← Boston Consulting Group Interview Insights

Boston Consulting Group·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
May 2026

Summary

BCG data scientist technical screen, heavy on probability and Bayesian reasoning. The whole thing was basically one multi-part coin problem that escalated from posterior computation into expected value derivation.

Questions Asked (3)

Q1

You have a fair coin and a biased coin where P(Heads) = 0.7. You pick one uniformly at random and flip it three times. Given that the first two flips are both Heads, what is the probability you picked the biased coin?

Product Analytics & MetricsAlgorithms & Data Structures
Author's notes

Classic Bayes setup.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Recognize this as a Bayesian inference problem where you need to update the prior probability of picking the biased coin (0.5) based on the observed data (two Heads in the first two flips). Compute the likelihood of observing two Heads under each coin (fair: 0.25, biased: 0.49) and apply Bayes' theorem to find the posterior probability.

Pro tip: State your assumptions clearly (e.g., independence of flips, uniform prior) and consider mentioning that the third flip is irrelevant to the conditional probability given the first two flips, showing you understand conditional independence.

1. Define events and prior probabilities

Let B be the event of picking the biased coin, F the fair coin. Prior: P(B) = P(F) = 0.5. Let E be the event that the first two flips are Heads.

2. Compute likelihoods

Under fair coin: P(E|F) = (0.5)^2 = 0.25. Under biased coin: P(E|B) = (0.7)^2 = 0.49.

3. Apply Bayes' theorem

Use P(B|E) = P(E|B)P(B) / [P(E|B)P(B) + P(E|F)P(F)] = (0.49*0.5) / (0.49*0.5 + 0.25*0.5) = 0.49 / (0.49+0.25) = 0.49/0.74 ≈ 0.662.

4. Interpret and communicate the result

The probability is approximately 66.2%. Explain that observing two Heads increases the likelihood of the biased coin from 50% to about 66%.

Key Points to Mention

  • Bayes' theorem and its application to conditional probability
  • Prior probability of picking each coin (uniform, so 0.5 each)
  • Likelihood of observing two Heads under each coin (0.25 for fair, 0.49 for biased)
  • Independence of coin flips (each flip is independent)
  • The third flip is not used in the calculation because the condition is only on the first two flips
  • The final posterior probability is approximately 0.662 or 66.2%

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

Q2

Using the same setup, what is the probability that the third flip is also Heads, given the first two were Heads?

Product Analytics & MetricsAlgorithms & Data Structures
Author's notes

You have to use the posterior from part one here, not the original priors.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, clarify the assumptions about the coin and the flipping process, especially whether the flips are independent and the coin is fair. Then, if independence holds, explain that the probability of the third flip being Heads is simply 1/2, regardless of previous outcomes. If there is any doubt, discuss how you would test for independence or update beliefs using Bayes' theorem.

Pro tip: Always state your assumptions explicitly—interviewers at BCG value clarity and rigor. Mention that in real-world data, you would check for independence and fairness before applying the simple probability.

1. Clarify assumptions

Ask or state whether the coin is fair and whether flips are independent. This sets the foundation for the probability calculation.

2. Apply independence

If flips are independent, the outcome of the first two flips does not affect the third. Thus, the probability remains 1/2 for a fair coin.

3. Consider alternative scenarios

If independence is not assumed, discuss how you would model dependence (e.g., a biased coin with unknown bias) and use Bayesian updating to compute the probability.

4. Connect to real-world data

Explain how you would test for independence and fairness using data, such as chi-square tests or Bayesian methods, to validate assumptions.

Key Points to Mention

  • Independence of coin flips
  • Fair coin assumption (probability = 1/2)
  • Conditional probability and how it simplifies under independence
  • Bayesian updating for unknown bias
  • Hypothesis testing for independence/fairness
  • Real-world implications and the need to validate assumptions

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

Q3

Still using the same coin from the earlier flips, let T be the total number of flips until you get the first Tail, counting that Tail flip. Derive E[T] given that the first two flips were both Heads.

Algorithms & Data StructuresProduct Analytics & Metrics
Author's notes

This is where it got genuinely hard.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Recognize that after observing two Heads, the process effectively restarts because the coin is memoryless. Let X be the additional flips needed to get a Tail, so T = 2 + X. Since each flip is independent with P(Tail)=p, X follows a geometric distribution with mean 1/p, hence E[T] = 2 + 1/p.

Pro tip: Explicitly state the memorylessness property and define p as the probability of Tails; this shows you understand the underlying assumption and avoids confusion if p is not 0.5.

1. Define the random variable

Let T be the total number of flips until the first Tail, including the Tail flip. Given that the first two flips are Heads, we can write T = 2 + X, where X is the number of additional flips needed to get a Tail.

2. Apply memorylessness

Because coin flips are independent, the fact that the first two flips were Heads does not affect future flips. Thus, X is the waiting time for the first Tail starting from scratch, which follows a geometric distribution.

3. Identify the distribution of X

X ~ Geometric(p), where p is the probability of getting a Tail on a single flip. The expected value of a geometric random variable is E[X] = 1/p.

4. Compute E[T]

Using linearity of expectation, E[T] = E[2 + X] = 2 + E[X] = 2 + 1/p. If the coin is fair (p=0.5), then E[T] = 2 + 2 = 4.

Key Points to Mention

  • Memorylessness of independent coin flips
  • Geometric distribution and its expected value
  • Linearity of expectation
  • Conditioning on past events does not change future probabilities
  • Definition of T including the first two flips
  • General formula in terms of p, with p=0.5 as a special case

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