Probability Intuition For Quants
This conversation was generated using Cook'd AI's career assistant and reviewed by Fallon Han from our editorial team for accuracy.
I'm a fourth-year maths student in Australia and I'm prepping for quant research interviews. I keep seeing probability questions in trading and risk, but I want the intuition first. How does a core probability idea actually show up in finance interviews?
A core idea is conditional probability. In interviews, it often shows up as: given new information, how should your belief change? In trading, that means updating the odds of an event after a signal arrives, like earnings news or an order flow imbalance. In risk, it is similar: the chance of a loss is not fixed if market conditions change. If you can explain that probabilities are not just static frequencies but beliefs updated by evidence, you are already thinking like a quant.
So if I hear a question about event A happening after event B, should I immediately think Bayes? How would that connect to a trading example instead of just a textbook setup?
Yes, Bayes is often the right frame. The interview version is usually less about writing the formula and more about naming the information flow. For example, suppose a stock jumps after a positive surprise in sector data. You might ask: does that surprise really increase the probability of a strong earnings report, or is it already priced in? Bayes helps you update the prior probability of a good outcome using the new signal. In trading, the key is not the math alone, but whether the signal adds genuine information.
That makes sense. But interviewers also ask about independence and correlation. I always get nervous because in markets things look connected even when they are not. How should I explain that cleanly?
A clean explanation is: independence means one event gives no information about another, while correlation measures linear co-movement, not full dependence. In finance, people often confuse the two. Two assets can have low correlation and still become linked during stress, which is why correlation can break down in a crisis. A good interview answer is to say that independence is a much stronger condition and is rarely safe to assume for markets. Then give a risk example, like diversification working in normal times but failing when everything sells off together.
Okay, so if they ask about variance or expected value, how do I tie that back to risk? I know the formulas, but I want the finance intuition for why a hedge fund cares.
Expected value is the long-run average payoff, so it tells you whether a strategy is profitable on average. Variance tells you how spread out outcomes are, which matters because a strategy with a good average can still be too unstable to hold. Hedge funds care about both because capital is constrained by drawdowns and tail risk, not just average return. A strong interview link is: expected value answers 'is this edge real?' and variance answers 'can we survive long enough to realize it?' That connection shows you understand both alpha and risk.
Could you give me a simple example that combines all of this, maybe one that sounds like a quant research screening question?
Sure. Imagine a signal that is correct 60 percent of the time and wrong 40 percent of the time. The payoff is large when it is right, but the loss is also sizable when it is wrong. An interviewer might ask whether this is a good trading signal. Your answer should combine conditional probability, expected value, and risk. First, ask what the hit rate means conditional on the market regime. Second, compute whether the average payoff is positive after costs. Third, check whether the variance or tail losses are so large that the strategy is impractical. That is the mindset they want: not just 'is the signal accurate,' but 'does it have usable edge after risk and frictions?'