Former Bank Trader Explains How AI Is Changing Retail Trading

Former institutional trader Matteo Hudson says the gap between Wall Street professionals and retail traders is shrinking rapidly.

After trading approximately €15 billion annually at a European bank, Hudson explains how artificial intelligence, backtesting, and automated strategies are giving everyday traders access to methods previously associated with institutions.

How Institutional Traders Actually Trade

Hudson says institutional trading is very different from what most retail traders imagine.

Rather than relying on secret indicators or manually predicting market movements, professional trading desks focus heavily on research and automation.

According to Hudson:

• Approximately 98% of trading activity is automated.

• Institutional traders may reject 90–95% of their strategy ideas before trading them live.

• Professional desks often run hundreds of algorithms across different markets.

He says the biggest advantage institutions have isn't simply capital. It's the process they follow before risking money.

How AI Is Changing Retail Trading

Tools such as ChatGPT and Claude are making it easier for retail traders to develop trading algorithms without advanced programming knowledge.

Hudson explains that traders can now use AI to help:

1. Turn a trading idea into clear rules.

2. Write code for the strategy.

3. Backtest it using historical market data.

4. Analyze profitability and potential drawdowns.

5. Automate execution once the strategy has been validated.

He mentions MultiCharts, TradingView, NinjaTrader, and Python as possible tools for building and testing strategies.

However, AI-generated code still needs careful testing before being trusted with real money.

Why Backtesting Matters More Than Trading Live

Hudson believes one of the biggest retail trading mistakes is risking real money just to discover whether a strategy works.

Instead, traders should test their ideas first.

He recommends analyzing win rates, average gains and losses, maximum drawdowns, and expected performance.

A strategy that fails during testing can be rejected without losing trading capital.

Why Simple Strategies Often Work Better

Hudson warns against building overly complicated trading systems.

Every additional rule creates another potential failure point.

A trading strategy should have three essential components:

• Entry rules: When to open a position.

• Exit rules: When to close it.

• Position sizing: How much capital to risk.

The strategy should also have a logical reason for why it could be profitable.

Hudson emphasizes avoiding overfitting, where a strategy is optimized so heavily for historical data that it may fail in future markets.

The Future of Retail Trading

Hudson believes traders will increasingly adopt one of two approaches:

• Hybrid trading: Using AI and statistical testing while executing trades manually.

• Automated trading: Letting software execute validated strategies according to predefined rules.

Automation can reduce emotional decisions, save time, and allow traders to manage multiple strategies simultaneously.

But it doesn't eliminate market risk or guarantee profitability.

The Business Lesson

Test your ideas before investing your money.

Hudson's approach applies beyond trading.

Whether launching a business, testing a marketing campaign, or developing a new product, use data to validate your assumptions before committing significant resources.

A failed experiment is much cheaper than a failed investment.

Watch the original interview: https://www.youtube.com/watch?v=iNsCeXCy1oQ

Disclaimer: This article summarizes Matteo Hudson's views from the interview. Trading involves financial risk, and historical backtesting does not guarantee future results.