Jane Street OA HackerRank: Format, Reported Questions and Probability Practice

A Jane Street OA on HackerRank can involve probability rather than coding. One public candidate account from 2016 describes four math questions in 30 minutes. It is useful practice material, but it does not establish the format of a current invitation. Check your role, question types and time limit before choosing what to study.

This guide separates three things: what Jane Street says on its own site, what candidates report about the HackerRank test, and original practice questions we wrote with worked answers. Our practice questions are modeled on the reported style. They are not Jane Street questions.

Research note: Checked November 7, 2026, against the linked Jane Street pages and a historical candidate account. This guide was edited with AI assistance; its four practice answers were independently checked with exact fractions in Python. We did not take a Jane Street assessment. SkillCopilotAI publishes this guide and sells an AI interview assistant. Use it for preparation, and follow the rules of your actual assessment.

What Jane Street officially says

The official pages linked below describe interviews and preparation. They do not confirm a current HackerRank OA format. They do provide useful guidance on what to study:

The clubs suit covers expectation and linearity; diamonds covers classical and conditional probability. Start with the first 5–6 cards in those suits, as Jane Street recommends. The harder cards are not a checklist of what you must solve to get hired.

A historical HackerRank report: what it does and does not show

In an October 2016 GitHub account, a candidate describes a Jane Street HackerRank test on math and probability. The account supports the following details for that sitting:

Jane Street's interviewing page says quantitative research interviews combine trading and software engineering. That does not specify a universal separate coding OA. Ask recruiting which format applies to your application.

Four questions recalled in 2016

These are short paraphrases of the candidate's historical account, not a current question forecast. For the dice and coin exercises, use the usual probability assumptions: a fair die, equally likely coin selection and independent flips given the selected coin.

1. Select four of the first 16 primes uniformly without replacement. What is the probability their sum is even?
2. Roll an eight-sided die, observe it, then stop or roll once more. Your payoff is the total unless it exceeds eight, when it becomes zero. Which first rolls should you keep?
3. Select one of three coins with heads probabilities 20%, 20% and 60%. After observing heads, what is the probability of heads on the next flip of that same coin?
4. Two indistinguishable urns contain four $1 and six $10 chips, and three $1 and seven $10 chips. Choose an urn uniformly and draw a $1 chip. Do not replace it. Should the next draw come from the same urn or the other urn, and what is its expected value?

Use these to practise updating probabilities after an observation and comparing expected payoffs before choosing a strategy.

Original practice set with worked answers

These are teaching exercises adapted from standard probability patterns, with their own numbers and worked explanations. They are not claimed as Jane Street questions. All four answers were checked with exact fractions in Python.

1. Parity of a prime sum

Choose a three-element subset uniformly from the first 10 primes (2, 3, 5, ..., 29). What is the probability its sum is odd?

Answer: 7/10. Every prime except 2 is odd. Three odd numbers sum to an odd number. If 2 is chosen, the sum is 2 plus two odds, which is even. So the sum is odd exactly when 2 is not chosen: C(9,3) / C(10,3) = 84 / 120 = 7/10.

Pattern: look for the one element that breaks symmetry, then count complements.

2. Roll again or stop?

Roll a fair 6-sided die. You may stop, or roll once more and add the result. If your total exceeds 7 you get nothing; otherwise you get the total. When should you roll again, and what is the game worth?

Answer: roll again on 1, 2 or 3; stop on 4, 5 or 6. The game is worth 41/9, about 4.56.

If your first roll is s, a second roll pays s + r for every r with s + r ≤ 7, and zero otherwise. The expected value of rolling again is:

Averaging the better choice over the six first rolls gives (4.5 + 25/6 + 11/3 + 4 + 5 + 6) / 6 = 41/9.

Pattern: for one-step stopping problems, build the table. It makes the comparison for each first roll explicit, so you can explain why the threshold falls between 3 and 4.

3. Which coin did you pick?

One coin is fair. The other lands heads 75% of the time. Choose either coin with probability 1/2, flip it and observe heads. Flips are independent given the chosen coin. What is the probability the next flip of that same coin is heads?

Answer: 13/20, or 0.65. After heads, the posterior weights are proportional to 0.5 and 0.75, so the fair coin has probability 0.4 and the biased coin 0.6. The next-flip probability is 0.4 x 0.5 + 0.6 x 0.75 = 0.65.

Pattern: update first, then average. A common mistake is to answer the posterior probability of the biased coin instead of the probability of the next head.

4. Quick expected value

You roll a fair die and may reroll once, keeping the second result. What is the game worth with best play?

Answer: 4.25. A reroll is worth 3.5, so keep 4, 5 or 6 and reroll 1, 2 or 3. The value is (3 x 3.5 + 4 + 5 + 6) / 6 = 25.5 / 6 = 4.25.

Pattern: compare each outcome with the value of continuing. This idea returns in harder forms as optimal stopping.

How to prepare in two weeks

Read the Probability & Markets guide linked from Jane Street's interviewing page alongside these exercises. On day 13, the four-question mock is a suggested practice session based on the historical report; adjust its length and content to match your invitation.

Checklist before you open the link

• Read the invitation for time limit, question count and whether explanations are required.
• Keep paper and a pen for trees and tables, unless the instructions say otherwise.
• Answer in exact fractions when you can; convert only if asked.
• Practise giving a short explanation; the historical questions ask for reasons or strategies. Follow your invitation's answer requirements.
• Budget reading and review time. For a 30-minute, four-question mock, eight minutes per question would already exceed the total; move on earlier if you are stuck.

Related guides

For a similar quant screen with different rules, read our Optiver OA guide. If your Jane Street role does involve a coding round, our HackerRank OA preparation guide and Citadel HackerRank guide cover coding-style tests.

FAQ

Is the Jane Street OA a coding test on HackerRank?

It can be a probability test, as in the linked 2016 account. That does not tell you whether your own invitation includes coding. Jane Street describes different interview preparation for trading, research and engineering; check your assigned track.

How long is the Jane Street HackerRank test?

The 2016 candidate account describes four questions in 30 minutes. We cannot confirm that as the current standard. Use the duration and question count on your invitation.

Does Jane Street confirm it uses HackerRank?

The official pages linked in this guide did not specify a HackerRank OA format when checked on November 7, 2026. The hosting claim here comes from the linked historical candidate account.

What should I study first?

Start with expected value and conditional probability. Jane Street recommends the first 5–6 questions in the clubs and diamonds suits of its probability playing cards.

Can I retake the Jane Street OA?

Jane Street's FAQ says candidates can reapply and generally suggests students wait about a year. That is reapplication guidance, not a specific OA retake policy. Ask recruiting about your case.