← Coinbase Interview Insights

Coinbase·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
Jun 2026

Summary

Probability question for a Data Scientist role at Coinbase, focused on a wallet risk scenario. Pretty conceptual, not a lot of coding involved.

Questions Asked (1)

Q1

What statistical assumption must hold for you to multiply individual event probabilities together to get the combined probability of all events occurring?

Product Analytics & MetricsA/B Testing & Experimentation
Author's notes

The answer is independence, which I knew, but the way it was framed around a real wallet risk scenario made me second-guess myself for a second.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by directly stating the core assumption: independence. Then explain what independence means in probabilistic terms and why it's necessary for the multiplication rule. Finally, connect it to practical implications in data science, especially in A/B testing and product analytics, where independence is often assumed but can be violated.

Pro tip: Mention that while independence is the key assumption, in real-world data, events are often conditionally independent given some factors; acknowledging this nuance shows depth. Also, briefly note that if events are not independent, you can use the chain rule of probability (P(A and B) = P(A) * P(B|A)) instead.

1. State the assumption clearly

Directly answer that the events must be independent. Define independence: the occurrence of one event does not affect the probability of the other.

2. Explain the multiplication rule

For independent events A and B, P(A and B) = P(A) * P(B). For more than two events, the product of their individual probabilities gives the joint probability.

3. Contrast with dependent events

If events are dependent, the multiplication rule does not hold; instead, you need conditional probabilities, e.g., P(A and B) = P(A) * P(B|A).

4. Connect to data science context

In A/B testing, independence is assumed between users or sessions; in product analytics, events like clicks might be assumed independent, but often they are not (e.g., user behavior). Highlight the importance of checking independence assumptions.

5. Discuss implications and alternatives

If independence fails, consider using models that account for correlation, such as Bayesian networks or conditional probability. Mention that independence is a simplifying assumption that may not hold in practice.

Key Points to Mention

  • Independence of events: P(A and B) = P(A) * P(B) only if A and B are independent.
  • Definition of independence: P(A|B) = P(A) and P(B|A) = P(B).
  • For dependent events, use the chain rule: P(A and B) = P(A) * P(B|A).
  • In A/B testing, independence is often assumed between experimental units (e.g., users).
  • Violations of independence can lead to incorrect probability calculations and biased results.
  • Conditional independence: events may be independent given a third variable, which is common in causal inference.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.