👉 **Learn AI trading bots and algo trading with us:**
Join the DaviddTech community: https://www.skool.com/daviddtech/about
I gave ChatGPT, Claude and Grok the same starting prompts and one week to build the most profitable trading strategies they could.
But finding promising backtests was only part of the challenge. Some agents stopped early, usage limits interrupted the experiment, and the leaderboard winner wasn’t necessarily my first choice for everyday use.
In this full step-by-step guide, I show you how to build your own AI trading desk, connect the backtesting tools, create a strategy dashboard and run the same experiment yourself.
📊 Strategy Incubation Tracker: https://docs.google.com/spreadsheets/d/1sMSKL-uul9R1QyIcDWV7SiO35JJQYuuaw9h3CWb51i4/edit?usp=sharing
🏆 **AI Trading Arena & Setup Prompts:**
Access the Arena and prompts: https://aitradingarena.com/
🛠️ **Free Backtesting Tools:**
TradingKit.com: https://tradingkit.com/
🔑 Watch for the password that appears on screen to access the prompts shown in the video.
—
## In this video you’ll discover:
✅ How to set up an AI trading desk with ChatGPT, Claude or Grok
✅ How to connect TradingKit’s MCP tools and run your first strategy backtest
✅ How to build a dashboard for returns, drawdown, trade history and strategy comparisons
✅ The dashboard, onboarding and one-week experiment prompts I use
✅ What happened when agents stopped working or hit their usage limits
✅ How the models compared on strategy results and reported token usage
✅ How to export a strategy into TradingView’s Pine Script editor
✅ Why out-of-sample checks, forward testing and buy-and-hold comparisons matter before risking real money
You can follow along with **one AI agent**—you don’t need subscriptions to all three. TradingKit is free; your chosen AI subscription and TradingView access may involve costs.
—
## ⏰ Video Chapters
0: 00 – Three AI Models, One Week: The Challenge
0: 30 – What We’re Building & the Roadmap
1: 33 – Inside My AI Trading Arena
2: 12 – The AI Models & Tools You Need
4: 22 – Connecting the Backtesting Tools
6: 07 – Testing the Connection With a BTC Strategy
6: 54 – Building Your AI Trading Dashboard
7: 59 – Onboarding Your AI Agents
9: 00 – Launching the One-Week Experiment
10: 23 – Midweek Problems: Persistence & Usage Limits
12: 48 – One-Week Results: What the Agents Built
14: 43 – Token Reporting & the Strategy Leaderboard
16: 23 – Exporting Strategies Into TradingView
17: 24 – My Verdict: Strategy Quality vs Value for Money
18: 17 – Before Going Live: Testing, Risk & Buy and Hold
19: 14 – Strategy Incubation Tracker & Community
—
## Before You Trade These Strategies
The one-week challenge is about **building and evaluating strategies**. Historical backtest returns are not one week of live trading profits.
I explain why I put strategies through at least three months of simulated live trading before considering real capital. I also look at profit factor, trade count, drawdown and performance against buying and holding the asset.
The token figures shown in the leaderboard are model-reported estimates, not independently verified billing data. The agents also needed intervention during the experiment.
⚠️ This video is for education, not financial advice. Backtested results do not guarantee future performance. Every strategy needs further validation and appropriate risk controls.
—
## 💡 Tools Featured
ChatGPT / GPT-6 Astra
Claude / Fable 5.1
Grok Bot
AI Trading Arena
TradingKit.com
TradingView
Pine Script
MCP Backtesting Tools
If you’re building AI trading bots, testing TradingView strategies or learning algorithmic trading, subscribe for more practical experiments and step-by-step guides.
Test the ideas.
Check the results.
Build a system you understand.
#AITrading #TradingBots #ChatGPT #Claude #Grok #TradingView #Backtesting #AlgorithmicTrading #DaviddTech