GATE (Core Branches) Exam Survival Guide

Window & Pattern: 3 hours, CBT, ~65 Q / 100 marks; sections: General Aptitude (15 marks) + Engineering Mathematics (~13 marks for most papers) + subject. Question types: MCQ (−1/3, −2/3), MSQ (no negative), NAT (no negative). Papers list & weight split vary by paper. 

Must-do (80/20 across branches)

• Gen Apt (GA): vocab-in-context, basic RC, arithmetic %, ratio, averages, data/graphs.
• Engg Math: limits/continuity, derivatives/integration (standard forms), linear algebra (system rank/eigen), probability (Bayes/binomial), complex numbers, ODE basics, numerical methods (Newton, trapezoid).
• DI/Units: dimensional checks, order-of-magnitude, log/semilog reading.
• Branch-agnostic skills: interpreting constraints, monotonicity, bounding, quick sketches.

Top traps (and fixes)

• Marking: treating MSQ like MCQ (there’s no negative for MSQ/NAT → attempt thoughtfully). 
• Algebra: squaring ⇒ extraneous roots; forgetting inequality flips.
• Calculus: wrong substitution bounds; mixing average rate vs slope.
• Probability: independence vs disjoint; wrong base in Bayes.
• Graphs: ignoring axis units / log scales; reading totals as rates.

Time split (3 h)

• Pass-0 (2–3 min): skim paper; flag NAT/MSQ freebies.
• Pass-1 (75–85 min): GA + Engg Math + single-step subject Qs.
• Pass-2 (50–55 min): medium multi-step Qs; 1–2 scoring MSQ/NAT clusters.
• Pass-3 (15–20 min): tough picks, sanity checks, units audit.

Last-48h checklist

• GA: 30 one-liners (RC + %/ratio + table/line/pie) with TAUSL (Title–Axes–Units–Scale–Legend).
• Math: 60 mixed (20 calc, 15 LA, 15 prob/stats, 10 numerics); re-do only the wrongs with a why-wrong tag.
• Branch core: 4 prior-year micro-topics you miss most (10 Q each).
• Run an official mock UI once (navigation + on-screen calc muscle memory). 

Quick frames

• Dimensional kill: eliminate any option with wrong dimensions/units first.
• Approx first: bound integrals/roots; refine only if needed.
• Linear algebra: check rank vs unknowns for consistency quickly.
• Probability: write sample space; Bayes: prior × likelihood → normalize.

Speed tactics

• Prefer ratios/normalization over raw arithmetic.
• Back-solve for algebra/optimization; test extrema/bounds.
• On MSQ: check each option independently; partial selection is allowed; no penalty. 

Exam-day mini-plan

• 3-min warm-up: 2 GA + 2 Math + 1 subject single.
• Keep a live error tag: [Units] [Sign] [Scope] [Read].
• Last 5–7 min: revisit flagged NAT/MSQ; unit/tolerance sanity pass.