UPSC Mains as an Optimisation Problem: A Strategy Guide

If you are wondering where to start for UPSC Mains, the most useful mindset shift is to treat UPSC Mains as an optimisation problem. You have limited time, limited resources and a fixed endgame, so your only job is to give yourself the best possible shot within those constraints. In this session, Neil Sir breaks Mains into two layers of optimisation, explains the distinct "flavour" of each GS paper, and shows — using real previous year questions — how repeated themes compress a huge syllabus into a handful of pages.

Key takeaways

• Mains is an optimisation problem with two layers: what you write in three hours, and what you read to prepare.
• The real constraint is not your three hours of writing — it is that your answer gets evaluated in under 15 minutes, so it must look and sound good fast.
• Only two things command truth in this exam: previous year questions and the syllabus. Everything else is opinion.
• Minimum sources, maximum revision: read just enough to have content for every syllabus head and to answer all PYQs from 2013 onwards.
• Each GS paper needs a different approach — GS1 (factual), GS2 (analytical), GS3 (data), GS4 (conceptual).
• Write at least 12 full-length tests; building answer-writing muscle memory matters more than the quality of the mock.
• Master a small set of frameworks and you will never run out of points.

The two-level optimisation problem in UPSC Mains

Neil Sir frames the entire strategy around two optimisation questions you must keep answering.

The first is about the exam hall. Your constraint is not "what gets the highest score in three hours" — it is "what, written in three hours, scores highest when evaluated in under 15 minutes." Drilled to the micro level: an answer you write in 11 minutes may be read in under a minute. So obsessing over a perfect introduction is wasted effort — the examiner will not weigh every detail, as long as what you have written sounds good. Understand this human vulnerability (availability) and design around it.

The second is about preparation. Because time is scarce, you must minimise sources and maximise revision. A "minimal source" is one that gives you enough content across every syllabus head and is sufficient to answer every previous year question on a topic from 2013 onwards. That is the standard — not how thick a compilation is.

The only two truths: previous year questions and syllabus

Don't take anyone's word for it — including, Neil Sir says, his own. In civil services preparation, only two documents command authority: the previous year questions and the syllabus. Everything else is opinion. So your reading should be calibrated against PYQs and the syllabus, not against the fear of missing out that a market flooded with compilations creates.

Each GS paper has its own flavour

The four GS papers share an exam but not a method. Each has a distinct flavour you must respect.

GS1 — factual and memory-intensive

GS1 rewards recall. History themes repeat relentlessly (think Mahatma Gandhi's role in the Non-Cooperation and Civil Disobedience movements), so don't read a thousand-page book cover to cover — prepare the repeating themes. For geography, stick to NCERT; a jet-streams question is either known or not, and anything outside NCERT bounds is a bouncer for everyone. Because memory decays, prepare GS1 near the end, in the last 15 days or so.

GS2 — analytical and about connections

GS2 is about interlinkages. The polity portion (roughly 40 marks) is factual — content on bodies like the CAG can almost be copy-pasted from a standard book. The rest is analytical: governance, civil services, SHGs, NGOs. Around 50 marks of IR repeat every year. Bilateral, multilateral and regional-organisation questions share fixed parameters — boundary issues, technology, climate justice, geopolitics, sea lanes of communication, adversarial versus collaborative ties — so a two-pronged approach works: ready-made framework points, plus a few specific substantiations (key India-Africa trade data, defence partnerships) from a compilation.

GS3 — objective and data-driven

GS3 is where data, committees and objective points pay off. Load answers with data across the economy, agriculture, S&T, environment, disaster management and security. As in IR, a large chunk of internal security repeats, so keep ready-made content for those and reserve your thinking time for the genuinely new questions.

GS4 — conceptual clarity over time

GS4 cannot be crammed in a block; conceptual clarity builds gradually. Neil Sir's three-step routine: spend about 45 minutes daily on introspection built around a single base note, then do answer-writing practice. Start this from day one.

A neat mental model ties it together: think like a humanities graduate in GS2 and GS4, think objectively like a science graduate in GS3, and ratify what you know in GS1. Do this and reaching around 420 marks is well within reach; how far beyond comes down to the art you develop.

Why full-length tests decide your Mains score

Reading more feels safe; it isn't. Mains is about writing, and writing has a strong recency effect — the more recently you have practised, the better your hand moves. Neil Sir recommends at least 12 full-length tests (about three per GS), starting around 15 August so the final month is pure practice. Expect inertia in the first four tests; better to burn through it at home than in the exam hall. (He shares candidly that, despite knowing the concepts, a poor psychological state and lack of writing practice left him blank in GS1 — a costly lesson in muscle memory.)

Because evaluation quality varies, you don't need a paid series to benefit — just sit and write, following the process. A study buddy helps: write the same paper, swap copies, and you start seeing answers through an examiner's eyes — spotting which structured, fully-addressed answer stands out and why.

Frameworks: how to never run out of points

Content gathering is step two; frameworks are step one. Four frameworks — stakeholder, GS-based, life cycle, and short-term/long-term — let you generate 15 points for a 10-marker and 25 for a 15-marker. The three stages are: know the frameworks, apply them, and make them work on the sheet (which is what FLTs train).

The art-and-culture illustration makes it concrete. PYQs show Bhakti literature recurring, Gandhara appearing verbatim across years, and Chola architecture surfacing again and again — so you prepare the repeating dynasties, not the entire set, and brainstorm one classification (performing/visual art, dance, painting) that retrofits to any dynasty. One page of frameworks plus two of substantiation covers art and culture in three pages. The same logic runs through modern history: even if the viceroy's name changes, a Lord Curzon-type question stays ~90% the same — only the substantiation and contextualisation shift.

The essay follows the same spirit. It is topic-neutral and likely read by an English professor, so it rewards connecting ideas to the theme, not subject depth. Build around 20 fresh examples (mined from newspapers or topper copies) you can retrofit anywhere, skip the synopsis-after-introduction habit, and accept that it comes down to practice.

Who should watch this

This is for serious aspirants — especially anyone appearing in the next Mains cycle — overwhelmed by what to read and how to start. If you keep accumulating sources but freeze on answer writing, this optimisation lens will reset your priorities.

Closing thoughts

The throughline is simple: define your endgame, respect your constraints, and trust the process. Internalise the two optimisation questions, write your tests, and let frameworks do the heavy lifting. To put this into action, start daily answer writing practice and study the method for writing Mains answers, then build muscle memory with the Mains test series. For more mentor-led guides, explore the blog.