Most candidates hit Optiver OA and realize "this is nothing like LeetCode". Optiver tests mental math speed + market-making intuition + probability improvisation. The exact mix shifts across Trader / SWE / QR, but the underlying logic is constant: pace beats per-question accuracy. This review structures common questions, how to prepare, and the OA assist playbook.
Optiver OA Surface by Role
| Role | Mental Math | Market Making | Probability | Algorithms |
|---|---|---|---|---|
| Trader Intern | ★★★★★ | ★★★★★ | ★★★★ | ★ |
| Trader NewGrad | ★★★★★ | ★★★★★ | ★★★★ | ★ |
| SWE | ★★ | ★ | ★★★ | ★★★★★ |
| QR / QT | ★★★ | ★★ | ★★★★★ | ★★★ |
Common Question 1: 80in8 Mental Math
Surface
- 80 questions in 8 minutes → 6 seconds each
- Add / subtract / multiply / divide + decimals + percentages + ratios
- e.g.
3.4 × 2.5 = ?,(160 - 17) / 11 = ?,12% of 850 = ?
How to prepare
- Master mental-math templates
- ×11: outer + inner pair (
23 × 11 = 253→ 2_5_3) - ×25: ÷4 then ×100
- Difference of squares:
(a-b)(a+b) = a² - b²
- ×11: outer + inner pair (
- 200 timed questions / day for two weeks → 6 seconds average
- Use tradermath.org / Rankyourbrain style simulators
- Train the skip reflex: anything over 8 seconds, skip without remorse
Bar
Community reports: 60+ correct passes, 65+ correct → ~70% onsite probability.
Common Question 2: Market Making
Surface
A virtual instrument with fair = 50. The market pushes bid/ask quotes every second; you must post two-sided quotes (the spread defines your inventory risk). 8 rounds × 3 minutes.
How to prepare
- Three-step chain: fair → spread → quote
- Inventory control:
inv > 0→ tighten ask;inv < 0→ tighten bid - Spread management: wider when volatility rises
- Avoid adverse selection: tighten or pull quotes if hit repeatedly
Simplified Python template (for understanding)
class SimpleMaker:
def __init__(self, base_spread=0.4):
self.base = base_spread
self.inv = 0
self.fair = 50.0
def quote(self, mid, vol_proxy):
spread = self.base + 0.05 * vol_proxy
skew = -0.02 * self.inv
return (mid + skew - spread / 2,
mid + skew + spread / 2)
Bar
Top-50% PnL + non-blowup inventory. Optiver's grader is not "highest PnL wins" — it's a weighted combo of PnL + inventory control + pace consistency.
Common Question 3: Probability Reasoning
Surface
- Cards / dice / urns
- Decision trees + expected value
- Conditional probability with Bayes
How to prepare
- First two chapters of Heard on the Street — 50 classic brain teasers
- 30–60 second oral solutions per problem — don't write the formula first; state symmetry / conservation / complement
- Core intuitions: Bayesian reverse, Markov stationary distribution, symmetry reduction
Examples
"Probability of at least one 6 in 4 fair die rolls?"
1 - (5/6)^4 ≈ 0.518
"Urn 1: 5 black, 5 white. Urn 2: 8 black, 2 white. Pick urn at random, draw a black ball. P(urn 2 | black)?"
P(urn 2 | black) = (0.5 × 0.8) / (0.5 × 0.5 + 0.5 × 0.8) = 8/13
Common Question 4 (SWE / QR): Algorithms + Probabilistic Code
For SWE / QR roles Optiver also runs algorithm + probability code:
- Order Book matching: two heaps + qty map
- Trading Sequences: sliding + average
- Allocation: rounding + leftover round-robin
- Monte Carlo convergence: bucketing + variance estimation
Full solutions in our Optiver SWE OA Order Book + Allocation guide.
How to Prepare: 5-Day Sprint
| Day | Task |
|---|---|
| D1 | 200 mental math questions / day, target ≤ 6 sec average |
| D2 | 4 market-making rounds, log PnL + inventory path |
| D3 | 25 probability brain teasers, ≤ 60 sec oral each |
| D4 | Full OA mock (mental math + market making + probability) |
| D5 | Weakness review + mental conditioning (pace; don't fixate) |
OA Assist Playbook
What oavoservice OA assist gives you
- Role-customized banks: Trader / SWE / QR with separate high-frequency sets
- 80in8 timed simulation: daily set + mistake review
- Market-making mentor sparring: mentor quotes / scores; 4 rounds simulated
- Probability oral training: simulated interviewer "why is it this number?" follow-ups
The biggest pitfall
The biggest pitfall isn't difficulty — it's pace. We've watched candidates score 100% on the first 30 questions, get stuck for 30 seconds on problem 31, and finish only 55. The OA is deliberately undoable in full. You must train the "skip after 8 seconds" reflex.
Add WeChat Coding0201 for pricing and scope.
FAQ
How long is the full Optiver OA?
Trader: ~60 minutes total (math 8 + market making 25 + probability 20 + buffer). SWE: 70–90 minutes algorithms, no math / market making.
Can I use a calculator?
Mental math round explicitly forbids it. Market-making round allows pen and paper but no on-screen calculator.
What tools for mental math practice?
tradermath.org, Rankyourbrain, Optiver's own simulator (when available). 200 questions / day for 14 days.
Cooldown after a fail?
Typically 6 months. Switching role family (SWE → Trader) often goes into a separate pool.
Preparing for Optiver / IMC / Akuna / SIG / Citadel Securities OA / VO?
oavoservice tracks market-makers and prop shops (Optiver / IMC / Akuna / SIG / Citadel Securities / DRW). Mentors come from live market-making / prop teams and provide role-customized banks, 80in8 timed simulation, market-making sparring, and probability oral training.
👉 Add WeChat: Coding0201 for the Optiver experience review and OA assist plan.
Contact
Email: [email protected]
Telegram: @OAVOProxy