Discussing the article: "Self-Learning Expert Advisor with a Neural Network Based on a Markov State-Transition Matrix"
A new word in the advertising of grid advisors!
I have never read such a sweet text on this portal.
What is missing is a comparison of the results of the EA with and without NS.
After that.
Forum on trading, automated trading systems and testing of trading strategies
fxsaber, 2025.03.25 08:52
As if the text was written by AI: so much water.I started to spend much less than a minute on some articles - I scroll through and read a brief conclusion with a final picture/video. If it is not interesting there, then the same attitude to the whole article is formed, unfortunately.
Because of AI texts I almost stopped reading long posts on the forum. Exception - a few authors I've known for a long time on the forum.
Conscious or unconscious, but blatant manipulation of test results. (many authors suffer from this).
The Expert Advisor was tested with a fixed lot, which completely kills any strategy - as the conditions of each next trade of the Expert Advisor become less risky. Hence the low drawdown percentage. For such a tester picture it is not necessary to make matrices, AI and so on, it is enough to find a convenient time point for the test.
It seems to me that an Expert Advisor (not only this one) should be tested at a lot size determined by the deposit size (percentage). Then each trade in the test will be like the first one. In fact, the conditions of each deal will always be the same in terms of risk. And here the picture will be completely different.
Strange, but the file from the box is already compiled with an error (DeInit )))). It is not clear at what settings it was tested - from the same "box" there are cosmic numbers. And if you remove the AI water, then in the end there is nothing to read. You can be more specific.
By the way, fill in the AI text"The current model uses mostly ATR to determine market conditions. But imagine what depth of understanding we will achieve by adding to this orchestra the sound of RSI, the melody of MACD, harmonic sequences of Fibonacci levels and rhythmic structures of volumes!". He'll give you such a story!!!!)))))
On the forum only sources can be posted, otherwise they can be banned.
Actually, what is the effect of additives?
Thefunny thing is that the Expert Advisor does not work on the tester. I don't understand what and how the author tested it. I put it on demo for all pairs, I will see what and how
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
Check out the new article: Self-Learning Expert Advisor with a Neural Network Based on a Markov State-Transition Matrix.
Imagine not just a program that executes an embedded algorithm, but a digital organism that continuously evolves, adapts, and, in a sense, understands the complex symphony of market movements. This article is devoted to such a system — a new-generation trading EA.
The key to this breakthrough lies at the intersection of three different fields of knowledge: the probabilistic mathematics of Markov processes, the intuitive power of neural networks, and the practical logic of hedging strategies. When these three forces come together, something greater than the sum of its parts is born — a qualitatively new system emerges, capable of thriving in the volatile and unpredictable environment of financial markets.
The results of our experiments speak for themselves: an average annual return of 28.7% with a maximum drawdown of only 14.2%, a Sharpe ratio of 1.65, and 62.3% of profitable trades. But behind these dry numbers lies a much more significant achievement: a system that performs with equal confidence in the quiet harbor of sideways movements and in the storm of high volatility.
Author: Yevgeniy Koshtenko