Quant Playbook: Solve Like a Manager, Not a Mathematician

What this section really tests

21 Problem Solving questions, 45 minutes, five options each, no calculator. The syllabus is deliberately narrow — arithmetic and algebra, dressed as word problems, number properties, rates, percents, sets and light statistics. There is no advanced mathematics, and that's the point: with the content held simple, the test can measure how you reason — and how efficiently.

The single biggest mindset upgrade: stop solving like a math student and start solving like a manager under deadline. The question is never "can you produce the textbook derivation?" but "what is the cheapest legitimate route to the answer?"

The efficiency toolkit

🛠️ Move	When it wins	The habit
Estimate first	Answer choices are spread apart	Round aggressively, compute the ballpark, kill non-contenders before doing exact work
Test the answers	"Which value of x…" / integer answers	Start from option (c); the result tells you which direction to move
Plug smart numbers	Variables in the choices, percents/fractions of unknowns	Choose 100 for percents, LCM-friendly values for fractions
Spot the structure	Symmetric expressions, differences of squares, telescoping	30 seconds of pattern-looking before 3 minutes of grinding
Do algebra last	Everything else failed	Textbook methods always work — they're just usually the slow lane

Because there's no calculator, the numbers are a message from the question-writer: they were chosen to be tameable. If your path produces brutal arithmetic, you're almost certainly on the wrong path — back up and look for the intended structure.

Scan the choices before you compute

The five options are part of the question, not an afterthought:

• Wide spread → estimation will be decisive.
• Pairs that differ by a sign or an inversion (x vs −x, ⅖ vs ⁵⁄₂) → the writer has booby-trapped a specific slip; name it and step around it.
• "Easy" numbers among ugly ones → test the easy ones first when back-solving.

The whiteboard is half your score

Mental math errors — not concept gaps — sink most Quant sections. Use the (physical or online) whiteboard relentlessly and identically in practice:

• ✍️ Write the target at the top ("max value of n", "% change") — mid-question goal amnesia is a classic killer.
• Organize work in columns, not a spiral of scribbles; carrying a wrong intermediate forward is the most expensive error type.
• If you're testing options, make a tiny table: option → outcome → verdict.
• Before exam day, rehearse on the same medium you'll use — the online whiteboard feels different from paper, and you don't want to discover that live.

Traps with names

1. The literal-answer trap — you solved for x; the question asked for 2x + 3. Re-read the final sentence before confirming. Always.
2. The unit shear — minutes vs hours, metres vs km, per-person vs total. Annotate units on the whiteboard.
3. The integer assumption — "x is a number" does not mean integer; "people/machines/marbles" does. The constraint (or its absence) is usually where the difficulty lives.
4. The favourable-rounding spiral — estimation is for eliminating, not for picking between two close options; when two survivors are close, switch to exact arithmetic on just those two.
5. The phantom-knowledge trap — using a half-remembered formula where basic reasoning suffices. If you can't re-derive it in ten seconds, reason instead.

Pacing notes specific to Quant

~2:09 per question. The hard ones are usually hard because the setup is hard, not the computation — so the 60-second rule applies cleanly: if after a minute you have no representation of the problem (equation, table, picture), eliminate, answer, 🔖 bookmark, and let Review & Edit (up to three changes) handle redemption.

How to drill

• Run topic sets in our question bank, untimed for method first, then timed for speed — in that order.
• Maintain an error log with a one-line "named trap" per miss; review it before every mock.
• In every review, ask: what was the 30-second route? If your correct answer took four minutes, that question still goes in the log.