Design dice betting strategy under time limit

Quick Overview

This question evaluates probabilistic reasoning, expected-value computation, risk management and staking-strategy concepts including Kelly-style bet sizing, confidence-interval estimation, and return–variance trade-offs in the Statistics & Math domain.

Design dice betting strategy under time limit

Company: Optiver

Role: Software Engineer

Category: Statistics & Math

Difficulty: medium

Interview Round: Technical Screen

You have T = 10 minutes and initial bankroll B = $1,000 to play a repeated betting game with a fair six‑sided die. Before each roll, you may stake any fraction of your current bankroll on exactly one offered wager: (A) “roll is 6” pays 6:1, (B) “roll is 5 or 6” pays 2.2:1, or (C) “roll is odd” pays 1.1:1. Rolls take 3 seconds and outcomes are independent. Unlimited number of bets within the time limit. a) Which wager (s) have positive expected value? Show EV calculations. b) For your chosen wager, compute the optimal bet fraction to maximize expected log wealth (Kelly) and propose a more conservative fraction to keep the probability of a 50% drawdown below 5%. c) Outline a practical staking policy for the time‑limited setting and explain when you would stop betting. d) After n = 60 “odd” bets you observed k = 36 wins. Provide a fast 95% confidence interval for p and interpret it. e) Discuss the return–variance trade‑offs and how your allocation changes if rolls slow down or T is shorter.

Quick Answer: This question evaluates probabilistic reasoning, expected-value computation, risk management and staking-strategy concepts including Kelly-style bet sizing, confidence-interval estimation, and return–variance trade-offs in the Statistics & Math domain.

|Home/Statistics & Math/Optiver
Optiver logo
Optiver
Aug 12, 2025, 12:00 AM
mediumSoftware EngineerTechnical ScreenStatistics & Math
6
0

Repeated Dice Betting with Time and Risk Constraints

You have T = 10 minutes and an initial bankroll B = $1,000 to play a repeated betting game with a fair six‑sided die. Each roll is independent and takes 3 seconds, so you can make at most N = floor(10×60 / 3) = 200 bets.

Before each roll, you may stake any fraction of your current bankroll on exactly one of the following wagers. Payoffs are net profits per $1 staked (you also lose your stake on a loss):

  • (A) “roll is 6” with probability p = 1/6, pays b = 6:1
  • (B) “roll is 5 or 6” with p = 2/6 = 1/3, pays b = 2.2:1
  • (C) “roll is odd” with p = 3/6 = 1/2, pays b = 1.1:1

Tasks:

  1. Which wager(s) have positive expected value? Show EV calculations.
  2. For your chosen wager, compute the optimal Kelly bet fraction that maximizes expected log wealth, and propose a more conservative fraction that keeps the probability of a 50% drawdown below 5%.
  3. Outline a practical staking policy for this time‑limited setting and explain when you would stop betting.
  4. After n = 60 “odd” bets, you observe k = 36 wins. Provide a fast 95% confidence interval for the win probability p and interpret it.
  5. Discuss the return–variance trade‑offs, and how your allocation changes if rolls slow down or if T is shorter.
Loading comments...