These are some visualisation of my trades for PE-SELL-ETH
I am no computer geeks ; in fact my major is mechnical eng
The only thing I can do is to clean the data
Any guide/tips on how to do the same ?
In my opinion, instead of trying to make an AI model replicate your strategy based purely on your past trade data, you may find it much more effective to provide the AI/ML model with the exact logic and guidelines of your strategy. This tells the AI why you made the decision, not just what the result was.
You can then ask the AI to create the trading strategy code using your concerned broker’s API to automate your trades. This approach is generally more robust and directly executable.
Here are a few important pointers to note:
-
Translate Your Logic: The AI needs clear, specific rules. Think about converting your strategy into simple “IF X, THEN Y” statements (e.g., IF volatility is low AND the price is at resistance, THEN Sell). The cleaner your rules, the better the output code will be.
-
AI is a Tool, Not a Trustee: Always remember that AIs are still non-deterministic. Instead of blindly trusting the AI’s output, you should leverage it as a powerful assistant. Always review the code it shares to ensure it perfectly matches your strategy, risk limits, and logic.
-
Backtesting is Non-Negotiable: Before ever executing the AI-generated code in the live market, you must thoroughly backtest the strategy using your historical data. This step confirms that the automated logic performs safely and profitably.
1 Like
This is past data only , 27 entry only trades to be precise on PE ETH for OCT month
I don’t to hard code rules , as i until the second I take trade I don’t if I would enter or not. It should learn when I take trade , Definitely there is a pattern , I would guide it to take trades based on that pattern
ok 
Never backtested any of my data in any asset class. But would test it in live market.
Would accept or reject trade entry so that it can learn from it.
Reinforced learning ig