← Sig Interview Insights

Sig·Software Engineer·Technical Phone Screen·Junior

Junior
Apr 2026

Summary

SIG quant researcher interview with a probability question that looks straightforward but trips you up if you're sloppy with the setup.

Questions Asked (1)

Q1

On a platform where 70% of stories are real and 30% are fake, a fake story has an 80% chance of getting 100+ likes while a real story only has an 8% chance. Given that a randomly selected story has 100+ likes, what's the probability it's real news?

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 Bayes' theorem problem and set up the calculation using the given probabilities. Compute the probability of a story being real and having 100+ likes, then divide by the total probability of having 100+ likes (summing over real and fake stories).

Pro tip: After calculating, sanity-check the result: since fake stories are much more likely to get 100+ likes, the posterior probability of real should be lower than the prior (70%). This helps catch errors.

1. Define events and probabilities

Let R be the event that a story is real, F that it's fake, and L that it has 100+ likes. Write down P(R)=0.7, P(F)=0.3, P(L|F)=0.8, P(L|R)=0.08.

2. Compute joint probabilities

Calculate P(R and L) = P(R)*P(L|R) = 0.7*0.08 = 0.056. Calculate P(F and L) = P(F)*P(L|F) = 0.3*0.8 = 0.24.

3. Find total probability of L

Sum the joint probabilities: P(L) = P(R and L) + P(F and L) = 0.056 + 0.24 = 0.296.

4. Apply Bayes' theorem

Compute P(R|L) = P(R and L) / P(L) = 0.056 / 0.296 ≈ 0.1892, or about 18.9%.

5. Interpret and sanity-check

The probability is about 18.9%, which is lower than the prior 70%, as expected because fake stories are much more likely to get 100+ likes. This makes sense.

Key Points to Mention

  • Bayes' theorem formula: P(R|L) = P(L|R)P(R) / P(L)
  • Law of total probability: P(L) = P(L|R)P(R) + P(L|F)P(F)
  • Correctly identifying the base rates (70% real, 30% fake)
  • Computing the joint probabilities accurately
  • Interpreting the result in context: the high likelihood of likes for fake stories skews the posterior
  • Sanity-checking the result against the prior probability

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