Wouldn’t it be great if you could lie back in your seat and let a computer program do all your trading for you? These programs are known as expert advisors and they can be used on most FX trading platforms to make more profits. It may sound too good to be true, earning money while you laid comfortably on your couch, but it is possible if you make the right selection. That is the big caveat that most companies selling these expert advisors fail to mention to you – that there are more crap EAs than there are good ones. After all, just like scam FX brokers promise to make you rich in weeks, so do many of the software developers. In this post, we shall be looking at the evolution of automated trading first and then dive into the complexity of expert advisors. By the end, you should be able to tell a good EA from one that will only cost you money. With this knowledge, you will become a much more successful trader than you were before. (Why FX Trading Should Not To Be Treated Like A Casino)
How automated trading has taken over finance
Today’s trading world is being dominated by trading robots as is being witnessed on Wall Street. Last year in February, there was a story that Goldman Sachs had fired 600 of their stock traders and replaced them with 200 computer engineers whose role would be to monitor the company’s automated trading programs. In fact, it was later revealed by the company’s CFO that a third of all Goldman staff are computer engineers. Keep in mind, these engineers had very little knowledge of the fiscal arenas, showing that computers are indeed taking over much of the fiscal arenas. (Do you ever wonder: What Would Happen To Bitcoin Price When It Is All Mined?)
The trend is not observed by Goldman Sachs alone, but many other banks offering FX trading. The research shows that there is a lot of trading activity on many trading floors that originates from computer programs – more than 50%. This activity is mainly focused on equities but also stocks and now currency trading. Moreover, about 40% of revenues on the trading floor is generated by the computer programs. Clearly, there has been an increase of automated program trading in the fiscal arenas, and the trend is only increasing. (If you’re interested in stocks, you should know: How to trade on the NYSE)
Should we be worried?
Talk about computers taking over humans is bound to trigger end-of-times scenarios in our minds, like many Sci-Fi movies depict. Such notions are not unfounded because there is indeed evidence across many industries that computers are replacing humans. On the other hand, there is another way to look at it, and that is to focus on the benefit automated trading can have for us. Inasmuch as we don’t like to admit it, computers are better than humans in many ways at analysing data. Most importantly, they don’t suffer from emotional changes as we do thus preventing one of the main reasons so many traders fail in the arenas. Now imagine if you could harness the power of these computers and use it in your own trading? (The most profitable traders know these: 10 steps of successful traders)
There is yet another reason not to be worried apart from the obvious, this one a bit more comforting to the psyche – computers are not always good. Even though the numbers mentioned above by Goldman may be troubling, ask yourself why they haven’t completely replaced human traders. Computers can only learn to trade the arenas by analysing historical data, similar to technical analysis. However, any trader knows that there is a lot more to the arenas and trading than just that. A lot of other factors have to be considered such as news announcements on the economic calendar FX, political events like the looming trade war, emergencies like floods and accidents, etc. Only a human can take all these into account, which is why even with computer programs being used, it is still often a human that pushes a button with information from the computer. (Let us try: Comparing fundamental and technical analysis)
Therefore, it’s still not time to worry about the singularity just yet, and for now, it is just more important to focus on the benefits we can all gain from using computer programs in our trading. Besides, there is a fear even among those using these computer programs that their influence could lead to an ‘uninspiring’ arena. If some day trading programs become readily available to everyone, then the arenas would be completely predictable thereby reducing profits. Given human nature and the need for self-preservation, you can sleep safe knowing the big players would not let this happen, at least not in their time or ours. (Meet other humans at: The Best FX Events And Expos To Attend Every Year)
What about the FX arena?
Automated trading did not begin with the FX arena, but it is now being widely used in the field. This is quite understandable especially when you consider how huge the FX arena is. For example, the top FX brokers offer more than 50 currency pairs to their clients. Can you imagine how difficult it is to monitor and analyse all these FX charts on different timeframes? Almost impossible for a human, but not so much for a computer. In the same symposium that revealed Goldman’s staff changes, they also said that a single computer could replace 4 traders! Remember that is just considering the amount of work a computer can do, not to mention the savings in revenue as a computer will never ask for a bonus. (Revealing FX Bonuses Of Brokers: How To Identify A Real Bonus)
Although the FX arena was not what the bankers intended to use computer algorithms for, they have spotted an opportunity and are taking it. Indeed, the FX arena is even more suited for computer programs than, say, the stock arena because it can be studied more in-depth. Furthermore, the arena is not as prone to price changes due to rumours as other fiscal arenas. (Anticipate these: Main central bank meetings)
Expert advisors in the FX arena?
Expert advisors are computer software designed specially to automate trading on MetaTrader platforms. Nevertheless, there are also EAs for other trading platforms like cTrader and others, except there may be fewer of them. These EAs are the exemplification of automated trading in the retail FX arena, and they have become as common among individual traders as algorithms for institutional investors like those mentioned before. If you have ever been in the search for an EA to use yourself, then you probably know that there are very many of them in the arena. One glance at a website like MQL5 or Google search for an EA will give you plenty of results, and therein lies the problem. (In case you didn’t know, this is: How to launch an expert advisor on MT4)
The problem with having numerous EAs is that they all promise huge returns at the comfort of your couch. What’s worse, many of them are not free, which means you could be paying to lose your own money. The truth is that, many of the EAs in the arena today are either totally useless or out of date, meaning that you will just lose your money in the process. To avoid becoming a statistic, you have to put in more work. Most people who get burned in the FX arena do so because they assume the FX arena is an easy business that can make you rich quickly. But the truth could not be more different. Successful traders know that you have to put in a lot of work to understand the arenas and its intricacies. When it comes to EAs, then you have to do the same, and understand how these programs work. Afterwards, you can then find a good source of these EAs and get started with the actual process. (You should know: How Not To Be Added To The 95% Of Losing Traders)
How do expert advisors work?
Before you run away from the page, this is not a lesson in programming or coding. You don’t need to be a coder to understand how EAs work, but the basic knowledge is essential. Basically, an expert advisor is a software that analyses the arenas using set parameters and then makes decisions from it. I know that anyone with the slightest knowledge in programming would be shaking their heads right now, but that is the most basic definition of an EA. To expound more on this, let’s take a look at some of the ways an EA might work:
- checking for breakouts in the arena – this kind of EA will look for any breakouts in the arena from support and resistance levels at pre-determined prices. For example, you could create an EA that placed a sell order at arena prices at the time whenever prices went below a specific support level. Of course, such a system could be vulnerable to fake-outs (Here is: How to draw S/R levels like a pro)
- spotting changes from news events – because any important news events are known beforehand, an EA can be used to monitor changes during the news announcement. Perhaps such an EA could be set to monitor all news announcements by central banks or other important news events like CPI, unemployment, etc. (Learn: How to work with the economic calendar)
- scalping – this kind of program uses different technical indicators like moving averages, MACD, RSI, etc. to make trades for you. As an example, imagine an EA that sold the currency pair whenever prices crossed beneath a moving average or vice versa. Because these signals are often triggered, there can be hundreds of trades executed every single day (These are the: Best technical indicators and how to use them)
EAs can also differ in the level of permission you grant them. Most EAs still require human assistance to trigger the trade, which means the EA is simply used as a guide. This is usually the most recommended use of EAs unless you are completely sure about one’s effectiveness. However, this would not work on an EA meant for scalping, hence the EAs that have the permission to trigger trades. Some go even as far as taking control of your entire account to make more ‘informed’ decisions about the trade. More on this later. Anyway, such an EA would be able to place orders and even set take profit and stop loss levels for you, even trailing stops. (There are some: MetaTrader 4 advanced features)
This may not fully explain how an EA works on a technical level, but it should at least remove some of the mystery around it. Leave the rest to the programmers to figure out how they can ‘instruct’ a program to check moving averages, S/R levels or whatever. (These are some: Uncommon technical indicators)
Where can you get a good EA
There are a lot of advisors available for use on MetaTrader on the internet as we mentioned before, the problem is actually getting a good source. One of the most popular and common source of EAs, signals and other automatic trading tools is MQL5, which is the official website for MetaTrader tools. However, just because it is the most common does not mean you can trust everything you find on it. Just like any other source, there are software developers who may not have fully perfected their EAs yet claim to have the best. We shall see later how you can distinguish the good EAs from the bad through multiple criteria. For now, you should just be aware that many EAs exist on MQL5 and you can grab one right now for testing in the coming sections. Another popular source of these EAs is Myfxbook that we highlighted some time back as an excellent source. Here too, there are many options you can get, both paid and free. (These are the: Best tips on working with Myfxbook)
If you’re using a different trading platform other than MetaTrader, you probably won’t find an EA on these aforementioned sources. cTrader provide EAs on their websites as cBots that you can download and then side-load using cAlgo. The process is different, but the EAs work just the same way. Although cTrader is not as popular as MetaTrader, the platform has been receiving a lot of attention and there is no shortage of EAs to choose from. (If you’re using it, then you ought to know: How to trade with cTrader)
Other trading platforms may even be more difficult to get EAs for, but then you can always hire a private developer. There are many freelance work websites from which you can hire a freelancer to create your own EA for a price. That may be the only option if you’re using a unique trading platform, yet it can also be the best way to go assuming you find the right guy for the job. Furthermore, keeping an EA effective over time needs tweaking and your developer can always be useful for updating and refining your EA as the major investors do by hiring software engineers. (Learn to use the: Ichimoku trading techniques for the FX arena)
Anyway, however you acquire your EA, the most important thing is to choose carefully based on the criteria we shall be exploring in the next section. First, though, some tips on how you can set up an EA to begin using it. (Are you aware: ESMA Finally Puts Its Foot Down On MiFID II Regulations)
Setting up your EA
Once you have got your EA, setting it up should not be a problem because we already showed you how you can set one up in MetaTrader. For the other trading platforms, you may have to find instructions from the developer themselves. Once you’re done, it is essential to use a VPS with your EA. MetaTrader 4 has a particular trading session period beyond which the EA could be disconnected from the server and require re-authentication with username and password. Additionally, your connection and execution speeds may not be suitable for using an EA especially if you’re using a scalping EA. This is high-speed trading and requires instantaneous execution where even seconds’ delays can cost you dearly. To avoid all this, having a VPS can ensure that your EA keeps running fulltime and even more effectively than from a personal computer. (You should know: Why Traders May Need To Use a VPS Service)
How to choose an EA for yourself
Before you have found an EA and set it up, the most important thing you should do is to analyse its performance and determine how good it is. If you manage to do this right, then you should have less problems and losses from underwhelming EAs. To determine if an EA is really effective, there are 4 main criteria you will need to consider about it. Whichever source you’re using for the EA, all of this data should be given to you as it is the basis from which you choose one. And in case you find that some of this information is missing, then think twice about why that is and why the developer is hesitant to give this information up. (Why you should: Think Twice When Making A Deposit In A FX Company)
Whenever you look at an EA’s performance, the most cited factor about it is the drawdown, but do you know exactly what it is? Simply put, it shows the percentage decline in the account balance from the previous peak. In this regard, there is both a maximum and average drawdown value. (FX Rigging And Manipulation: How The Major Investors Pull It Off)
- Maximum drawdown - this is the largest decline in an account’s balance from the most recent peak. Assuming you found an EA that had a maximum drawdown of 20%; this would mean that, at some point, the EA lost 20% of the account balance, but never more. This is just the worst case scenario you can expect from the EA when you start to use it. Thus, if you start with a balance of $10,000 and after a few days or weeks have just $8,000 left in your account, you can still consider everything to be ‘normal’ before you panic and shut it down (Have you made any of the: 10 most common mistakes FX traders make)
- Average drawdown – no one ever made consistent profits all the time, and neither do robots, so they too will have several losses. The average drawdown takes every period of loss where there was a drawdown and calculates an average. The result should inform you of the ‘normal’ amounts of losses you can expect throughout the use of the EA (Learn: How to protect yourself from margin call)
- Drawdown recovery – apart from the negative, this data tells you how long it took the robot to get out of its ‘funk’ and get back to making profits (Finally: Putting an end to broker bashing)
The idea behind understanding drawdown is to establish how much of a loss you can expect when using the EA at any given time. Most importantly, this shows you how volatile the EA is, but there is also a psychological effect involved. Volatile EAs make rather rash decisions based on the parameters they observe. For example, a volatile EA may be very sensitive to price changes that it makes many trades that end up in losses because of false signals. Such volatile programs may register excellent returns, but they will also often have large drawdown values, meaning that they are much riskier to use as they pose the danger of huge unexpected losses even from a few bad trades. On the contrary, a more reserved program will have only minimal drawdown values. Even though the account’s growth rate may not be exponential, such EAs are safer because there are little chances of blowing the entire account on a few trades. (An entrepreneur must know: How to create their own FX brokerage firm)
The drawdown values are expressed as a percentage, with average maximum drawdown amounts being between 15% and 20%. Any more than that can be considered to be volatile and risky. Keep in mind that risk management tips advise that you should never risk more than 2% of your capital on a single trade and not more than 5% on multiple trades, so even such a maximum drawdown is not the best. All the same, as long as the average drawdown remains below 10%, that could still be a safe bet. A short drawdown recovery period is also excellent, but you should be wary of abnormally fast recovery. Often, a very short recovery period means that the program is volatile and is making rash decisions that helped it recover quickly. Things sometimes go exceedingly bad beyond the maximum drawdown sometimes, and that is how people wake up in the morning to find their accounts wiped out. (Some of the: Most common questions FX traders ask)
People usually opt for trading robots and EAs because they don’t experience emotions as we do. Can you imagine losing 10% or even 20% of your capital and still continue trading normally? Of course you wouldn’t because you would begin to fear for your entire account and make mistakes. Knowing the drawdown values beforehand prepares you psychologically for this so that you select the EA that suits the amount of a hit you can take before hitting the panic button. This is why knowing the drawdown is so important. (Some: Simple Day Trading Strategies You Can Use In The FX Arena)
There is a lot more to drawdown than meets the eye too, which is why you should really be critical when studying it. Below are 4 tables representing different EAs. All of them had a maximum drawdown of $10,000 and a profit of $50,000 in the end, but notice the trend:
Here, the EA experienced dwindling profits over the years and increasing drawdown. This is an indication that perhaps it was getting out of date toward the end and probably needed be refined.
This is the typical volatile EA that received all its profits in a single year and lost in the other 4 years. As you can see, such an EA is not reliable and would not be recommended especially for someone saving for the long term. (Some things to consider when: Investing for the future)
Now this is the kind of EA you should be looking for because not only did the drawdown continue to decrease, but the largest drawdown also coincided with the period of least activity, 2014. This showed that, even though there were losses that year, the EA still minimized trading activity to reduce the losses.
In this example, the values have been inversed from the first table. Although it looks good, it may not be the best EA either. However, if you’re still interested in such a program, it is crucial to ask yourself why it performed so badly initially and what made the program better over time. (You should know: How to complain against a broker)
Average profit to expect per trade
Isn’t it nice knowing how much money you’re going to be making, that way you can start to plan on the day you finally buy that Lamborghini. The average profit to expect is calculated by subtracting the average loss per trade from the average profit per trade. Of course, the value you get does not accurately predict your future earning, but it is merely a benchmark of how profitable the EA can be. There is something more to consider when it comes to expectancy, though, and that concerns other factors not usually mentioned by software developers. (Know more about: Looking for arena correction)
Take the case of an EA with a $3 average profit per trade. That sounds great and it would be if there were hundreds of trades per day, but there are several things that may not have been taken into account like spread, slippage, commission, etc. It doesn’t mean that the developer is cheating you, but maybe they tested it on an ideal arena situation. In actual arena situations, spreads change with volatility and sometimes slippage occurs, then the broker may charge a commission if they are the most popular FX ECN brokers. Once these are taken into account, that average profit could end up in the negative meaning that it’s a losing EA. That is why it is recommended to go for EAs with a high average profit above $5 and even as much as $10 expectancy just to cover those situations when arenas are choppy. (All about: Using trendlines in your FX trading strategy)
The percentage of winning trades
This is also another very important factor that is used even by professional traders. Unlike what many people would think, professional traders actually don’t have an astronomical percentage of wins, and it is usually under 50%. The trick is to make sure the profits run all the way and cut losing trades early, that way the profits always outweigh the losses and make up for it. To calculate the same for an EA, take the ratio of the average size of a winning trade against the average amount of a losing trade. Just like the professional traders, an EA with a percentage win of 50% or below is good when it is a trend trading system because the profits are usually large to make up for the losses. (Here is how to use the: Trend rider V3 FX trading strategy)
When it comes to scalping EAs, on the other hand, always try to shoot for EAs with a high percentage of wins between 60% and 70%. The problem is that it is difficult to find such an EA and to still have it stable and not volatile. Therefore, just try to balance the percentage of wins with the drawdown value to make sure the percentage is not just high because of unnecessary risks. (All scalpers should master these: 10 rules of how to earn money with scalping)
The risk-reward ratio
Every time we mentioned volatility previously on this post, this is what we were referring to – the level of risk taken to earn the reward. A basic rule for investors is never to risk money unless there is the possibility of at least a 2:1 risk-reward ratio. Assuming an EA uses a 10-pip take profit and a 5-pip stop loss target, this would be a 2:1 risk-reward ratio. However, many EAs try to achieve a much higher ratio and that tends to complicate things. Let’s say another EA has a 10:1 risk-reward ratio, then that would sound great until you consider that it would need to have more than a 90% success rate in order to be profitable. Otherwise, it would just keep hitting stop losses and losing money. To ensure you pick the most reasonable risk-reward ratio, look at the percentage of winning trades mentioned earlier to determine the best ratio. (You could be: Profiting From Carry Trade Candidates)
How do you tell a good EA and one that is out-of-date?
Even when you know an expert advisor ticks all the boxes for you based on the criteria above, you have to confirm that the results still work presently and that the data is not out of date. To do so, you have to perform some tests on your own to determine how good the EA is. (Ask yourself: Should You Invest In CFDs Or Stocks To Make More Money?)
This is the most common test conducted on EAs to confirm that it works well and profitably. These tests are done by taking historical data and allowing the EA to run using the pre-determined parameters over a specific period in the past. If, say, the EA is triggered by crosses over and under the moving average, then each time the arenas crossed a moving average a virtual trade would be initiated. At the end of the test, you can see the potential profits had the EA been running during that time. It is often this data that is used to calculate all the data sets mentioned in the previous section. The good news is that many trading platforms allow for back testing including MetaTrader 5, which is better than its predecessor at this. (Is It Time To Upgrade To Metatrader 5: Features Of MT5)
To perform proper and effective back tests, always make sure to test different arenas, currency pairs and timeframes to ensure the results are accurate consistently. Also go for the most detailed tests that take every tick into account instead of the candle-by-candle tests. A lot of activity occurs in between the candlesticks that could trigger false breakouts. Not taking these into account could make you think an EA was excellent when actually there could have been more losses that were ignored. Great EAs are those with many parameters taken into account. For example, a good EA may be one that uses pivot points, moving averages and MACD before initiating a trade. (This is: All you need to know about pivot points)
However, back testing is not always effective because of two main reasons:
- arena conditions change constantly and data from 5 years back does not reflect what will happen in the future
- back tests don’t account for news announcements
Therefore, forward testing is done to eliminate these weaknesses. There are two ways of doing it, the first being a bit less cumbersome. Perform back tests, but on different time periods. For example, you could do one back test for the year 2016 and another for 2017, then compare how the EA did in both years. If the results are inconsistent, then you know that the EA is not well optimized and just faced better arena conditions in one year. Such an EA could also be volatile and not reliable. (Do you know: How Is Spread Betting Different From FX Trading?)
The second more effective method but also tiresome is to perform the test on current arenas. This takes a lot of time because you have to run the EA for days or even weeks to determine how effective it is. Even so, this is the best possible test you can conduct on an EA since you get current results with current arena prices. The main idea behind forward testing is to ensure that the results you got from the software developer are really consistent with current arena situations. Nevertheless, many people forgo this test because it is tiresome and just go by the data provided. That is not a good way to protect your investment, hence why so many have been burned. (Concepts Every Trader Should Understand: Leverage, Margin And Hedging)
When do you know it doesn’t work?
Both of the above tests rely mostly on demo accounts because they are safe. After all, you wouldn’t want to risk your money on an EA you just acquired online, would you? Nonetheless, you can’t really take a demo account seriously because there is nothing at stake hence no psychological pressure. While we would never advise anyone to risk their money on a real account to test an EA, it would be much better to test it on a cent or micro account or just start with a small deposit first. That way, you can have some pressure that pushes you to pay keen attention to the EA and its performance. (Some of the: Differences between demo and real accounts)
On top of that, FX brokers have a tendency to make trading conditions on demo accounts very nice. Spreads are very tight and there is no slippage because there is no actual volatility. Unfortunately, all these could be disastrous for the testing of an EA. This is why we advise that you use a real account with a small account balance instead. (Pluses and minuses of trading advisors)
Now it’s finally time for you to choose an EA that you will feel comfortable working with and also one that actually generates money. If you follow all the tips mentioned in this article, then there is no way you will suffer from a losing EA like many have before you. And when you do find it, remember to always keep tweaking it to adjust the parameters otherwise it will stop becoming effective. Who knows, do it long enough and you could become a developer yourself. Passive income, eh!
You must be curious to know what the arena experts think, so just watch this short video to know if robots are better than people: