Why does answer writing — not knowledge — decide your UPSC Mains marks?

TL;DR

Every Mains topper knows the same content as 500 other aspirants. What separates a 110/250 GS paper from a 90/250 paper is how that content lands on paper: structure, directive-word fidelity, value-addition, and presentation. UPSC evaluators get ~7-8 minutes per script — they reward clarity, not erudition.

The brutal truth about Mains evaluation

If you spoke to any UPSC examiner, they would tell you the same thing: by the time they reach script number 200 on a given day, they are scanning for structure, keywords, and presentation — not deep philosophical insight. A candidate who knows the Directive Principles inside out but writes a wall-of-text paragraph will lose to one who knows half as much but uses subheadings, underlines key terms, and answers the exact directive word asked.

Why this happens

  • Content is roughly 60% of marks, presentation is 40%. That 40% is where you actually beat the competition — because everybody has the content.
  • The same syllabus, the same textbooks (Laxmikanth, Ramesh Singh, Shankar IAS). When 10,000 candidates write on Article 356, the differentiator is who frames it as federalism + judicial review + S.R. Bommai (1994) + a way-forward line, versus who just dumps facts.
  • Time pressure is real. 20 questions in 180 minutes leaves exactly 9 minutes per question — and that includes thinking, structuring, and writing. Aspirants who never practised under timed conditions write beautiful first answers and rushed gibberish for the last five.

The arithmetic of a 3-hour GS paper

Component10-marker (×10)15-marker (×10)Total
Marks100150250
Words150 each250 each~4,000
Pages on UPSC booklet~1 page~1.5 pages~12-14 pages
Ideal time per Q7 minutes11 minutes180 min
Buffer time0-3 min

The math is brutal. If you spend 12 minutes on a 15-marker, you have stolen 1 minute from another question. Five such overruns and you skip an entire question — losing 10-15 marks not from ignorance, but from arithmetic.

What CSE 2024 GS-2 actually demanded

The September 2024 GS-2 paper carried 20 questions (10 × 10-marker + 10 × 15-marker), with around 11 questions linked to current affairs (Vision IAS analysis). Of the analytical questions, the directive distribution skewed toward "discuss" and "examine" (the most frequent), followed by "critically examine," "comment," and "evaluate." Questions like "What changes has the Union Government recently introduced in the domain of Centre-State relations? Suggest measures..." combined description + suggestion — a structure most candidates botched by forgetting the second half.

What a 'good' answer actually looks like

A 10-mark question on cooperative federalism gets a 2-line intro defining it via Article 263, a body with 3 dimensions (institutional — Inter-State Council; fiscal — GST Council; policy — NITI Aayog), a flowchart if time permits, and a 2-line forward-looking conclusion citing the Punchhi Commission. Same content, structured presentation — that is the difference between rank 800 and rank 80.

What CSE 2026 candidates are racing toward

The CSE 2026 Notification was released on 4 February 2026 with ~933 vacancies. Prelims is on 24 May 2026 (Sunday) and Mains begins on 21 August 2026 (Friday) — 9 descriptive papers across 5 consecutive days. That gives Mains aspirants approximately 90 days between Prelims result and Mains — barely enough time to write 200 fresh answers if you have not built the muscle through Prelims prep. The candidates who clear Mains 2026 are already writing 2 answers daily as of May 2026, not waiting for Prelims results.

The Shubham Kumar test (AIR 1, CSE 2020)

In his widely circulated GS strategy note, Shubham Kumar — IIT Bombay civil engineering graduate who cleared on his third attempt — wrote: "Try to keep your answers in points and very precise. Focus on conveying your ideas rather than just filling pages." He took 1-hour mock tests daily during Mains prep and a full 3-hour mock every third day. That is roughly 30 answers per week for 4 months — about 480 timed answers before the real paper.

Shruti Sharma's contrast — same syllabus, different presentation

Shruti Sharma (AIR 1, CSE 2021) scored 1105/2025 — and her marksheet has no single outstanding paper and no weak paper. She did not master one paper; she eliminated weak ones. Her widely circulated GS answer copies (released by Forum IAS and Vajiram) show consistent structure across all four GS papers: 2-3 sub-headings even in a 10-marker, short direct sentences, sparing use of quotes (reserved for essay), and a forward-looking conclusion every time. The lesson: presentation is the system that scales across all four GS papers; content varies by topic, but the system does not.

The marks delta — content vs. presentation

Candidate typeKnowledge depthStructure disciplineLikely GS score (out of 250)
Read everything, wrote 30 answersHighLow75-90
Read selectively, wrote 200 answersMediumHigh95-115
Read everything, wrote 200 answersHighHigh110-130
Read selectively, wrote 30 answersMediumLow60-75

The biggest delta is not between low-knowledge and high-knowledge — it is between low-practice and high-practice. UPSC rewards trained execution, not raw knowledge.

Mentor takeaway

Stop reading more. Start writing more. A candidate who has revised Laxmikanth twice and written 200 answers will outscore one who has revised it five times and written 20. Knowledge is the price of entry. Answer writing is the game.

Sources: · · · ·

What is the ideal 150-word answer structure for a 10-mark UPSC question?

TL;DR

2 + 10 + 2. Intro of 20-25 words (≈2 lines), body of 100-110 words (≈10 lines, 2-3 thematic sub-headed paragraphs or bullets), conclusion of 20-25 words (≈2 lines). Target time: 7 minutes. UPSC accepts ±10-15% on word count, so 135-165 words is the safe band.

The 2-10-2 blueprint

A 10-mark answer is a sprint, not a marathon. You have roughly 7 minutes and 150 words. Every word must earn its place.

Intro (20-25 words / ~2 lines)

Define the core concept or contextualise with a recent event/data point. Skip the dictionary definition out of context and ban phrases like "Since time immemorial" or "In today's globalised world."

Good intro on Article 32: "Article 32 — termed the 'heart and soul of the Constitution' by Dr. B.R. Ambedkar — empowers the Supreme Court to issue writs for the enforcement of fundamental rights."

That is 28 words. One quote (with attribution), one constitutional anchor, one functional definition. Done.

Body (100-110 words / ~10 lines)

For a 10-marker, aim for 2-3 thematic sub-points, each with a mini-heading or underlined keyword, followed by 2-3 lines of explanation. Use bullet points if the question is descriptive ("enumerate," "list") and short paragraphs if analytical ("discuss," "examine").

Within the body, hit at least one credibility marker: a committee name (2nd ARC, Punchhi), a case (Kesavananda Bharati 1973), a scheme with year (PM-KISAN, 2019), or a data point (Economic Survey 2024-25).

Conclusion (20-25 words / ~2 lines)

Forward-looking. Suggest a reform, cite the SDG target, or invoke a constitutional ideal. Never repeat the intro.

Good conclusion: "Strengthening judicial review under Article 32, alongside the 2nd ARC's recommendation for grievance redressal, can transform rights into lived realities."

The word-limit tolerance bands

UPSC publishes a target word count below each question but does not penalise mechanically. Based on multi-year coaching aggregate analysis, examiner tolerance follows this pattern:

TargetLower safe (–10%)Upper safe (+10%)Caution zoneMarks impact
150 words135165165–180Mild — examiner skims
150 words>180Above 180Loss of focus marks (~1-2)
250 words225275275–290Padding penalty risk
250 words>290Above 290Examiner stops reading body

Underwriting (below the lower band) hurts more than overwriting in moderation — fewer words signal under-prepared content. The optimal target for a 10-marker is 145-155 words.

The marks-per-word economy

10-marker (150 words)Marks per word
Intro (25 words → 1.5 marks)~0.06
Body (105 words → 7 marks)~0.067
Conclusion (20 words → 1.5 marks)~0.075

The conclusion is your highest-paying real estate per word — and yet it is what 60% of candidates botch. Treat it as gold.

What Shruti Sharma (AIR 1, CSE 2021) did differently

Shruti Sharma — who scored 1105/2025 in CSE 2021 (second attempt) — built her brand on short, direct points and clear structure that reads quickly. Her widely studied answer copies (released by Forum IAS / Vajiram) show 2-3 sub-headings even in a 150-word answer, with each sub-point taking 3-4 lines maximum. She did not write more; she wrote more visibly.

In her Vajiram answer-writing session, she emphasised: "My answers had no decorative language. No attempt to sound impressive. They read like someone who understands governance, not someone trying to perform understanding." For a 10-marker, that translates to: cut every adjective that does not earn its presence.

Time-on-the-clock breakdown for a 10-marker

MinuteActivity
0:00–0:30Read question, underline directive word + scope words
0:30–1:30Rough-sheet skeleton: 3 sub-points + 1 fact-anchor per point
1:30–2:30Write intro (25 words)
2:30–5:30Write body (105 words across 2-3 sub-headings)
5:30–6:30Write conclusion (20-25 words, forward-looking)
6:30–7:00Underline keywords, box diagram if drawn, move on

If you cross 7:30 on a 10-marker, stop and move on — the marginal mark from finishing is less than the marginal loss from skipping a question later.

The directive-marks calibration for 10-markers

Most 10-markers in CSE 2020-2024 carried descriptive or moderate-analytical directives — "discuss," "examine," "comment," "elucidate." Heavyweight directives ("critically analyse," "evaluate") tend to attach to 15-markers. This means your 10-mark answer can usually skip exhaustive merit-demerit treatment and go straight to describe + 2-3 examples + one forward line.

Directive in 10-markerBody structureSub-headings needed
DiscussFor + Against in 50:502 (Pros, Cons)
ExamineCauses + Effects2 (Why, So-what)
CommentOpinion + Reasoning + Caveat2-3
ElucidateDefinition + Examples + Significance3
Enumerate / ListDirect bullet enumeration0 (bullets only)

A real 10-marker walk-through

"Discuss the role of the Inter-State Council in Indian federalism." (10 marks, 150 words)

Intro (24 w): The Inter-State Council, constituted under Article 263 on the Sarkaria Commission's (1988) recommendation, is India's principal forum for resolving Centre-State and inter-State disputes.

Body (105 w):

Composition & mandate — Headed by PM; CMs, six Union Cabinet Ministers, lieutenant governors as members. Mandate covers inquiry into inter-State disputes, policy coordination, and inter-State subjects.

Working & limitations — Met only 13 times since 1990 (against the recommended thrice-a-year schedule). Standing Committee active; sub-committees thinly empowered. Punchhi Commission (2010) flagged irregular meetings as a structural weakness.

Recent revival — President-led reconstitution (2022) signals renewed federal dialogue.

Conclusion (21 w): Operationalising the Inter-State Council quarterly, as Punchhi recommended, can transform Article 263 from dormant text into living cooperative federalism.

That is 150 words flat, 2 sub-headings, 3 credibility markers (Article 263, Sarkaria 1988, Punchhi 2010), one numerical anchor (13 meetings), and a forward-looking conclusion.

Mentor tip

Draw two horizontal lines on your answer sheet — one after the intro, one before the conclusion. This forces the examiner's eye to register the structure within the first 3 seconds. You score before they read a single sentence of body.

Sources: · · · ·

How do you structure a 250-word, 15-mark answer with multi-dimensional depth?

TL;DR

Intro (30-35 words) + Body (180-200 words across 3-4 thematic dimensions) + Way Forward (30-35 words). Target time: 11 minutes. The 15-marker is where multi-dimensionality, data, and a diagram/flowchart actively buy you marks.

The 30-200-30 framework

A 15-mark question is where examiners separate the rank-100 candidates from the rank-1000 ones. You have more real estate (250 words, 11 minutes), so you must show range, not just recall.

Intro (30-35 words / ~3 lines)

Open with one of three hooks:

  1. Data hook — "India loses ~16% of its GDP annually to air pollution (World Bank, 2023)..."
  2. Quote hook — sparingly, only if it fits the directive word.
  3. Constitutional/legal anchor — "Article 21, as interpreted in M.C. Mehta v. Union of India (1987)..."

Body (180-200 words / ~15-18 lines)

This is the heart of the answer. Break it into 3-4 dimensions using PESTLE-lite framing:

  • Political/Institutional — schemes, ministries, constitutional provisions
  • Economic — Budget allocation, Economic Survey data, sectoral impact
  • Social — caste/gender/regional disparity, NFHS data
  • Environmental — sustainability angle, IPCC/CPCB linkage
  • Ethical/Technological (where relevant)

Each dimension gets a bold sub-heading, 3-4 lines of content, and at least one source-cited fact ("NITI Aayog SDG Index 2023-24," "15th Finance Commission," "Justice B.N. Srikrishna Committee").

If the topic permits — climate, federalism, judicial pendency, urbanisation — insert one small flowchart or map. Toppers use visuals once per 15-marker, never on every page.

Way Forward (30-35 words / ~3 lines)

This is non-negotiable for 15-markers. Pivot from analysis to action: cite a committee recommendation (2nd ARC, Sarkaria/Punchhi, Justice Verma, M.S. Swaminathan), an SDG target, or a constitutional ideal (Article 38 socio-economic justice).

Worked scenario — CSE 2024 GS-2 federalism question

Actual UPSC 2024 question (15 marks, 250 words): "What changes has the Union Government recently introduced in the domain of Centre-State relations? Suggest measures to be adopted to build the trust between the Centre and the States and for strengthening federalism."

Skeleton answer (250 words):

Intro (32 words): Indian federalism, characterised by Granville Austin as a system of "cooperative federalism," has witnessed institutional and fiscal recalibration since 2014 — shifting from competitive to competitive-cooperative paradigms anchored in Articles 245-263.

Body — Recent Union Government changes (90 words):

Fiscal & institutional — Replacement of Planning Commission with NITI Aayog (2015) gave States a Governing Council voice; 14th and 15th Finance Commissions raised vertical devolution to 41-42% of divisible pool.

Cooperative platforms — GST Council under Article 279A operationalised "pooled sovereignty"; institutionalised meetings of Chief Secretaries (since 2022) and the National Conference of Governors.

Friction areas — Centralisation of cesses/surcharges (outside divisible pool, ~15-18% of gross tax revenue), All-India Services postings disputes, Governor's discretion (Tamil Nadu vs. Punjab references, 2023), CBI consent withdrawals.

Body — Measures to build trust (78 words):

  1. Punchhi Commission (2010) — implement recommendations on Article 355/356, Inter-State Council activation under Article 263.
  2. Sarkaria Commission — consult States on All-India Services postings.
  3. Fiscal — cap cesses/surcharges via constitutional amendment; revive Article 282 grants transparently.
  4. Institutional — quarterly Zonal Council meetings (currently irregular); judicialise Governor's discretion per Sarkaria's threshold tests.

Way Forward (30 words): As B.R. Ambedkar noted, the Constitution is "federal in normal times, unitary in emergency" — strengthening trust requires operationalising Article 263 Inter-State Council and rule-bound asymmetric federalism.

Scoring rubric for the above answer (out of 15)

ParameterWeightThis answer scores
Directive fidelity (described and suggested)33/3
Multi-dimensionality (fiscal + institutional + political)33/3
Value-addition (Punchhi, Sarkaria, GST Council, NITI Aayog, articles)33/3
Data/factual anchors (41%, 14th/15th FC, Article 279A, 263, 355)22/2
Structure (intro + 2 sub-sections + way-forward)22/2
Conclusion forward-looking21.5/2
Total1514.5/15

A typical aspirant scores 7-8/15 on this question because they describe only the changes (first half) and forget suggestions (second half). Reading the question twice — slowly — is worth 2 marks more than any value-addition trick.

The 11-minute clock for a 15-marker

MinuteActivity
0:00–0:45Read question twice, underline directive + scope, identify if compound ("and suggest...")
0:45–2:00Skeleton on rough sheet — 4 dimensions + 1 anchor each
2:00–3:15Write intro (32 words, one hook)
3:15–8:00Write body — 4 sub-headings × 3 lines each
8:00–9:00Draw diagram if relevant (box + label)
9:00–10:30Write way-forward (32 words)
10:30–11:00Underline keywords, scan for word-count compliance

If you cross 12 minutes consistently, your speed is the bottleneck — not your knowledge. Drill timed writing 30 times and the clock self-corrects.

Mentor tip

If you cannot honestly write 3 dimensions + 1 data point + 1 committee/case + 1 way-forward line, you do not know the topic well enough — go back and re-read. The structure is a diagnostic, not a decoration.

Sources: · · · ·

What do directive words like 'discuss', 'examine', 'critically analyse', 'elucidate', and 'evaluate' actually demand?

TL;DR

Misreading the directive is the #1 reason aspirants lose marks despite knowing the content. Discuss = debate both sides. Examine = establish facts. Critically analyse = break down + judge with evidence. Elucidate = explain in depth. Evaluate = pros + cons → verdict.

Directive words are instructions, not decoration

UPSC's question stem is a contract. If the question says "critically examine" and you only describe, you've breached the contract — examiner caps your marks at 50% regardless of content quality.

The directive frequency table — CSE 2020-2024

Aggregated across GS-2 and GS-3 papers from 2020 to 2024, the directive distribution looks like this:

DirectiveApprox. frequencyMarks-pattern (most common)What it demands
Discuss~25-28%10 + 15 mixBoth sides, balanced
Examine~15-18%10 + 15 mixEstablish facts, cause-effect
Critically examine / analyse~14-16%Predominantly 15Strengths + weaknesses + verdict
Comment~8-10%Usually 10Informed opinion
Evaluate / assess~6-8%Usually 15Weigh + verdict
Elucidate / explain~6-8%Usually 10Detailed exposition
Suggest / recommend~4-6%Part of 15-markerAction-oriented
Justify / substantiate~3-5%EitherDefend with evidence

In CSE 2024 GS-2 specifically, "discuss" and "examine" dominated, with "critically examine" appearing on the heaviest-weight analytical questions. Around 11 of 20 questions had explicit current-affairs anchoring (Vision IAS analysis).

The five directives that appear most often

Discuss

A debate. Present arguments for and against, weigh them, arrive at a balanced conclusion. Example: "Discuss the relevance of Panchayati Raj institutions in 2025." → Pros (decentralisation, 73rd Amendment), cons (capacity gaps, parallel bodies), verdict.

Examine

Establish facts. Less judgmental than "critically examine." Focus on what is, with cause-effect linkage. Example: "Examine the working of the GST Council." → Structure, voting weights, friction points — described, not judged.

Critically analyse / Critically examine

The heavyweight directive. Break the topic into components, examine each, deliver a judgment backed by evidence. Both strengths and weaknesses must appear — a one-sided answer is auto-capped. Use phrases like "While X is true, Y limits its effectiveness because..."

Elucidate

"Make clear, explain in detail." More expository than argumentative. Used for conceptual questions ("Elucidate the doctrine of basic structure"). Build the answer like a layered explanation: definition → origin (Kesavananda Bharati, 1973) → evolution (Minerva Mills, 1980; I.R. Coelho, 2007) → contemporary relevance.

Evaluate

Weigh merits and demerits to arrive at an overall verdict. Similar to critically analyse, but with a stronger expectation of a final judgment line. Example: "Evaluate the impact of MGNREGA on rural livelihoods." → Achievements (wages, women's participation), shortcomings (delayed payments, asset quality), verdict.

Two more worth knowing

  • Comment — share an informed opinion with reasoning.
  • Substantiate — defend a given statement with evidence; you must agree, then justify.

The directive-marks-format matrix

DirectiveLikely marksWord targetStructure
Discuss10 or 15150 or 250Intro → For → Against → Balanced verdict
Examine10 or 15150 or 250Intro → Components → Cause-effect → Conclusion
Critically examineAlmost always 15250Intro → Merits → Demerits → Judgment + Way forward
ElucidateUsually 10150Intro → Layered explanation → Contemporary relevance
EvaluateUsually 15250Intro → Pros → Cons → Final verdict + Way forward
CommentUsually 10150Intro → Your reasoned opinion → Caveat → Forward line
SuggestPart-question, 15-marker80-120 of the 250Bullet-form, action-oriented

What Anudeep Durishetty (AIR 1, CSE 2017) emphasised

Anudeep — who cleared on his 5th attempt and now publishes a widely-read essay/answer-writing primer — argues that the single highest-leverage habit is reading the question stem twice. On his blog he writes: "Present both the positive and negative side of the subject... use subheadings to make points clear and important... if data wasn't given in the introduction, suffice arguments with data in the body." The implicit instruction: the directive word dictates whether you balance, describe, or judge — never assume.

Compound directives — the silent killer

UPSC increasingly asks compound questions with two directives glued together: "Discuss... and suggest...", "Examine... and evaluate...", "Critically analyse... in light of...". The CSE 2024 GS-2 federalism question ("What changes... Suggest measures...") is the canonical recent example.

Compound typeExampleBoth halves must appear
Describe + Suggest"What is X? Suggest reforms."50:50 split of body
Examine + Way forward"Examine challenges... way forward."60:40 (analysis-heavy)
Discuss + Comment"Discuss... Comment on its impact."60:40 (description-heavy)
Critically examine + In light of"Critically examine X in light of Y."Y is the lens, not a second part

The rule: count the verbs in the question stem before writing. One verb = simple directive. Two verbs = compound — body splits 50:50 or 60:40 depending on weight.

The 5-second directive drill

Every time you sit down to write, do this in 5 seconds before any pen-on-paper:

  1. Read the question stem aloud (in your head if in exam hall).
  2. Underline every verb ("discuss," "suggest," "examine," "evaluate").
  3. Mark the scope ("in India," "since 2014," "in the digital age").
  4. Write one sentence on your rough sheet — "The question wants me to ___ and ___."

This is the single highest-leverage 5 seconds in the whole answer. Skip it and you risk an autocap on directive miss; do it and you guarantee directive fidelity.

Mentor tip

Underline the directive word in the question paper before you start writing. Then write one line at the top of your rough sheet: "This question wants me to ____." That 10-second discipline saves entire marks. For "critically" directives, force at least one explicit pivot sentence: "However, this institutional design suffers from three structural limitations..."

Sources: · · · ·

How do you write a multi-dimensional answer — covering political, economic, social, environmental, and ethical angles?

TL;DR

UPSC questions are rarely uni-dimensional. A topic like urbanisation has political (74th Amendment), economic (informal sector), social (migration), environmental (urban heat island), and ethical (right-to-the-city) dimensions. The PESTLE framework — or the simpler 7-5-3 rule (7 min, 5 points, 3 dimensions each) — gives you a reusable scaffold.

Why one-dimensional answers cap your marks

UPSC explicitly tests integrative thinking. A question on "renewable energy transition" that you answer only economically (cost, investment) will score 5/10. Add land acquisition (political), displacement (social), grid stability (technological), intergenerational equity (ethical), and you score 8/10 — same word count, more dimensions.

The PESTLE-E scaffold

Memorise these six lenses and apply them to every issue:

  • Political/Constitutional — Articles, amendments, schemes, ministries, federalism angle
  • Economic — Budget %, Economic Survey data, sectoral impact, GDP linkage
  • Social — Caste, gender, region, demographic, NFHS-5 data
  • Technological — Digital divide, automation, AI, infrastructure
  • Legal/Ethical — Constitutional morality, judicial pronouncements, ethical principles (utilitarian vs. deontological)
  • Environmental — Climate, biodiversity, IPCC linkage, sustainability

You will not use all six in every answer — pick the 3-4 most relevant to the question.

The 7-5-3 rule for execution

Popularised by topper interviews: in 7 minutes (for a 10-marker), write 5 substantive points, each elaborated through 3 dimensions. For a 15-marker, scale to 11 minutes / 6 points / 3-4 dimensions.

Worked example — "Discuss the implications of farm-to-fork supply chains." (15 marks)

  • Economic — Reduced intermediaries, 15-25% farmer income gain (NABARD studies).
  • Social — Empowers FPOs, reduces distress migration; but excludes smallholders without digital access.
  • Environmental — Cold-chain energy intensity, food-mile reduction trade-off.
  • Political/Institutional — PM Kisan Sampada Yojana, e-NAM, APMC reforms.
  • Ethical — Right to food (Article 47), corporate concentration vs. farmer autonomy.

Five dimensions, each with a fact-anchor — examiner cannot deny full marks.

The dimension-density curve

Topper answer-copy analyses (Shruti Sharma 2021, Shubham Kumar 2020, Srushti Deshmukh 2018) show a consistent pattern: their 15-mark answers average 3.5-4 dimensions, with 2-3 sentences per dimension. They do not try to cram all six PESTLE lenses — they pick the ones with strongest factual anchors.

Aspirant tierAvg. dimensions / 15-markerAvg. fact-anchors
Rank 1-1003.5–44–5 (cases, data, committees)
Rank 101-50032–3
Rank 501-100021–2
Below rank 10001–20–1

The top tier is not writing more — they are anchoring each dimension with one verifiable fact, which signals depth.

What Shubham Kumar (AIR 1, CSE 2020) actually wrote

In his GS strategy note (posted September 2023 on his Telegram channel and widely reposted), Shubham Kumar advised: "While reading the question paper (before start of examination, papers are distributed 2-5 minutes before) — try to link questions with your short notes of value addition. If you are able to remember committee name/articles/case studies — try to mark it on question paper itself."

Subject-specific dimension-building from his notes:

  • History/Art-Culture — "enough scope to make maps, diagrams like Ashokan pillar, Stupa, Temples"
  • Geography — "make diagrams in most of the questions"
  • International Relations — "political diagram must, historical dimension, use of key words (like track 2 diplomacy), G20"
  • GS-3 Economy — "Use of Graphs/pie charts (which I had prepared beforehand) and statistics: economists love stats/data"
  • Disaster Management — "Use of diagrams and maps, best practices"
  • Security — "use of bodies/Acts/Institutions"

Notice that each subject area maps to a different dominant dimension — history to visual/cultural, economy to data, IR to historical-political. Multi-dimensionality is not abstract; it is subject-specific shorthand that you internalise through 100+ practice answers.

The dimension-to-paper mapping

Not every paper rewards every PESTLE lens equally. Use this paper-specific guide:

GS PaperStrongest dimensionsWeakest dimensions
GS-1 (History, Geog, Society)Social, Cultural, Environmental, HistoricalEconomic-fiscal (limited)
GS-2 (Polity, IR, Governance)Political, Legal, Institutional, InternationalEnvironmental (occasional)
GS-3 (Economy, S&T, Env, Security)Economic, Technological, EnvironmentalCultural (rarely)
GS-4 (Ethics)Ethical, Constitutional-values, PsychologicalEconomic (avoid unless directly asked)

A worked CSE 2024 dimension audit

The 2024 GS-2 paper had a question on one nation, one election — a topic that screams multi-dimensionality. A top-tier answer covered:

  • Political — Reduces poll fatigue, frees parties for governance; but reduces State-level autonomy and accountability.
  • Constitutional — Requires amendments to Articles 83, 85, 172, 174, 356; touches federal basic structure.
  • Economic — ECI estimates ~Rs 4,500 crore savings per cycle; but one-time EVM/VVPAT cost ~Rs 10,000 crore.
  • Administrative — Frees Model Code of Conduct cycles; but logistical strain on ECI manpower.
  • Federal — Asymmetric impact — States with mid-term dissolutions lose tenure parity.

Five dimensions, each with a concrete anchor — this is what scoring 12+/15 looks like.

Mentor tip

Keep a one-page "Dimensions Cheat Sheet" in your revision folder. Whenever you read a new topic, force yourself to list its P-E-S-T-L-E lenses in writing. Within a month, this becomes muscle memory — and your answers automatically gain depth. By CSE 2026 Mains (21 August 2026), you should be able to generate 4 dimensions for any GS topic within 30 seconds of reading the question.

Sources: · · · ·

When should you add diagrams, flowcharts, and maps — and when should you skip them?

TL;DR

Use visuals once or twice per 15-marker, never on every answer. Add them when they genuinely simplify (climatic regions, judicial hierarchy, river systems) — skip them when they're decorative filler. Neat + labelled beats artistic. Visuals can buy you 1-2 bonus marks per answer, but cluttered/irrelevant ones backfire.

The visual decision tree

Before drawing anything, ask one question: "Does this diagram replace 3-4 lines of prose more clearly than I can write them?" If yes, draw. If no, skip.

When visuals genuinely help

Maps (Geography, IR, Polity-federalism)

  • Indian physiography — Himalayan ranges, peninsular drainage, Western/Eastern Ghats
  • Strategic geography — Quad, BIMSTEC, Strait of Hormuz, Indo-Pacific chokepoints
  • Resource distribution — coal belt, lithium triangle, monsoon trajectory

A rough outline map of India with 4-5 labelled features takes 90 seconds and replaces an entire paragraph.

Flowcharts (Polity, Governance, GS2/GS3)

  • Bill → Law process in Parliament
  • Judicial hierarchy — SC → HC → District → Tribunals
  • Scheme implementation chain — Central Ministry → State → District → Block → Beneficiary
  • Disaster management cycle — Prevention → Mitigation → Preparedness → Response → Recovery

Diagrams (Economy, Environment, GS3)

  • Demand-supply, Laffer curve, Phillips curve for macro questions
  • Carbon cycle, nitrogen cycle, food web for environment questions
  • Five-Year Plan focus areas, fiscal federalism flow for economy

When to skip visuals

  • Ethics (GS4) — visuals rarely add value to ethical reasoning; stick to prose + quotes.
  • Abstract polity questions — "Discuss judicial activism vs. overreach" doesn't need a diagram.
  • When you're running out of time — a half-finished, unlabelled doodle hurts more than no diagram.

The visual marks-buy table

Visual typeTime to drawReplaces wordsMarks boughtRisk if botched
Outline map of India (4-5 features)60-90 sec~40-50 words+1 to +2Low (recognisable shape)
Flowchart (3-5 boxes, arrows)45-60 sec~30-40 words+1 to +1.5Medium (label dependency)
Mind-map / spider diagram60 sec~30 words+0.5 to +1High (looks decorative)
Pie/bar chart with %s45 sec~20-30 words+1 (only if data is accurate)High (wrong data = negative marker)
Cycle diagram (e.g., DM cycle)60 sec~40 words+1 to +1.5Low
Hand-drawn graph (Phillips, Laffer)45 sec~25 words+1Medium (axes mislabel risk)

Anudeep Durishetty's diagram philosophy

Anudeep (AIR 1, CSE 2017) was famous for his clean diagrams — particularly his GS-3 economy graphs and GS-1 geography maps. In his published answer copies (compiled by upscprep.com) he uses roughly 1 diagram per 2-3 answers, never more. On his blog he writes: "Add diagrams to make answers look different and easy to understand, relate answers to ongoing current events wherever possible, and have a multidimensional approach."

The operative phrase is "look different" — diagrams help the examiner's eye lock on, but only if they are clean, labelled, and box-bounded. Shubham Kumar (AIR 1, CSE 2020) similarly noted: "Use of Graphs/pie charts (which I had prepared beforehand) and statistics — economists love stats/data." The phrase "prepared beforehand" is critical — toppers do not invent diagrams in the exam hall; they reproduce ones they have practised 30+ times.

Execution rules

  1. Always use a pencil or single pen — no colours, no shading. UPSC scripts are black-and-white scanned.
  2. Label everything in short, precise terms. Unlabelled = invisible.
  3. Box your diagram with a thin border so the examiner's eye locks on it.
  4. Caption it — "Fig: Stages of Disaster Management Cycle."
  5. Reference it in the prose — "As shown in Fig 1, the cycle has five phases..."
  6. Keep it small — 1/4 of a page max for a 15-marker, 1/6 for a 10-marker. A diagram that eats half a page steals body-text marks.

A 30-diagram "bank" for CSE 2026

Build a personal sketch-bank covering all four GS papers. Indicative coverage:

PaperSuggested diagrams
GS-1India physical map, monsoon trajectory, Indus drainage, Western Ghats biodiversity, demographic dividend curve, urbanisation cone
GS-2Bill-to-Law flow, Indian judicial hierarchy, scheme implementation chain, federalism Venn (Central + State + Concurrent), Inter-State Council structure
GS-3Disaster management cycle, Phillips curve, Laffer curve, FRBM glide path, energy mix pie, water cycle, carbon cycle
GS-4(Avoid; use only "ethical decision-making" mind-map sparingly)

Time each one under 90 seconds. By August 2026, they should appear automatically — like reciting a multiplication table.

The diagram-frequency sweet spot

From topper answer-copy aggregation (Anudeep 2017, Shubham Kumar 2020, Shruti Sharma 2021):

Per GS paper (20 Qs)Diagrams usedWhy
Anudeep CSE 2017 GS-35-6Economy graphs, environment cycles
Shubham Kumar CSE 2020 GS-16-7History temples, geography maps
Shubham Kumar CSE 2020 GS-23-4Federalism Venn, scheme chains
Shubham Kumar CSE 2020 GS-35-6Disaster cycle, economy pies
Shubham Kumar CSE 2020 GS-40-1Ethics avoids visuals
Shruti Sharma CSE 2021 GS-22-3Used sparingly; structure-heavy

The pattern: GS-1 and GS-3 reward visuals more; GS-2 rewards them moderately; GS-4 almost never. Plan your Diagram Bank with this distribution in mind — don't waste prep time on GS-4 visuals.

Mentor tip

Maintain a "Diagram Bank" notebook — 30-40 reusable rough sketches drawn in your own hand, timed at under 90 seconds each. Practise them like multiplication tables. On exam day, they appear effortlessly — and that effortlessness is what scores.

Sources: · · · ·

How do you add credibility with quotes, data, and committee names without sounding like a name-dropper?

TL;DR

Three rules: (1) Always cite the source — "Economic Survey 2024-25," "NFHS-5," "2nd ARC." (2) Use quotes sparingly, and never without explaining their relevance. (3) Memorise 1-2 anchor reports per GS paper (2nd ARC for GS2, Economic Survey for GS3, M.S. Swaminathan for agriculture) — depth beats breadth.

Why credibility markers move marks

When 5,000 candidates write on "police reforms," the ones who cite Prakash Singh v. Union of India (2006) + 2nd ARC's 5th Report + Padmanabhaiah Committee score in the top decile. The content is the same — the attribution is different. Examiners read these markers as "this candidate has gone beyond NCERT."

The four credibility tiers

Tier 1 — Government primary sources (highest weight)

  • Economic Survey (latest: 2024-25, presented Jan 2025) — for any GS3 economy answer
  • Union Budget — for fiscal/scheme questions
  • NITI Aayog reports — SDG India Index, Aspirational Districts data
  • NFHS-5 (2019-21) — health, gender, nutrition
  • RBI Annual Report / Monetary Policy — financial sector
  • PIB press releases — latest scheme announcements

Tier 2 — Constitutional and committee anchors

  • 2nd ARC (2005-09) — governance, RTI, ethics, local government, public order
  • Sarkaria (1988) + Punchhi (2010) Commissions — Centre-State relations
  • Justice Verma Committee (2013) — women safety
  • M.S. Swaminathan Committee (2004-06) — agriculture, MSP
  • B.N. Srikrishna Committee (2017-18) — data protection

Tier 3 — Landmark case law

Kesavananda Bharati (1973), Maneka Gandhi (1978), S.R. Bommai (1994), Vishaka (1997), Puttaswamy (2017), Sabarimala (2018), Navtej Johar (2018). Always include year + one-line ratio.

Tier 4 — Thinkers and quotes (use sparingly)

For GS4/Essay: Gandhi, Ambedkar, Tagore, Vivekananda, Kautilya, Plato, Kant. Rule: maximum 1-2 quotes per essay, never more than 1 per GS answer. Always follow a quote with one line of your own contextualisation.

The credibility-density table (target per answer)

Answer typeMin. credibility markersOptimal mix
10-marker (150 words)1-21 article/scheme + 1 committee OR case
15-marker (250 words)3-41 article + 1 case + 1 committee + 1 data point
GS-4 ethics case study2-31 thinker quote + 1 constitutional value + 1 code-of-conduct provision
Essay (1000-1200 words)8-122 quotes + 3-4 data + 2-3 committee/case + 2-3 thinkers

How to deploy them

  • Open with one — "According to the Economic Survey 2024-25, India's services exports crossed $340 billion..."
  • Anchor a sub-point with one — "The 2nd ARC's 4th Report on Ethics in Governance recommended a Public Service Values law..."
  • Close with a forward-looking one — "Implementing the Punchhi Commission's recommendation on Article 355 can strengthen cooperative federalism."

What Shruti Sharma (AIR 1, CSE 2021) said about value-addition

In her Forum IAS "Basics of Answer Writing" session (widely circulated on YouTube), Shruti Sharma was explicit about restraint: "My answers had no decorative language. No attempt to sound impressive. They read like someone who understands governance, not someone trying to perform understanding." She used data, reports, and examples only where they added substance — never as ornament. For essays, she added quotes; for GS, she relied on committees, articles, and recent schemes over thinker quotes.

This is the inversion most aspirants get wrong: they pepper GS answers with Gandhi/Tagore quotes (which belong in essays/GS4) and forget the committee names (which belong in GS-2/GS-3).

Anudeep Durishetty's anchor-report list

Anudeep's blog repeatedly recommends "reports released by national and international organizations to add validity to your points, such as SC judgments, Economic Survey reports, ARC reports, WTO reports, and UNO reports." His CSE 2017 GS-3 answer copies (publicly archived) show 3-5 such anchors per 15-marker — never random, always tied to the sub-point's argument.

The "specificity test"

Before submitting an answer (in practice), apply this 3-question test:

  1. Did I name at least one article, scheme, or law with year?
  2. Did I name at least one committee, commission, or case?
  3. Did I include at least one data point or % figure?

A 10-marker should pass 2/3. A 15-marker should pass 3/3. If you fail, your answer is generic — the kiss of death in UPSC evaluation.

The 2nd ARC quick-recall table (governance answers)

The 2nd Administrative Reforms Commission (2005-09) produced 15 reports — and 3 of them appear in 80% of GS-2 governance answers.

Report #TopicUse it when
1Right to InformationRTI, transparency, info-rights
4Ethics in GovernanceGS-4, public service values, code of ethics
5Public OrderPolice reforms, internal security
6Local GovernancePanchayats, municipalities, 73rd/74th Amendments
11Promoting e-GovernanceDigital India, JAM trinity, service delivery
12Citizen Centric AdministrationService delivery, citizen charters

Memorise these six. They cover 80% of governance question-types UPSC has asked since 2013.

Mentor tip

Maintain a one-page "Anchor Sheet" per GS paper with 15-20 reports, 10 committees, 10 court cases, and 5 quotes. Revise it weekly. Within 8 weeks, these become reflexes — and reflexes are what survive exam-hall pressure. By CSE 2026 Mains, you should be able to recall the relevant 2nd ARC report number within 5 seconds of seeing a governance question.

Sources: · · · ·

What is a realistic daily and weekly answer-writing routine — without burning out?

TL;DR

2 GS questions daily (one 10-marker + one 15-marker) in the static/current theme you studied that day, written under timed conditions. 1 essay every weekend (3 hours, 1200 words). Review same-day if self-evaluating, weekly if mentor-reviewed. Sustained for 4-6 months pre-Mains, this beats every coaching test series.

Why 'daily' beats 'binge'

Answer writing is a motor skill, like cricket batting. You cannot bat 200 balls on Sunday and skip Mon-Sat — your timing decays. The same physiology applies here: 20 minutes daily for 180 days beats 6 hours weekly for 30 weeks.

The weekday routine (Mon-Fri)

Morning study block

Read one static topic + linked current affairs (45-60 min).

Answer writing block (25-30 min)

  • One 10-marker, 7 minutes — on the topic you just studied
  • One 15-marker, 11 minutes — on the same topic or a current-affairs angle
  • Use actual A4 sheets, write with your exam pen, time yourself with a phone stopwatch face-down.

Review block (10-15 min)

  • Compare against a model answer (Drishti, Insights, Forum IAS)
  • Mark in red: missing dimensions, missing data, missing way-forward
  • Note 2-3 specific corrections in a "Errors Logbook"

The weekend routine

Saturday — Full-length essay (3 hours, 1200 words)

Pick one of the 8 themes UPSC rotates (philosophical, social, economic, polity, S&T, environment, federalism, ethics). Write under exam conditions. Review Sunday morning.

Sunday — Sectional test

Pick one GS paper (rotate weekly). Attempt 5-10 questions in 90 minutes. This builds exam-day stamina, which 20 minutes daily cannot.

The Shubham Kumar volume benchmark

Shubham Kumar (AIR 1, CSE 2020) — IIT Bombay civil engineering, third attempt — practised on this volume: "During my mains preparation, I used to take 1-hour mock tests daily and a full length 3 hour mock test every third day." That translates to:

CadenceVolumeApprox. weekly answers
Daily 1-hour mock4-5 questions28-35
Every 3rd day, full 3-hour mock20 questions~45
Combined weekly70-80 answers

Most aspirants cannot sustain 70-80/week. The realistic baseline is 14 answers/week (2/day × 7 days) — and even this puts you ahead of 90% of competition by Mains.

The progression curve for CSE 2026 (Mains: 21 August 2026)

Working backwards from 21 August 2026, today (May 2026) is roughly 15 weeks before Mains assuming you clear Prelims (24 May 2026). Use this calendar:

PhaseWeeks before MainsFocusAnswers/day
Foundation16+ (i.e., now)Structure — 2-10-2 frame2
Pre-Prelims pause8-10 weeks (Apr-May)Reduce to 3-4 answers/week0.5
Post-Prelims sprint0-12 weeks (Jun-Aug 2026)Daily writing + 2 sectional tests/week4-5
Final 2 weeksAug 7-20, 2026Full mock simulation, 3 papers/weekMock-only

The candidates who clear Mains 2026 are already writing daily as of May 2026. The 90 days between Prelims result and Mains is not enough to build the habit from scratch.

Burnout-proofing

  • Take one full day off weekly — non-negotiable.
  • Cap daily writing at 30 minutes on weekdays. More than that, and quality collapses.
  • Track inputs (minutes written, answers reviewed), not outputs (marks). Marks fluctuate; consistency compounds.
  • Rotate themes weekly — don't write 14 answers on polity in one week. Cycle through GS-1 / GS-2 / GS-3 / GS-4 / Essay.

What sustained practice looks like in numbers

Months of daily practiceTotal answers writtenLikely rank tier
1 month (14 × 4 ≈ 60)60No measurable impact
3 months (~180)180Structural fluency achieved
6 months (~360)360Top decile readiness
9-12 months (~500-700)500+Topper-tier muscle memory

Shubham Kumar's volume (70-80/week × 16 weeks ≈ 1200+ answers) is the upper bound. You do not need that volume — but the relationship between answers written and marks scored is linear, not logarithmic.

What Anudeep Durishetty did between his 4th and 5th attempt

Anudeep (AIR 1, CSE 2017) failed four times before clearing — and the inflection in his 5th attempt was not fresh content but deliberate, repeated answer writing under timed conditions. His blog notes that he wrote on an actual UPSC-style answer booklet (not on plain sheets), with a real exam-pen, against a phone stopwatch — every single day. The simulation, not the content, was the unlock.

For a CSE 2026 candidate today (May 2026), the equivalent move is: buy a stack of UPSC answer booklets (or print 12-page A4 ruled sheets), write on those exclusively for 3 months, and your hand learns the spatial dimensions of a real answer — how 150 words occupies one page, how 250 fills 1.5, where the diagram fits, how long 11 minutes feels. This haptic memory cannot be replicated on a laptop.

Mentor tip

Keep a physical "Answer Writing Streak Calendar" on your wall — mark an X for every day you wrote two answers. Don't break the chain. By the time you sit for Mains on 21 August 2026, you will have written 300+ answers — more than 95% of your competition.

Sources: · · · ·

Self-review, peer review, or mentor review — which actually improves your answers?

TL;DR

All three have a role, but in different phases. Self-review builds structural awareness (months 1-2). Peer review exposes you to diverse approaches (months 2-4). Mentor review gives expert calibration on content depth and presentation (months 4-6, pre-Mains). The myth is that mentor review is needed from day one — it isn't.

The three-stage review pyramid

Stage 1 — Self-review (foundation phase, months 1-2)

Goal: Internalise structure.

  • Write the answer
  • Wait 30 minutes (cognitive distance helps)
  • Compare with the model answer line-by-line
  • Mark in red: missing intro hook, missing sub-headings, missing data, missing way-forward
  • Maintain an Errors Logbook — 2-3 corrections per answer, reviewed weekly

Why it works: No one knows your knowledge gaps better than you. Self-review catches the 60-70% of issues that are structural and discipline-based.

Where it fails: You cannot judge content depth or factual nuance. You will mark your own "according to Economic Survey" as a credible citation even if you got the figure wrong.

Stage 2 — Peer review (months 2-4)

Goal: Exposure to diverse framings.

  • Form a 3-4 person WhatsApp/Telegram group
  • Share scanned answers daily
  • Rotate the evaluator role — each person reviews one peer's answer per day
  • Use a standard rubric: structure /3, content /3, value-addition /2, presentation /2

Why it works: You see how 3 other minds attack the same question. You spot dimensions you missed and adopt phrasings that work.

Where it fails: Peers reinforce each other's mistakes. Without an expert anchor, the group can spiral into mutual back-patting.

Stage 3 — Mentor review (months 4-6, pre-Mains)

Goal: Calibrate to UPSC's actual evaluation standard.

  • Choose a mentor who has either cleared UPSC or trained scorers (rank holders, retired examiners, established faculty)
  • Submit 2-3 answers per week, not more — quality of feedback matters more than volume
  • Demand line-by-line feedback, not just a score
  • Specifically ask: "Is my directive-word fidelity correct? Is my conclusion forward-looking? Is my value-addition exam-grade or coaching-grade?"

Why it works: Mentors catch the subtle 10-15% gap between a "good" answer (rank 1000) and a "top" answer (rank 100).

Where it fails: Over-reliance. If you outsource judgment to a mentor, you stop developing your own examiner-eye — which is fatal in the exam hall when no mentor is reading over your shoulder.

The review rubric — a 10-point scoresheet

Use this on every answer you self-review:

ParameterMaxWhat to check
Directive fidelity2Did I do what the directive asked (discuss/examine/critically)?
Intro quality12-3 lines, no fluff, anchored in fact or quote?
Sub-heading structure12-4 visible sub-points?
Multi-dimensionality23+ dimensions (P-E-S-T-L-E)?
Credibility markers2At least 1 article + 1 committee/case + 1 data point?
Conclusion (forward-looking)1Way-forward / reform / SDG / constitutional ideal?
Word limit compliance1Within ±10% of target?
Total10

Score yourself honestly. Anything under 7/10 — rewrite tomorrow. Maintain a rolling 7-day average; it should climb from ~5/10 (week 1) to ~8/10 (month 3).

The ideal blend

  • Daily: self-review (every answer)
  • Weekly: peer review (2-3 answers, structured rubric)
  • Fortnightly: mentor review (2 answers, deep feedback)

What Shruti Sharma (AIR 1, CSE 2021) leaned on

Shruti Sharma joined Forum IAS's MGP (Mains Guidance Programme) for structured mentor review — her published test copies show weekly mentor feedback that focused on structure first, content second. Her testimonial reads: "I practiced answer writing daily to improve my speed and articulation." The combination of daily self-writing + weekly external review compounded into AIR 1.

What Anudeep Durishetty did instead

Anudeep cleared on his 5th attempt — earlier attempts taught him that over-dependence on mentor review delays writing volume. By his successful attempt (CSE 2017), he wrote 200+ answers with mostly self-review against model answers, supplementing with occasional peer discussion. The lesson: there is no single right model — but no successful candidate skips self-review.

The review-channel cost-benefit table

ChannelTime cost / answerMoney cost / monthQuality of feedbackBest for
Self-review (against model)10-15 min₹060-70% of structural issues caughtMonths 1-3
Peer review (rotating)15-20 min₹070-80% (depends on peer quality)Months 2-5
Coaching test series0 (outsourced)₹6,000-15,000Variable — depends on evaluatorMonths 4-6
1-on-1 mentor (rank holder / faculty)0 (outsourced)₹15,000-30,00090%+ for content + structureMonths 4-6 only

Note: most aspirants over-spend on mentor review in months 1-3 (when self-review is sufficient) and under-invest in months 4-6 (when mentor review's marginal return peaks).

Mentor tip

The single most underrated technique is reading your own answer aloud the next morning. Overnight, your brain detaches from the writing — what felt brilliant at 11 PM reveals its filler, repetition, and vague generalities at 7 AM. Do this for 30 days and you will not need any mentor for the structural layer. Pair it with the 10-point rubric above and your self-review becomes 80% as good as a mentor's — at zero cost.

Sources: · · · ·

What are the top mistakes — introduction-heavy, no conclusion, vague generalities — and how do you fix them?

TL;DR

The five repeat offenders: (1) 5-line introductions that steal body space, (2) missing or generic conclusions, (3) vague phrases like "India is progressing fast," (4) ignoring the directive word, (5) wall-of-text with no sub-headings. Each is fixable in one writing session if you know the diagnostic.

The Top 5 Mistakes — diagnosed and treated

Mistake 1 — Introduction-heavy answers

Symptom: Intros that spill into 5-6 lines, often opening with "Since time immemorial" or a dictionary definition.

Why it hurts: Every word in the intro is a word stolen from the body — which is where 70% of marks live.

Fix: Cap intro at 2 lines for 10-markers, 3 lines for 15-markers. Use one of three hooks: definition + constitutional anchor, recent data point, or short quote with attribution. Practise the "20-word intro" drill — write a one-line intro every morning for 30 days.

Mistake 2 — Missing or generic conclusion

Symptom: The answer trails off, or ends with "Hence, it is important to address this issue."

Why it hurts: Examiners read intros and conclusions most carefully. A weak conclusion lowers the impression marks.

Fix: Every conclusion must be forward-looking. Pivot to: a committee recommendation, an SDG target, a constitutional ideal (Article 38/Article 51A), or a specific reform. Banish "Hence," "Thus," "In conclusion" — start with "A way forward," "Going ahead," or directly with the action.

Mistake 3 — Vague generalities

Symptom: "India is a vast and diverse country." "Many challenges remain." "Several initiatives have been taken."

Why it hurts: Every vague sentence is a missed opportunity to demonstrate specificity — which is what scores.

Fix: Apply the "specificity test" — every sentence must name a year, a scheme, a number, a person, a place, or a case. "Several initiatives have been taken" → "Initiatives like Smart Cities Mission (2015) and AMRUT 2.0 (2021) have addressed urban infrastructure."

Mistake 4 — Ignoring the directive word

Symptom: Question says "critically examine," candidate writes a one-sided descriptive answer.

Why it hurts: Auto-cap at 50% marks regardless of content.

Fix: Underline the directive word in the question before writing. For "critically" directives, force yourself to include at least one "however / yet / on the other hand" pivot sentence. Maintain a one-page directive cheat sheet on your study desk.

Mistake 5 — Wall-of-text

Symptom: A 250-word paragraph with no breaks, no bolding, no sub-headings.

Why it hurts: Examiners with 200 scripts to grade in a day scan, not read. A wall of text gets a scan, not a read — and you lose 40% presentation marks.

Fix: Every body must have 2-4 sub-headings, each underlined or bolded. Use bullets for descriptive sub-points and short paragraphs for analytical ones. Leave a clear blank line between intro/body/conclusion.

The mistake-frequency table (from coaching evaluator surveys)

Mistake% of aspirants making itAvg. marks lost / answer
Introduction over 3 lines65-70%1-2
Generic/missing conclusion55-60%1-2
Vague phrases (no specifics)70-75%2-3
Directive-word miss30-40%3-4 (autocap)
Wall-of-text50-55%2-3 (presentation)
No way-forward in 15-marker60-65%2-3
Over-quoting in GS (not essay)25-30%1-2
Padding to hit word count40-45%1-2

Fix the top 3 and you gain 5-8 marks per answer — across 20 questions in a GS paper, that is 100-160 marks across all four GS papers combined. That gap alone is the difference between rank 800 and rank 80.

Two bonus mistakes

  • No way-forward in 15-markers — costs 2-3 marks per answer.
  • Over-quoting — more than 1-2 quotes per essay/answer signals weak independent thinking.

Anudeep Durishetty's "mistakes to avoid" list

In his widely-watched video lecture "Mistakes you should avoid during your answer writing," Anudeep (AIR 1, CSE 2017) flagged five repeat sins he himself made in his first four attempts:

  1. Writing what you know rather than what is asked — directive miss.
  2. No structure — paragraphs without sub-headings.
  3. Generic introductions — "In today's globalised world..."
  4. No conclusion or way-forward — answers that just stop.
  5. Padding for word limit — diluting strong points with filler.

His fix: "Read the question twice. Underline the directive. Spend the first minute on a skeleton, not on writing." That one-minute rough-sheet skeleton is what separates a 7/15 from a 12/15 on the same content.

The CSE 2024 GS-2 cautionary tale

In the 2024 GS-2 paper, the question "What changes has the Union Government recently introduced in the domain of Centre-State relations? Suggest measures to be adopted to build the trust..." had two equal parts. Coaching post-mortems showed that roughly 50% of test-takers wrote only on the first half (recent changes), missing the second half (suggested measures). That single directive miss cost them 4-5 marks on one question — and similar misses on 4-5 questions across the paper cost ~20 marks total. Rank-100-tier candidates always finish the second clause of compound questions.

Mentor tip

At the end of every answer, do a 30-second self-audit: Did I (a) hit the directive word, (b) cap my intro, (c) use 3+ sub-headings, (d) cite at least one source, (e) end with a forward-looking line? Five yeses = solid answer. Anything less, mark it for rewrite tomorrow. Build this audit into your daily routine now — by 21 August 2026 (CSE 2026 Mains), it should be automatic, not deliberate.

Sources: · · · ·
Ujiyari Ujiyari — Current Affairs