I have been in the currency market for several years, mainly engaged in EA development and trading. Let me talk about my views and understanding of EA, including how to develop a good EA and how to evaluate an EA. I hope it will be of some help to EA fans and EA users.
EA is roughly divided into 5 segments:
1. Positive expectations cannot be met, that is to say, they cannot stand the test of historical data. It can be said that 80% of the EAs on the currency market belong to this lowest level of EA.
Specific points include:
a. After long-term testing, the Martin strategy directly broke the position.
b. After long-term testing, the Martin strategy has a stable loss with a stop loss.
c. If the gap between the test and the firm offer is too large, the test is perfect and the firm offer is stable. Mainly manifested as breakthrough scalp, early scalp, and other ultra-short-term trading strategies.
d. Periodic profit, long-term loss or loss, represented by Martin and most of the indicator EA.
e. Special market filtering, over-optimized EA, the test is very good, but it will lose money once it is firm.
f. The EA with direct data fraud writes the historical trend into the EA to make orders, and there is no trading model and operation logic. Usually, the file size of this type of EA is more than 400KB. The test is perfect, and the real offer is completely inconsistent with the test. The author often updates the EA.
2. An EA that has achieved positive expectations but has poor stability and yield. About 15% are this kind of EA.
This kind of EA has found a little rule of market movement, and the EA builds a model based on this rule. During the long-term test process, it can achieve profitability, but the stability and yield are poor.
The momentum system and the morning scalp system are typical examples, and these systems are relatively common on MQL5.
The reason for the poor stability is that the developers do not handle the details of the regular application well, such as using simple moving stop loss, fixed stop loss and take profit, and being unable to follow the market to make effective changes, which may lead to poor stability of the strategy Obvious fluctuation period.
The reason for the poor yield is mainly linked to the frequency and effectiveness of regular exercise. Any signal has a probability, and this probability determines how large a position investors can use.
For example, when the winning rate is 100%, you can invest all your capital in this signal, that is to say, work with a full position, while for a signal with a 50% winning rate and a profit-loss ratio of 1:1, you can only invest a small amount of money at a time. game.
Frequency is also very important. If the frequency is too low, rapid growth cannot be achieved.
3. Positive expectations, good stability and low returns. About 4% are this kind of EA.
This kind of EA handles the law of market movement in a more detailed manner. Select a signal with good stability to do it, and at the same time control the details of position management and entry and exit methods, so that the periodic retracement of EA is effectively controlled. The most ideal representative is the EA of SFE Price Action. Of course, some signals are also the path of this momentum system, which can be made more stable after processing. The main factor limiting the rate of return of EA is the signal frequency. The market has various movement laws, but the laws discovered by people will gradually weaken. Now it is extremely difficult to find a new quantifiable movement law of.
4. Positive expectations, good stability and high yield. It is extremely rare for such EA to be listed on the market, less than 1%.
The stability we are talking about is at least based on historical data of more than 5 years, transaction records of more than 3,000 orders, the retracement will not exceed 30%, and the annual return can reach more than 50%. There is a specific operating logic, and the real offer is highly consistent with the test The EA.
We can see that there are very few EAs that can meet such conditions. Needless to say, Martin’s positive expectation is a problem. Put a stop loss and use every real-time price test, no special filter, and no EA fraud. was passed out. Due to the enlargement of the point difference in the early trading scalp system, the firm offer and the test will be far apart, and most of them are less stable. Breaking through the scalp and ultra-short-term (holding time of one minute or even one or two seconds), the platform slippage will be unbearable, simulation can make money, and the real offer will be defeated.
Many EAs cannot break through to the third level. This is determined by the laws of motion. Only by discovering new and efficient laws can they break through and enter the fourth level.
5. Finally, let’s talk about the EA on the fifth floor, which is the high-frequency EA.
It can be said that ordinary people can't see it at all. The nuclear weapons in the trading tools are not restricted by time or capital. They can achieve sudden wealth through rapid compound interest with small funds in a short period of time. Some people will not disclose any details.
This kind of thing is not without it. You can use your browser to search for news that Russian traders have earned 3 billion with hundreds of thousands of high-frequency trading. This market really exists. There is also a story about a Taiwanese rookie circulating in the currency world, who said he made hundreds of millions through high-frequency trading. Of course, the legend is not credible and has no practical significance. Anyone who has been around for a while knows that there are people who do delay arbitrage. On a small platform, arbitrage trading through platform loopholes is actually a high-frequency trading. It has a high accuracy rate and a very short position, but it is real and effective and can be realized quickly. Getting rich, of course, the platform is not a fool, it will restrict these people from doing so. Therefore, as a technical researcher, we still need to pursue high frequency at the transaction level.
The above is my general classification of EA ratings. Next, I will talk about my understanding and suggestions on trading and development.
One way is to learn the existing EA in the currency market, and the other way is to explore market laws independently.
Let’s talk about traders first. Most traders actually go the second way, which is to explore the system by themselves. This path is almost difficult to navigate, the specific reasons are as follows:
a. No statistical thinking, no statistical tools. Most traders make orders based on their impressions, and the market changes relatively quickly, so it is easy to have periodic profits and periodic losses, and it is difficult to form an effective model in their minds.
b. Limited executive ability, affected by emotions and time constraints. After a loss, it is completely negated, and the profit is over-inflated. It is impossible to keep an eye on the market in time when entering and exiting the market, and opportunities are missed, which leads to the inability to establish effective rules.
c. Decision-making management is not scientific enough. Seemingly simple trading, in fact, contains some iron-clad rules. If it cannot be handled and managed scientifically, even if it is right, it is wrong. A simple example is a heavy position. The position exceeds the capacity of the system, and the profit-making strategy will turn into a loss. , the Kelly formula that every trader should know.
Let's talk about the first road. I believe that most of the friends who do strategy development will take the first road, that is, the idea of learning first. By learning the existing strategies in the market, they will first realize profits and live first in this market. Come down and talk.
Learning strategies need to use:
a. Will program. This can help us have an advantage in statistics and analysis.
b. Trading experience. The more details you know, the easier it is to understand the data and understand the principles. Some strategies are related to time, some are related to price, whether the strategy is volatile or unilateral, these things are all experience, if one does not understand, one may always be confused.
c. Industry Convergence. In-depth integration of the market allows us to obtain more resource information from the market. If there is no research source, we will not be able to get in touch with more excellent ideas, and naturally progress will be slow. If you know a certain great god, you can learn real things from him, which may be better than studying by yourself for 5 or even 10 years.
The above is my understanding, and there may be one-sidedness and deficiencies, but the original intention is to help everyone.
Add
