ChatGPT for Traders and Investors: How AI Now Predicts Market Moves Like Language

Also about AI for Finance:

AI pattern Recognition for Financial Markets

ChatGPT predicts words. Forecaster Projection tool predict markets.

Introduction

After years of relentless work, countless attempts, moments of frustration, and flashes of inspiration, we can finally unveil something that has long been a dream: Pattern, the new feature inside the Forecaster Terminal.

This is not just another update. This is the product of an idea we carried for years: what if we could give traders and investors a tool that recognizes market patterns the way the human brain recognizes language, and turns decades of market experience into a split-second insight?

Today, that idea has become reality.

From Predicting Words to Predicting Prices

Let’s start with something simple. If we give you the phrase: “Mom went to the store and bought some …”, what comes to mind?

For most people, the answer is “milk.” Some might say “bread” or “apples,” but overwhelmingly, the choice is milk. Why? Because it’s the most statistically likely continuation of that sentence.

That’s exactly how ChatGPT works. It doesn’t think like a human. Instead, it uses probability and mathematics to predict the most likely word to follow in a given context, based on everything it has learned from billions of texts.

Projection applies the same concept to financial markets. Instead of predicting the next word, it predicts the next movement in prices. It looks at how an asset has been moving recently, searches through decades of historical data to find similar patterns, and then tells us what happened next in those cases.

In other words: where ChatGPT continues sentences, Pattern continues price charts.

How Projection Works

At its core, Projection is powered by machine learning and pattern recognition. Here’s the process:

Analyze the recent movement. For example, take the last three months of the DAX index.
Scan decades of history. Projection goes back as far as 30 years, searching for similar trajectories in the past.
Build a probability map. It shows what happened next in those historical cases, transforming history into a probabilistic forecast.

The output is displayed directly on the chart:

The average of bullish cases (green line).
The average of bearish cases (red line).
The most correlated single case (blue line).

Here’s a real example: On the DAX, Projection identified a recurring pattern. Every time the index behaved this way in the past, the outcome was a subsequent rally. In the current projection, the model highlights a possible target at 25,716 points (roughly +8% from current levels). But it also reveals a consistent detail: before rising, the market often dipped by around -4%.

That subtlety – anticipating the pullback before the move – is the kind of insight that makes Pattern so powerful for trade planning.

Probability Is Not Enough: The Power of Robustness

But we knew from the start that probability alone wasn’t enough. Saying “there’s an 80% chance the market goes up” isn’t useful if the data behind it is weak.

So we created a second measure: Robustness.

Every Pattern comes with a rating from 0 to 5 stars, which reflects the quality and consistency of the forecast. To earn five stars, several conditions must be met:

The dominant scenario must hold a probability above 70%.
Historical cases must show internal consistency (similar follow-through).
The most correlated case must align with the average projection.
In most cases, prices must have gone beyond the boundaries of the original range (showing decisive movement, not just noise).
The structure of the most important case must closely mirror the overall pattern.

When all of these line up, as they did in the DAX example, you get a projection that isn’t just likely, but also statistically solid.

From Individual Charts to a Global View: The Quantum Screener

Now imagine applying this not to just one chart, but to the entire market universe. That’s where the Quantum Screener comes in.

We integrated Pattern directly into the Screener. This means you can now scan entire asset classes. indices, stocks, currencies, commodities, and instantly see where Pattern identifies the strongest probabilities.

Here’s what I’ve already found in practice at this moment:

The DAX emerged as the index with the strongest bullish projection.
The NZD/JPY currency pair showing clear bearish tendencies.
Meta, where historical outcomes lean bearish in two out of three similar cases.

With just a couple of clicks, you can now ask: “Which markets have more than 80% probability of a bullish move?” or “Which assets have the most bearish projections?” And the Forecaster delivers.

This is like having a market radar that scans decades of history and tells you where the winds are blowing.

Synergy With Other Tools

One of the most exciting things for me is how Pattern works in synergy with the tools you already use in Forecaster.

Combine it with seasonality: if Projection shows a pullback before a rise, and seasonality also expects a weak September, you’ve got alignment.
Add fundamentals: if the DAX is fundamentally undervalued, a projected pullback only makes the eventual rally more compelling.

Use it in trading systems: for example, trade only when Pattern probabilities exceed 80%, and filter signals with other criteria.

And there’s another powerful feature we are particularly proud of: the Macro Comparison. For instruments that allow it – such as commodities, and gold in particular – Forecaster automatically retrieves the macroeconomic environment of the most correlated historical case, even if it’s 20 or 30 years old, and places it side by side with today’s situation. This lets you directly compare monetary policy, inflation, the labor market, consumer confidence, investor sentiment, and even geopolitical events from then and now.

In short, Pattern isn’t meant to replace analysis. It’s designed to augment it, giving you the equivalent of thirty years of chart experience condensed into half a second.

Real Results: The S&P 500 Backtest

Skeptical? Fair enough. We asked ourselves the same question: “Does it actually work?”

So I ran a simple test on the S&P 500:

Go long when the bullish probability ≥ 80%.
Go short when the bearish probability ≥ 80%.

The results were clear:

Win rate: 61%
Profit factor: 2.37 (meaning profits were, on average, 2.37x greater than losses).

That’s with a basic, rules-only approach. And in some cases, the signals were what I call “open-goal trades.” For instance, when Nvidia was at its lows, Projection screamed bullish. The outcome? A powerful rally that matched the projection almost perfectly.

Limitations and Edge Cases

Pattern isn’t magic. It has limits.

If an asset has little price history, Projection may not find enough data for a 3-month pattern. In those cases, shorter time frames (like one month) can still yield meaningful results.

For Bitcoin, for example, a 3-month pattern sometimes finds no matches. But at 1-month intervals, it produces valid probabilities – such as 60% bearish vs. 40% bullish with 4/5 robustness.

This is important: Pattern is not about forcing predictions, but about uncovering what history can reliably tell us.

The Philosophy: No Black Boxes

I want to make one thing absolutely clear: Pattern is not a black box. I don’t expect you to blindly trust an algorithm.

That’s why I designed it so you can:

Inspect every historical case.
Hover over tooltips to see probabilities and numbers.
Save every projection as an image with date and instrument name for future review.

The purpose of Pattern is not to hide the logic, it’s to give you clarity and confidence.

Conclusion: Turning Vision Into Action

If ChatGPT has revolutionized the way we interact with text, Pattern aims to do the same for financial markets.

It takes what only seasoned traders with decades of experience can see at a glance: patterns, recurrences, probabilities. And it makes this available to everyone, instantly, across every market in the world.

It’s more than a tool. It’s the realization of a vision: harnessing the power of AI, machine learning, and pattern recognition to turn raw history into actionable insight.

And it’s already inside your Forecaster Terminal.