I built a trading bot with Claude that made +$168,236.

It's a complete cheat code, and the entire strategy was built using Fable 5.
No dev experience required.
Here's exactly how to build & backtest profitable trading strategies using AI (nfa):
Step 1. Extract your strategy rules
Start with any trading strategy you want to test, and drop this prompt into Fable 5:
"Take the classic [strategy name] and turn it into a fully objective rule set on the [insert timeframe].
Give me: exact entry rule, exact exit rule, stop logic, and position sizing.
No discretion - every rule must be computable."
This could be an RSI strategy, EMA strategy, etc.
The goal here is to get a clean set of rules you can actually backtest.
Step 2. Convert to Pine Script
Once you have your rules, paste this follow-up:
"Now write this as TradingView Pine Script v6 strategy code.
Include a 0.1% commission per side, and code ready to paste into the Pine Editor.
Copy the entire code block it gives you.
Step 3. Import into TradingView
Open TradingView → Pine Editor → New Script. Paste the code and click Add to Chart.
You'll instantly see every historical entry, exit, win, loss, and drawdown.
If you hit errors, screenshot them, paste back into Fable, ask for a fix.
Step 4. Refine the strategy
This is where the real alpha lives.
Download the CSV file from your TradingView backtest results. Import it back into Fable and prompt:
"Analyse this trade history and tell me exactly how to improve this strategy."
Fable will identify patterns in your losing trades, suggest rule adjustments, and help you rebuild a stronger version.
Step 5. Automate it
Connect Claude Code to your exchange via its MCP server. Generate an API key from your exchange and hand it to your agent.
Now the loop closes: Fable backtests the strategy, you validate the results, and the system watches the market and executes automatically whenever your conditions are met.
Don't want full automation? You can just set it up to flag you whenever a trade opportunity meets your parameters.