Optiver OA has always been known as "speed + correctness", this year's version of the set of 12.4 is still relentless: three Coding, all trading scenarios simulation, the amount of code is not large, but the logic is very round, any step missed edge case will hang. Our students were assisted by Programhelp OA on-line without any trace, and the pace of the whole process was pressed very steadily, and the three questions were passed at once.
Below is a review of the content and difficulty distribution of the questions in order.
Problem 1: Counting Trading Sequences
The question is not long, but very "Optiver-like". Here's the title:
- A price series
- asking you count all valid profitable trading sequences
Core rules:
"A valid trading sequence is defined as a pair (i, j) such that
i < j and price[i] < price[j].
Count how many such valid trading sequences exist."
It looks like counting increasing pairs, but in reality there are a lot of pitfalls:
- Do duplicate prices count as valid
- Whether successive downward intervals affect statistics
- Intervals of length 1 must be ignored
- Non-strict ascent vs strict ascent to be distinguished
- It takes O(n) or O(n log n) to pass for large inputs.
Simple and violent will be TLE, in the end, our participants are made in a way that "maintains the prefix smaller number of elements".
Problem 2: Backtesting a Proportional Allocation Strategy
The second question is a typical backtest question:
"Given daily returns for multiple assets and a fixed proportional allocation vector.
compute the cumulative return of the portfolio from day 1 to day N."
Essentially, it simulates portfolio returns:
- Input:
returns[day][asset] - Daily:
portfolio_return = sum(weight[i] * returns[day][i]) - Output: cumulative returns (usually multiply, not sum)
The real point of this question is not the math:
- Floating Point Error Accumulation
- weight + return is a decimal at the same time
- Insufficient precision will cause the last two hidden tests to fail outright.
The trainee started with Python float and got stuck at 89% test, Programhelp reminded him to "recalculate the daily accumulation with Decimal", and then he got an AC without any problem.
Problem 3: Order Book Matching Simulation
This is the most difficult question in the entire OA set and is a complete engineering-level simulation. The question provides a string of orders, each in the following format:
{ "type": "buy" / "sell", "price": x, "size": y }
"Simulate an order book and return all unmatched orders after processing the list."
You need to implement a Two-way order book:
- buy book: high to low
- sell book: from low to high
- Immediate matchmaking for every new order.
- Support for partial match
- FIFO for orders at the same price
- leftover before adding book
The whole process is very much a test of detail:
- Buy to eat the cheapest seller
- Sellers want to eat the highest buyers
- You can't break after a partial fill, you need to continue the loop.
- price Orders that come in at the same time but at different times should be kept in order.
- Delete order node with size=0
Many students fall in a classic pit:
Continue summarizing when remaining size > 0, instead of pushing directly to book.
Our trainee was pausing in this same logic, and the Programhelp voice reminded of the critical step: "don't break - continue matching until size becomes zero ", and all hidden use cases passed after matching.
Summarize
Optiver The difficulty level of OA this year was high overall, but the questions were very consistent:
| Problem | Type | Difficulty | exam point |
|---|---|---|---|
| P1 | Counting sequences | Middle | Monotonic Logic + Boundaries |
| P2 | Portfolio backtest | upper middle class | Floating Point Precision + Cumulative Calculation |
| P3 | Order book simulation | your (honorific) | Trading system simulation, queuing, sorting |
Our trainees ended up ACing three questions at a time in about 50 minutes.Optiver's OA features are:
The code has to be written fast, but even more so, it has to be written accurately, with no room for minor logic errors.
Why you need ProgramHelp to aids?
Starting salary for New Grad Package at Optiver is usually in the range of $200k - $300k (Total Comp). This is one of the most expensive exams of your career.
The most complete Quant question bank on the net: We cover the latest questions from top companies like Optiver, Jane Street, Citadel, Akuna and many more. The three questions mentioned earlier have been fully solved in our Backend system.
Ex-FAANG & Quant mentor sitting: Our assistance is not ordinary "ghostwriting", but by experienced engineers with real-world experience. Real-time Coding Support.
100% Safety and Originality: We not only guarantee All Cases Passed, but also focus on the naturalness of code style. We will adjust variable naming and logic structure according to your past code habits, and completely avoid Code Signal and HackerRank checking and AI judgment.
