Neural networks on Forex
Neural networks on Forex - do they make sense?
In today 's world, artificial intelligence, self-learning voice assistants and Big Data analysis are penetrating our lives. Ordinary people face this most often in smartphone application format, but the fact remains - neural networks, or self-learning computer systems are already everywhere, even if we don 't see them.
For many years, since 2006, programmers have tried to introduce neural networks into trading as well. The idea seems interesting - to adjust trade automatically to the constantly changing market. But how are things in practice?
Many of us are constantly looking for new strategies and tactics to trade in the Forex market. Each system found undergoes testing on a historical period of different lengths. Ideally, testing should identify patterns that work over a sufficiently long period of time.
In reality, this is an unresolved task - trading systems "drain," having worked from three months to several years. Extending "lifespan" helps optimization, but ultimately you have to look for a different approach to the Forex market.
The phenomenon of the failure of Forex strategies is given many explanations, but it is worth paying special attention to one of the reasons that can over time nullify all efforts to earn on currency speculation - the evolution of neural networks.
The Forex market of the nineties literally "handed out" money to the first bidders to install trading terminals, connecting to the Internet, and using quite simple tactics described in the books of traders of the eighties. Earnings were not hampered even by temporary lags, platform bugs, large spreads, low speed of connection to the World Wide Web.
The neural network will change currency speculation, brokers can return to tactics similar to "kitchens" only on a global scale. In fact, the limitless capabilities of neural networks can be applied against the crowd, predicting not the exchange rates, but the pattern of behavior of each individual trader. Market makers and prime brokers will be able to pick counter-strategies, hunting for footsteps, expanding spreads at the time of market applications, putting phantom volumes in glasses ahead, not in fact.
Designed and launched by the startup Sentinent Technologies, the neural network can already emulate 1,800 working sessions, predicting with high accuracy to trillion (!) cognitive patterns of behavior of real traders. The system trained on flows of warrants taken from a book of exchange applications and broker servers.
Quality and quantity of data - the guarantee of successful training of neural networks; Archives of teak transactions broken down into specific accounts are the most running goods in the data-mining market. This term refers to a separate industry extracting, analyzing, and formatting primary input information for a neural network.
Even if brokers fail to study the traders "pattern of behavior and successfully play against crowd strategies, they will create high-end prognostic systems that are no longer repeated in trading terminals. Modern trading systems operating in the markets of shares, goods and currencies read and understand news, recognize patterns, that is, represent an analyst with the brain of a supercomputer.