Discussing the article: "Employing Game Theory Approaches in Trading Algorithms"
Tsetlin Machine is also interesting for small data but less known: https: //github.com/cair/TsetlinMachine
https://www.literal-labs.ai/tsetlin-machines/ but I find it difficult to implement.
- cair
- github.com
Tsetlin Machine is also interesting for small data but less well known: https: //github.com/cair/TsetlinMachine
https://www.literal-labs.ai/tsetlin-machines/ but I find it difficult to implement.
Original thing, I am overflowing with delight as from an object of art, thank you :) But it is desirable to test it on real ticks, because it is shallow with deals.
Greetings, I am very interested in your project, but I am new to this field. I can't understand how to run the Expert Advisor in the strategy tester. As I understand it is impossible to fully configure and train it through the tester? Or am I doing something wrong? I would be grateful for the OS
Where I have relatives in the Netherlands from? 👀
Ahahahahah, not in the Netherlands)))) VPN is such a thing)))))
PS: bottom line in the strategy tester is it possible to run training or not? According to the balance chart screenshot it's a strategy tester, but whatever I do I don't even get close to + in it
The author presents “AdaptiveQ Enhanced,” a multi-symbol FX EA that claims to combine DQN, Nash equilibrium, causal analysis, seven major currency pairs, six actions, and 531,441 states. The action set includes buy, sell, add to buy, add to sell, and closing only profitable buys or sells.
My main problem: the article uses fancy labels more than real substance. Its “Nash equilibrium” is not an actual equilibrium solve; it is just taking a symbol’s Q-scores and nudging them with other symbols’ Q-scores weighted by rolling correlation when |corr| > 0.3 . That is not game theory in any serious sense. The same issue exists with the “causal” language: the cross-symbol update is literally based on reward times correlation when |corr| > 0.2 . Correlation is not causality.
The state design also looks shaky. The article says it builds a rich multi-timeframe state from prices, MA differences, RSI, stochastic, and MA flags, then hashes that multidimensional information and reduces it with hash % TOTAL_STATES . So different market situations can collapse into the same bucket. Calling that “531,441 unique states” sounds more impressive than it really is.
The position logic is the ugliest part. The EA can run in multi-position or opposite-position mode, add volume to existing positions, allow up to 5 positions per symbol, and selectively close only profitable positions while leaving losers open “to recover.” That is not smart inventory management to me; that is a dressed-up path to ugly exposure.
What is good: the engineering side is more serious than average MQL5 fluff. Caching indicator handles, periodic updates, and saving/loading the Q-matrix are practical implementation details.
So in my opinion an interesting experiment and weak trading design. Too much branding around “AI/game theory/causal inference,” not enough proof that it has a real edge. I would not build on this logic directly.
- 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: Employing Game Theory Approaches in Trading Algorithms.
In conditions where the speed of decision-making is critical and the market is characterized by a high degree of uncertainty, a different approach to creating trading systems is required. AdaptiveQ Enhanced is a trading Expert Advisor developed based on deep reinforcement learning (DQN) methods, game theory, and causal analysis.
The Expert Advisor analyzes the market by modeling 531,441 unique states, taking into account interrelationships between the seven major currency pairs. The key element of the algorithm is the Nash equilibrium, which is used to select the optimal strategy under conditions of mutual influence of symbols.
The article examines practical implementation of these approaches in MQL5 and demonstrates how the combination of adaptive learning, game theory and AI allows you to build more accurate and sustainable trading strategies.
Author: Yevgeniy Koshtenko