Gabe's Guess Probability

This classic quant prep question is about updating beliefs when you see only the outcome, not the process that produced it. You observe that an answer on a test is correct, but you do not directly observe whether it came from genuine knowledge or from guessing. You have prior information about how often each hidden cause occurs, and you must infer which cause is more plausible given the success you saw.

It trains conditional probability, working with priors, likelihoods, and a careful normalization step. You practice distinguishing between the chance of an event under a scenario and the chance of the scenario given the event, a confusion that often appears in quant interviews and probability reasoning.

This matters for quant interviews because market data are noisy signals of hidden states. Strong conditional reasoning under uncertainty is central to trading decisions, model calibration, and risk inference.

What it tests

This problem class is governed by Bayes' Theorem, which allows us to reverse conditional probabilities when direct information is unavailable. The key structure is that we observe an outcome (such as a correct answer) that can arise from multiple underlying causes (knowing or guessing), each with their own likelihood and prior probability. Bayes' Theorem combines the prior probability of each cause with the likelihood of the observed outcome under each cause, then normalizes over all possible causes. This pattern holds because, in many real-world scenarios, we see effects and want to infer the probability of their hidden causes, not just predict effects from known causes. The normalization step ensures that the probabilities sum to one, reflecting all possible ways the observed event could have occurred.