What the test actually measures
The Optiver Mental Math Test is a timed online assessment that shows up during the quantitative reasoning stage of their trader candidate pipeline. It consists of four rounds of rapid arithmetic—mostly addition and subtraction, with a few multiplication and division questions mixed in. You get roughly twenty minutes for about forty questions. The interface is basic: a question appears, four answer options are listed, and you click the one you think is right. There is no scratch paper. There is no calculator. You have to compute and select fast enough that your brain doesn't freeze mid-way through a long chain of operations. The scores are bucketed. They don't give you a raw number out of forty. Instead you land in a band—something like Tier 3, Tier 4, Tier 5—and that band determines whether you get called back. Tier 5 is the bottom, Tier 1 is the top. A lot of people think this is a test of calculation ability alone. It isn't. It's a test of staying calm when you know you're going to miss something and still pushing through the rest of the questions without spiraling.
Where to find the Optiver Mental Math Test practice materials
The practice version most people use isn't officially from Optiver. It's a PDF that circulates on trading forum communities and some prep sites. It contains around forty questions in a similar format. You can download it by searching for the PDF directly—sites like Prepskills and similar quant prep platforms host it. The questions in the practice PDF closely mirror the style and difficulty of the actual assessment. They're not identical, but the computational range is the same. Addition chains, subtraction with large numbers, a handful of multiplications like 47 times 53, a few divisions. The difficulty is calibrated for someone who can do mental math at a reasonable speed but isn't a savant. Most candidates waste the first five minutes because they treat every question like it's equally important. That is the wrong instinct. The test is long enough that you will hit a question you cannot solve in your head within ten seconds, and if you sit on it, you will not finish the round. Move on. Mark it mentally and come back if you have time. The scoring structure means that getting thirty solid answers right beats getting twelve perfect answers and leaving half the paper blank. For addition and subtraction, break the numbers apart by place value in your head. Do not try to hold the entire sum in working memory. For something like 487 plus 364, think 400 plus 300 equals 700, then 80 plus 60 equals 140, then 7 plus 4 equals 11. Add those mentally: 700 plus 140 is 840, plus 11 is 851. You do this in about four seconds. The answer options on the test will usually include 841 and 861 as distractors, so if you are sloppy with your place-value carry, you will pick the wrong one and waste another twenty seconds figuring out why.
For multiplication, look for friendly numbers. 47 times 53 is annoying until you notice it is 50 squared minus 3 squared, which is 2500 minus 9, or 2491. You should spot these patterns quickly if you have practiced them. The test authors know this, and they include several of these near-square problems specifically to reward people who recognize the difference of squares shortcut. If you just multiply digit by digit in your head, you will make an error and move on angry, which is exactly what they want. One specific problem I ran into that catches people out is a subtraction round where the numbers look like they require borrowing across multiple columns, but the answer choices are spaced so tightly that any small carry mistake sends you to the wrong option. The exact question was something like 1024 minus 678 with options ranging from 344 to 356. The trick here is not to do the borrowing vertically in your head. Subtract 678 from 1000 first—that is 322. Then add the remaining 24. You get 346. If you tried to borrow across three columns digit by digit, you would likely fumble the middle zero and end up at 344 or 356. This is the kind of question where the workaround is to reframe the operation entirely rather than brute-force it.
Get the Full Details

Common mistakes candidates make
The biggest issue is option clustering. The answer choices are deliberately close together—sometimes only five or ten apart. This forces precision. You cannot eyeball an approximate answer and feel confident. If you are guessing, you are guessing between four numbers that are nearly identical, which means your error rate skyrockets even if your raw math skill is fine. Another mistake is trying to maintain a steady pace from the start. The first round is usually easy, and candidates get complacent. By round three, when the multiplication and division questions appear, they are already fatigued. The questions do not get harder. The candidate does. Sleep, stress, and even just the rhythm of clicking through questions affects your calculation accuracy. I have seen people who could do mental math quickly before the test start making errors on simple problems by question thirty because their attention had dropped. A third issue is not using the answer options as a tool. When you see four options that end in different digits, the last digit of your calculation should match the last digit of the correct answer. If your mental math says the result ends in 3 but three of the four options end in 7, you already know you made a carry error somewhere. This can save you from picking the wrong answer and spending extra time verifying it.
What the test cannot measure
The Optiver Mental Math Test is a filter, not a complete assessment. It tells you whether someone can handle basic arithmetic under time pressure. It does not tell you anything about market intuition, risk management, pattern recognition, or the ability to make decisions with incomplete information. Candidates who score very high on this test sometimes struggle later in the selection process because the subsequent stages demand different skills. Conversely, people who score in the middle bands often perform well in the interviews and trading simulations that follow. The test also has a ceiling. Once your mental math speed tops out around thirty seconds per question, grinding more practice questions gives diminishing returns. The real improvement comes from learning the shortcuts—difference of squares, breaking multiplication into distributive steps, reframing subtractions away from borrowing. These techniques cut your per-question time down from roughly thirty seconds to fifteen or twenty seconds on the easier items, which frees up mental capacity for the harder questions later in the round. Practice with the PDF for about a week, timing yourself. Aim to finish the forty questions in under fifteen minutes with a high accuracy rate. That leaves buffer time for the actual test where things will feel slower due to the pressure. If you are finishing comfortably in twenty minutes during practice, you are in the right range.