Bootcamp | Trading From Scratch

This is not a get-rich-quick track. Trading is a craft, and as with every craft, you can only get better by relentless practise. The bootcamp sets the foundation so you can build upon it and refine it through iterations.

Expect to invest 5 to 8 hours per week.

What you will learn

The bootcamp is structured into nine lessons that take you from first principles to a running strategy.

Breaking problems into questions — Every complex situation becomes manageable when you decompose it into the right questions. You start with three: what should you trade, when should you trade it, and how should you trade it.
Trading as a craft — Risk management, clear rules, and a focus on systems over results. You learn position sizing, stop loss and take profit calculations, and why protecting your downside is more important than chasing profits.
The journey to profitability — The three pillars of consistent trading: finding edge, rigorous validation, and trade execution. Each pillar is broken down into its prerequisites, giving you a clear roadmap for what to learn and in which order.
Leveraging AI — How to treat AI agents as human peers, manage expectations, and iterate on prompts. You learn when LLMs shine (reasoning, analysis, creative problem-solving) and when deterministic code is more efficient.
Magnus, the mentor — A personal coaching agent that helps you grow as a trader through the Socratic method. Magnus asks the hard questions, challenges your assumptions, and guides you to your own answers.
Arvid, the analyst — A technical analyst agent that helps you compose strategy definitions with clear, quantitative rules. You walk through a complete example of building a strength and compression breakout strategy from a vague idea to a fully specified rule set.
Quentin, the quant — A quantitative analyst agent that takes your strategy definition and assesses its robustness through backtests and simulations. You see how Quentin downloads data, writes scripts, runs analyses, and produces a verdict on whether a strategy is worth pursuing.
Tudor, the trader — An execution agent that trades your strategy on the Binance Testnet. You learn how to give an agent the skills it needs (API access, market data, position sizing) and how to run periodic trading sessions.
Manual trading exercise — A hands-on practise routine using TradingView. You learn to read the story of a chart, identify swing highs and lows, draw positions, and review your trades systematically.

Key insights

Your growth as a trader directly fuels the quality of your AI agents. You can only recognise quality in what they produce when you understand the craft yourself. Mastery is not optional; it is the prerequisite.
Risk management is non-negotiable. Risk only 1% of your available funds per trade. Set daily and weekly loss limits. Your funds are your oxygen. The less you have, the harder it gets.
Focus on process, not results. If you constantly refine your process, it is a question of time until you reach consistent profitability. View losing trades as the overhead cost of finding the winning ones.
Every trade needs clear, repeatable rules before entry. No trade without a defined setup, entry trigger, stop loss, and take profit. Clear rules reduce moving parts to a minimum and enable true refinement.
Micro strategies reduce risk and speed up learning. Short-term, clearly defined trades give you fast feedback loops and allow you to start from scratch whenever a trade ends. Profit is secondary at this stage; execution quality matters.
Treat AI agents as human peers. They are generalists with excellent reasoning. Your responsibility is to equip them with the necessary context and skills. The better you articulate yourself, the better they deliver.
Context is key when working with LLMs. The richer the context of your conversation, the better the output. An agent with the exact strategy, full assessment data, and historical prices in context produces dramatically better results than one with a generic prompt.
Volatility is opportunity, not a problem. The same moves that can cause losses can generate profits if you are on the right side. Recalibrate your approach for the environment you are trading in.
Verified metrics from real exchanges beat theoretical backtests. There is a gap between theory and practise. Slippage, delays, and execution imperfections mean your actual risk/reward ratio will differ from the theoretical one. Only confirmed metrics from production count as proof.
Break every complex problem into questions. The only limit you have is the quality of the questions you ask.

Key questions answered

How do I find a trading edge, and what does "edge" actually mean in measurable terms?
How do I define clear, quantitative rules for a strategy so a quant can assess them?
How do I assess whether a strategy is worth trading with real money using backtests and performance metrics?
How do I use AI agents to mentor me, analyse charts, compose strategy definitions, and execute trades?
How do I calculate position size, stop loss, and take profit based on my risk tolerance?
How do I read the story of a chart using swing highs, swing lows, and conflict zones?
What does the journey from beginner to consistently profitable trader look like, and what are the prerequisite chains?
How do I set up and interact with Claude Code agents, and how do I run them in parallel?
When should I use reasoning (LLMs) versus deterministic code (trading bots)?
How do I review my trades to learn the most from winners and losers?

What is not covered

The bootcamp sets the base. The following topics are intentionally left for later stages.

Full, sophisticated strategy assessments such as walk-forward validation and out-of-sample testing at scale
Trading with real money on Binance Mainnet — the bootcamp uses Testnet exclusively
Agent orchestration and autonomous agent-to-agent workflows
Building and maintaining a knowledge repository
Advanced agent definitions with skills, rules, and subagents
Secure, production-grade trade execution with proper isolation
Sophisticated backtesting infrastructure and automated pipelines
Order types beyond market orders (limit, OCO, trailing)
Fundamental analysis or news-based trading strategies
Multi-position portfolio management and correlation analysis

Who benefits the most

Aspiring traders who want to learn the craft from first principles with a structured, methodical approach instead of gambling on tips
People interested in AI-assisted trading who want to understand how to use Claude Code agents as a team of analysts, quants, and traders
Technically minded people comfortable with terminal tools and willing to set up agents, read code, and iterate on prompts
Traders with some experience who have been trading without a systematic process and want to build one from the ground up
Anyone willing to invest the time. This is not passive consumption. You get out what you put in, and the bootcamp demands deliberate, hands-on practise

If you are looking for a signal service, copy trading, or a bot that trades for you without understanding what it does, this is not the right fit.

Requirements

Claude Code installed on your computer (or a similar LLM tooling that supports agent workflows)
A Binance account — Testnet is sufficient. No real money is needed to complete the bootcamp
A TradingView account for charting and the manual trading exercise
5 to 8 hours per week to go through the lessons, practise, and interact with your agents
Willingness to practise and iterate — the material is designed to be applied, not consumed passively

No prior trading experience is required. Basic comfort with a terminal or command-line interface is helpful but not mandatory. If you can install software and navigate a file system, you are ready.