Key Takeaways
- A bold prediction that an under-the-radar AI stock will double in 12 months leans on the story of the AI boom, but a good story and a rising price moving together does not prove one drives the other.
- Correlation measures how two variables move together. It says nothing about which one causes the other, or whether a third factor drives both. This distinction is central to correlation vs causation investing.
- A past relationship is a description, not a forecast. Candidates should ask how long the data covers, whether the link is stable, and what could break it.
A Prediction That Starts With ChatGPT
A new prediction argues that an under-the-radar artificial intelligence stock is going to double over the next 12 months. The case opens with the origin story of the sector. The AI boom, as the argument goes, unofficially began in January 2023, after OpenAI's ChatGPT reached 100 million users faster than any software application in history.
That is a striking fact, and it is a fair way to date the moment when investors began paying attention to AI. But notice what it does and does not establish. It shows that adoption of one product was extremely fast. It does not, on its own, tell you that a particular company will earn more, that its shares will rise, or that they will rise by 100% within a year. The prediction asks the reader to jump from a sector-wide story to a precise price target for one stock.
What Correlation Actually Tells You
Correlation is a measure of how closely two variables move together, usually summarised in a single number between -1 and +1. A reading near +1 means they tend to rise and fall together. That is all it means. It is a description of a pattern in past data.
In the AI story, you could easily imagine a strong positive correlation between measures of AI enthusiasm (user numbers, media attention, spending announcements) and the share prices of AI-linked companies since early 2023. A high number would look persuasive. But the number cannot tell you the direction of influence. Did user growth lift the stock, or did rising prices attract attention and more investor interest in the story? It also cannot rule out a third factor, such as falling interest rate expectations or a broad rally in technology shares, pushing everything up at once.
This is why correlation vs causation investing matters in practice. A causal claim needs a mechanism: a reason why one thing makes the other happen, supported by more than co-movement.
Why a Doubling Forecast Needs More Than a Pattern
A forecast is a statement about the future, and it needs assumptions about what stays the same. If the boom years of 2023 onward shaped the relationship you are relying on, you have to ask whether those conditions will repeat. The period since the boom began is a short sample, and a short sample dominated by one theme can make almost any relationship look strong.
There are a few classic traps worth keeping in mind:
Spurious relationships
Two series can trend upward together for no meaningful reason. When you search many possible variables, some will line up by chance. Finding a match after the fact is very different from predicting it beforehand.
Outliers and survivorship
A handful of spectacular winners can drag a correlation upward. Stories about stocks that doubled are easy to find, while the ones that fizzled attract far less attention. A prediction built on the winners tells you little about the odds for any single pick.
Unstable relationships
A link that held in one market regime can weaken or vanish in another. A stock that rose with AI enthusiasm can fall just as quickly with it, because the same shared driver works in both directions.
How to Read the Next AI Prediction
When you meet a confident forecast, whether in a news feed or on an exam vignette, run a short checklist. First, what are the two things being linked, and is the link measured or merely asserted? Second, could the arrow run the other way, or could something else be driving both? Third, how much data supports it, and does it cover more than one kind of market? Fourth, what would have to be true for the forecast to fail, and has the author said so?
These questions also work on the exam. The quantitative material at Level 1 often tests whether you can tell what a statistic does and does not support. A candidate who reads a correlation as proof of cause will lose marks on exactly the questions designed to catch that habit. The same discipline protects real portfolios, because position sizing should reflect how uncertain a forecast really is, not how exciting its story sounds.
Test Your Own Understanding
The gap between spotting a correlation and proving a cause is easy to nod along to and surprisingly easy to get wrong under exam pressure. If you want to see how you would handle it, try a free EduFite mini mock exam and check whether you can separate what a statistic describes from what it predicts. The full EduFite course then builds the mechanics and practice you need to answer these questions with confidence.
FAQ
What is the difference between correlation and causation in investing?
Correlation means two variables tend to move together in past data. Causation means a change in one actually produces a change in the other. Investors often see strong correlations, but proving causation requires a clear mechanism and evidence beyond co-movement.
Does a high correlation help forecast a stock price?
Not by itself. A high correlation describes a historical pattern. It can be a starting point, but a forecast also needs the relationship to be stable, supported by enough data, and not driven by chance or a hidden third factor.
Why is a spurious correlation a risk for investors?
Spurious correlations appear when two series line up by coincidence or because both trend over time. Acting on them can lead to confident decisions built on a relationship that has no real economic link.
Is correlation vs causation tested on the CFA Level 1 exam?
Yes. The quantitative methods material covers what correlation measures and its limitations, and questions often test whether you can tell what a statistic does and does not show.

