I Tested Letting Claude Trade For A Month and Made $102k

How I Made $100,000 in a Month Using Claude

Overview of the Trading Challenge

• The speaker shares their experience using Claude as a personal trader, achieving $100,000 in profits within May.

• The initial investment was approximately $66,000, which grew to around $169,000 by month-end, resulting in 155% gains.

• Claude managed the portfolio daily, monitoring trades and making adjustments based on market conditions.

Approach to Trading with Claude

• Unlike previous automated systems built with deterministic rules, this approach utilized a more user-friendly platform (Robinhood).

• The goal was to demonstrate how non-technical traders could integrate AI into their trading strategies without complex coding.

• Claude acted as both a screener and portfolio manager while keeping human oversight involved.

Strategy Development

• The speaker outlines their background in math and economics from UCLA and experience in investment banking.

• They explain that the strategy for May was crafted by Claude based on specific constraints like account size and risk tolerance.

Research Phase with Claude

• The challenge combined deterministic (code-based systems) and non-deterministic (analytical responses from Claude).

• Initial research relied heavily on qualitative analysis rather than back-testable algorithms.

Key Insights from Strategy Implementation

• Context is crucial; specific directives provided better outputs from Claude compared to generic queries.

• Cost considerations led to using the web UI of Claude initially instead of API usage for affordability.

Crafting an Effective Trading Strategy

Defining the Strategy Parameters

• Before identifying tickers, the speaker had Claude outline a strategy focused on balancing risk and upside potential.

Options Trading Focus

• The primary strategy involved LEAP options—long-term call options designed to capitalize on stock price increases over time.

Market Conditions Consideration

• Smaller market cap companies were targeted due to lower liquidity issues affecting option pricing.

Identifying Trade Candidates

Screening Process with Claude

• Criteria for selecting stocks included those that had recently sold off but had catalysts for recovery.

Analysis Depth

• Each candidate was scored based on various factors such as catalyst timing and implied volatility environment.

Finalizing Trades

Selecting Specific Contracts

• For each trade identified by Claude, specific strike prices and expiration dates were chosen for optimal leverage management.

Building an Automated Monitoring System

Portfolio Dashboard Creation

• A dashboard was created using Claude to track performance metrics like net asset values and alerts for target hits or risks.

Features of the Dashboard

• It includes theta decay tracking for positions affected by time decay alongside general news updates relevant to holdings.

Developing Layered Monitoring System Prompts

Four-Layer System Design

• Each layer builds upon the last: data valuation, portfolio analytics, market context/news analysis, followed by alerts/dashboards.

Daily Operations

• This system runs daily checks on positions while providing insights into necessary adjustments based on market changes.

Conclusion

• The video emphasizes combining deterministic coding with non-deterministic analysis through AI tools like Claude for effective trading strategies.