Discussing the article: "Neural Networks in Trading: Heterogeneity-Informed Meta-Parameter Learning (Conclusion)"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Check out the new article: Neural Networks in Trading: Heterogeneity-Informed Meta-Parameter Learning (Conclusion).
Training is organized into two complementary phases, which provide both a solid foundation and the flexibility needed to operate in real-world market conditions. In the first, offline phase, we conducted extensive training using historical data for the EURUSD pair on the H1 timeframe over all of 2024. This period covered the full range of market regimes—calm flat markets, sustained trends, sharp spikes in volatility, and periods of heightened noise—so it served as an excellent training ground for the model.
The second stage is online fine-tuning. The model processed the stream of candlesticks sequentially in the MetaTrader 5 Strategy Tester, in conditions as close as possible to real-world trading. This stage reveals properties that are entirely different from those of offline training: the ability to tolerate noise, respond appropriately to changes in liquidity, and account for delays and slippage. We carefully simulated real execution conditions so that the model's behavior would remain predictable when transferred to real market conditions.
The final and most rigorous validation stage was conducted on a completely external sample—quotes from January through March 2025. All model parameters remained frozen during this process. This type of testing provides an objective picture of practical effectiveness: the algorithm’s ability to maintain stability and predictability under new conditions.
The test results are shown below.
Author: Dmitriy Gizlyk