert these market nuances into clearly defined entry and exit conditions for their trading systems and perform adequate backtesting and papertrading.
The problem lies in the innate human tendency to idealise things. Backtesting or papertrading should be precise and accurate enough, actually you cannot afford such a luxury as a bad backtest in today’s competition. Yes, you can artificially worsen the results by 30-40%, but how can you know that the worsened results correspond to reality? I recommend a very useful book by David Aronson that deals with this topic: Evidence-Based Technical Analysis, which provides extensive and statistically relevant studies proving that traders tend to distort the results of manual backtests. It is simply their natural characteristics – an exaggerated optimism and belief in own abilities and luck.
Yet I do not say that the discretionary approach does not work! There are many people who earn a lot of money by exchange trading thanks to their strong will and self-discipline. They are honest and strict to themselves and they are able to see and analyse the market sentiment and patiently verify their observations on historical data. For example, discretionary trading strategies such as long-term trading of commodity spreads and options can be very profitable. In addition, today´s software unfortunately does not offer automation of commodity spread trading and therefore you have no choice but to trade them in a discretionary manner.
I admire intraday discretionary traders who keep their positions for minutes or hours only and who are able to adapt to the constantly changing market dynamics and volatility. In terms of managing your emotions during trading this is absolutely the hardest trading approach. It is also probably the reason why I know so little successful intraday discretionary traders.
Each novice trader realises very soon that, willing or not, he will have to face his own ego and the two biggest enemies – his own greed and fear of losing. Traders often experience greed when they have an open profit, i.e. when they are in an open position on the right side of the market. It is because they often get into a situation when their system says to take the profit but they don´t do it in expectation of a further market movement and even higher profit. Yet the market suddenly turns against their open position and they suffer a loss instead. Imagine that you experience such several times. Your losses will reach thousands of crowns or even more (depending on your trading approach).
In contrast, a trader using ATS lets the computer to perform all the calculations and trading actions. He just watches the markets from afar. Traders utilizing ATS often let their systems to run completely independently and only check the results in the evening. Thus they do not have to deal with the emotions that adversely affect the actual trading. Of course, they may experience an inner emotional struggle if they suffer several consecutive losses and their account balance decreases. This is an aspect that relates both to discretionary and ATS-using traders. It is important to be sure about your trading system´s logic, robustness, and statistical relevance so that you know that you will overcome these unpleasant losing periods (which are experienced by all traders) and generate profits that will exceed the losses in the future. The most important thing it is to manage your own emotions. One of the main reasons why many traders fail (including traders using ATS) is a failure to stay disciplined. When getting afraid of failure traders begin to interfere in their trading systems so that they lose the edge they found by a previous tests and verified on historical data. It is no surprise then that they get into a vicious circle of failure within which they open completely random positions without any statistical relevance.
A discretionary trader must stay emotionally calm during the whole time he has an open position, i.e. he faces his emotions during the trading session. In contrast, a trader using ATS usually faces his emotions after the end of the trading session. It is not difficult to assess who has a better starting position.
On the other hand, there is one unshaken evidence that discretionary trading can be very profitable. The most famous discretionary traders like George Soros, Warren Buffett, or Bruce Kovner are multibillionaires. Yet it is clear that these people have an extraordinary talent in terms of knowledge of markets´ principles combined with an extraordinary emotional resistance. The decisive argument why to choose ATS is the incomparably higher speed of building, backtesting, optimisation, and robustness testing of trading systems (for example by employing genetic algorithms). We will explain all these issues later.
Successful discretionary traders are characterised by an in-depth knowledge of markets and trading approaches and the ability to choose the most ideal approach at any moment. They are able to quickly respond to the dynamically changing market environment and they never act under duress. Executions of their trades are precise and very well planned. They are true masters at managing their emotions. You could say that a successful discretionary trader should always stay on top of things and do not act according to his emotions.
In contrast, traders using ATS do not have to solve their emotions regarding proper executions of trades. It is ideal to monitor markets only briefly. Of course, it is necessary to be aware of what is happening in the markets but it is certainly not a good idea to watch them constantly. This is the domain of discretionary traders. Another advantage of using ATS is the possibility of building, backtesting, optimisation, and robustness testing of trading systems. We will tell you more about these procedures in the following chapters.
Previous: 13. Technical Analysis in Trading