Have you ever heard about trading robots? Well, this is a kind of program for trading in the financial markets. You install it on a trading terminal and then drink a cocktail on the beach while that trading software sells and brings a steady income. Sounds like a complete scam, doesn’t it? Or not? Read more in our blog.
Trading robots are programs able to analyze the market and signal when there is a favorable situation for buying or selling an asset. Some robots are even trained to make decisions and open deals independently – thus, the trading process becomes fully automated. When such programs appeared on the open market, they were in massive demand.
However, quite soon, the enthusiasm of buyers gave way to negativity – robots very quickly stopped working “as they should” and only brought losses. So what’s the problem with such programs? Are they really only meant to be sold for big bucks? Or are they just not exactly the kind of robots that can generate income? We discussed this issue with Ilya Matantsev, our Business Analyst, who has first-hand experience in investigating programs for trading in the stock markets.
Why Forex robots don’t work
To begin with, we want to clarify: we are not claiming that none of the existing Forex robots work. We just don’t know any examples of a truly working robot that does bring a stable income regardless of the market conditions. There are programs that can be effective here and now, in the current market situation. But the market is constantly changing, and in a month or two, any robot stops being profitable. And guessing the moment when the program will lose its effectiveness is no easier than understanding when an asset will rise or when it will fall. So what’s the point of such “automation”?
There is one more nuance: robots being fiercely criticized on the Internet are almost always free. Sometimes these can be versions of paid advisors leaked by pirates, sometimes a program is originally placed in the public domain. It’s quite obvious that no one will just hand out the notorious trader’s “grail”, and those who believe that this is possible just lose money in the end. On the other hand, information about robots created in the strictest secrecy slips on to the Internet from time to time. Such robots are sold, if at all, for a lot of money, and a leaked version is impossible to find. These unique programs, which are developed for a particular client, are most often ordered by large companies specializing in trading exchange.
How fundamental analysis programs for the stock market work
One of the key features of programs designed specifically for large financial companies is that this trading software is an auxiliary tool and doesn’t make decisions for the trader. A digital advisor can process a huge amount of data and provide a trader with a short report, but opening or closing a trade is left for the human.
The program that was built with the participation of Ilya analyzes news related to companies whose shares are traded on the world’s major stock exchanges. The robot independently searches for information on specific resources. These are reputable aggregators like Reuters. The list of aggregators is formed according to the ratio of information flow speed to its reliability. This approach helps to minimize the likelihood of errors due to reacting to fake news. The program independently determines what influence this news is more likely to have on the company’s shares. Thus, an overall assessment of the asset’s sentiment is formed: if most of the news is positive, the shares are more likely to be bought, and their price will rise. Otherwise, a downward correction can be expected.
Ilya Matantsev, Business Analyst at Andersen:
The system searches for news by keywords that are associated with the company. Based on text recognition algorithms, the direction and degree of influence of this news on the company are determined. For example, it comes to light that Tesla’s self-driving car hit a person – the algorithm, firstly, immediately finds this news, and secondly, recognizes the negativity for the company in it. And a few seconds after the news appeared, the trader receives a signal that there is a new negative factor for Tesla’s shares. Thus, the trader has accurate information and some amount of time to make a decision, unlike other market participants who monitor the news feed “manually”.
On the other hand, we may receive news regarding the company, but so insignificant that, given other news, it is unlikely to affect the price. The program is able to analyze this as well and then give it an appropriate assessment. This is an example of a simple yet effective way to work with Big Data. We collect chaotic data, process it using Data Science, organize it, and get the result.
The news analysis software is also a kind of robot, but not one of those Forex robots that are advertised on some specific sites. Perhaps, over time, the level of AI will reach the point of making more optimal trading decisions than people make now. At a minimum, forecasts will become more accurate, and robots will be able to successfully predict not only the direction of movement of an asset but also the duration of this movement. But so far, the final decision is still up to the human. Clumsy algorithms that are not able to learn on their own are doomed to failure in a sphere where psychology and the human factor often turn out to be much more significant than mathematical analysis.