GATE's numerical-heavy paper rewards problem mileage. One formula sheet per subject, 15 timed problems a day, and every error logged the same evening.
GATE is 3 hours of 65 questions worth 100 marks: 15 marks of General Aptitude (verbal + quantitative) plus two subject papers, split into a 45-question Part A (one-paper format) — the current structure rolls your two subject papers into a single 3-hour CBT. Question types: 1-mark and 2-mark MCQs, 1-mark and 2-mark NATs (numerical answer type — no negative marking, which makes them the highest expected-value questions in the entire Indian exam calendar), and a small set of multiple-correct with full marks only for exact attempts and 1/4 negative on wrong single-correct. The scoring twist that changes strategy: GATE normalises across papers — your raw score is adjusted so that top performers in each branch are comparable (the 'top-10% rule' of normalisation). In plain terms: your target is not 70 marks, it's being in the top cohort of your branch — which means you should study to the branch's known high-yield clusters, not to an imagined average.
Three layers: (1) the 15-mark GA — cheap, consistent, and where weak candidates give away 5+ marks; (2) the 'core 60%' — the 30–40 questions that repeat the same problem families year after year (every branch has its permanent high-yield chapters; in CSE it's GfK + Algorithms + DBMS + OS + COA + Networks; in EE it's Networks + Signals + Control + Machines; in ME it's SOM + Thermodynamics + Fluids + Manufacturing); (3) the 'differentiator 15%' — 5–8 questions designed to separate 700 from 900 ranks, usually one deep multi-concept problem per subject. GATE rewards breadth-with-depth: you don't need to be a 99-percentile expert in your branch, you need 85%+ on the core 60% and GA, and you must not panic on the differentiators. The exam is a distribution test, not a knowledge ceiling test — and distributions are predictable from PYQs.
The plan assumes 2 focused hours on weekdays + 4–5 on weekends — about 900 hours total, which is sufficient for a serious first attempt. Months 1–2 — core syllabus rebuild: one chapter per 3–4 days, each chapter = concept pass + 50 PYQs (2015–2026) graded by difficulty, errors into a log. Months 3–4 — full PYQ sweep: every past paper's questions solved under time, chapter-by-chapter, with a 'year of last appearance' tag per question family (families that appear in 4+ of the last 10 years are mandatory; 2–3 years are high-yield; 0–1 years are last). Month 5 — mock season: 8–10 full 3-hour mocks (standard test-series + previous GATE papers), each analysed for 90 minutes; the NAT accuracy target is 80%+ (no negative marking — there is no strategic reason to miss them). Month 6 — final 30 days: formula + 1-page-per-subject sheets, 50-question mixed sets daily at exam pace, GA locked (13–15/15 target), mocks taper to one per week, sleep protected. The working-professional rules: the weekday block is fixed (6–8 AM beats the evening — the evening loses to fatigue and group chat), Sunday is a mock day, and a skipped weekday is made up by a Saturday half, never silently.
2 focused hours on weekdays + 4–5 on weekends, consistently for 6 months (~900 hours), is a serious first-attempt budget. The non-negotiables: a fixed weekday block (morning wins), one mock Sunday per week, and the full two-pass PYQ sweep.
Your raw score is adjusted against your branch's top cohort, so the target is percentile, not marks. Study the branch's known high-yield clusters (from PYQ frequency), maximise NAT accuracy (no negative marking), and bank GA — the percentile difference at the top is made of 2–3 marks, all of which are plan-able.
PYQs are the strategy's backbone but not the entire syllabus: use them to map question families and difficulty trends, then close any concept gaps with one standard book per subject. The two-pass PYQ method (timed, then as families) is worth more than any third-party question bank.
Start focusing free