How to Prepare for GATE ST 2027: Syllabus, Strategy and a 14-Week Plan

ST is the Statistics paper. Its syllabus reads like a good B.Sc. and M.Sc. statistics programme compressed into twelve sections: real analysis and matrix theory, probability and distribution theory, inference, stochastic processes, non-parametric methods, multivariate analysis and linear regression.

Expect proofs in disguise. A question may hand you a density and ask whether a statistic is complete, or give a transition matrix and ask for the long-run probability of state 2. Very little of it is data crunching; most marks come from knowing a theorem's exact conditions and being quick and accurate with the integrals that follow from them.

It suits students of statistics, mathematics and related subjects. If data science also interests you, much of the probability, linear algebra and regression overlaps with GATE DA preparation, and DA is one of the papers you may pair with ST.

All of this is from the official GATE 2027 site at IIT Madras. Dates there can change, so check before relying on them.

Organising institute	IIT Madras
Exam dates	6, 7, 13, 14, 20 and 21 February 2027 (ST's date is in the official schedule)
Result	19 March 2027
Mode	Computer-based, English only
Duration	3 hours
Questions	65: 10 General Aptitude, 55 statistics
Marks	100: General Aptitude 15, ST 85
Question types	MCQ, MSQ, NAT
Wrong answers	MCQs only lose marks: 1/3 (1-mark), 2/3 (2-mark)
Calculator	On-screen virtual calculator in recent GATE exams; 2027 rules in the information brochure
Score validity	Three years from the date the result is announced
Two-paper options with ST	DA, CS, MA, XH

Eligibility covers anyone in the third or a higher year of an undergraduate degree, or who has completed a government-approved degree in engineering, technology, architecture, science, commerce, arts or humanities. The official table adds that 3-year B.Sc. or B.A. students may appear in their third year, and M.Sc. or M.A. students in any branch, Mathematics and Statistics included, from their first year. The eligibility page has the full table.

General Aptitude sits on top of these twelve. Download the official ST syllabus PDF and treat it as the checklist; the table only sketches each section.

Section	Main content
1. Calculus	Countable and uncountable sets; sequences, series and convergence tests; power series; continuity and uniform continuity; mean value and Taylor theorems; Riemann and improper integrals; several variables, Lagrange multipliers, double and triple integrals
2. Matrix Theory	Subspaces, basis, rank and nullity; determinants and linear systems; Gram-Schmidt; eigenvalues and Cayley-Hamilton; symmetric, Hermitian, orthogonal and unitary matrices; diagonalisability; positive definite matrices, quadratic forms, SVD
3. Probability	Axioms, conditional probability, Bayes' theorem, independence; CDF, PMF, PDF; moments, MGF and PGF; quantiles; functions of a random variable; Chebyshev, Markov and Jensen inequalities
4. Standard Univariate Distributions	Bernoulli, binomial, geometric, negative binomial, hypergeometric, discrete uniform, Poisson; continuous uniform, exponential, double exponential, gamma, beta of both kinds, Weibull, normal, Cauchy
5. Joint Distributions	Joint, marginal and conditional distributions; conditional expectation; correlation; joint MGF; functions of random vectors; order statistics; multinomial and bivariate normal; central chi-square, t and F
6. Convergence of Random Variables	Modes of convergence and their relations; Slutsky's lemma; Borel-Cantelli lemma; weak and strong laws; CLT for i.i.d. variables
7. Stochastic Processes	Markov chains on finite and countable spaces, state classification, stationary distributions; Poisson process; birth-and-death, pure-birth and pure-death processes; Brownian motion
8. Estimation	Sufficiency, minimal sufficiency, factorisation, completeness; ancillary statistics and Basu's theorem; UMVUE, Rao-Blackwell, Lehmann-Scheffe; Cramer-Rao; consistency; moments and maximum likelihood; pivots and confidence intervals
9. Testing of Hypotheses	Neyman-Pearson lemma; MP and UMP tests; MLR; UMP unbiased tests including exponential families; likelihood ratio and large sample tests
10. Non-parametric Statistics	Empirical distribution function; goodness of fit; chi-square and Kolmogorov-Smirnov; run, sign and Wilcoxon signed rank; Mann-Whitney U and Kruskal-Wallis; Spearman and Kendall correlation
11. Multivariate Analysis	Multivariate normal with marginals and conditionals; MLE of mean vector and dispersion matrix; Hotelling's T-squared; Wishart; multiple and partial correlation
12. Regression Analysis	Simple and multiple linear regression; R-squared and adjusted R-squared; quadratic forms; Fisher-Cochran and Gauss-Markov; tests and intervals for coefficients

Small items such as Basu's theorem, Brownian motion and SVD are easy to skip on a quick read. Go through the PDF once, slowly.

Sixty-five questions, each worth 1 or 2 marks. Ten are General Aptitude for 15 marks. The other 55 carry 85 marks and come entirely from the twelve ST sections, since this paper has no separate Engineering Mathematics component.

Three formats: MCQ with one correct option, MSQ with one or more, and NAT where you type the value. Only MCQs carry a penalty, 1/3 for a wrong 1-mark question and 2/3 for a wrong 2-mark one. MSQs give no partial credit.

That MSQ rule shapes how ST must be learnt. A typical MSQ lists four statements about an estimator or a mode of convergence, and you must pick every true one. Half-remembered theorems fail here. A typical NAT asks for something concrete: the variance of an MLE at a given sample size, a probability from a Poisson process, the size of a test.

On calculators, recent GATE exams provided an on-screen virtual scientific calculator and did not allow physical calculators. The GATE 2027 information brochure will confirm the rules for this year. Practise on a virtual calculator, and get comfortable working with exact expressions rather than looked-up values.

Think of the syllabus as a ladder. Probability and the standard distributions are the bottom rungs: expectations, MGFs and transformations are tools you will use in every later section. Calculus and matrix theory run beside them at about an hour a day, because quadratic forms, the multivariate normal and regression all need the matrix work later.

Joint distributions come next, and order statistics with them. Once you can derive chi-square, t and F from normal samples yourself, the tests and intervals in the inference sections stop feeling like a list of arbitrary recipes. Convergence of random variables is short but subtle and supports consistency and large sample tests.

Then the heart of the paper. Estimation first, in the order sufficiency, completeness, UMVUE. Testing second, with Neyman-Pearson before UMP tests.

Regression and multivariate analysis share their machinery, so study them as a pair. Stochastic processes and non-parametric statistics stand largely on their own and can come late. Late is fine. Skipping them is not.

General Aptitude takes 20 to 30 minutes on three or four days a week, from the start.

For about 4 to 5 focused hours a day with a statistics or mathematics degree behind you. Stretch blocks if you have less time.

Weeks	Statistics	Mathematics alongside	Test yourself on
1	Axioms, conditional probability, Bayes, random variables	Sequences and series	Recent probability questions
2	Expectation, MGF and PGF, inequalities, transformations; discrete distributions	Continuity, mean value theorems	Start a distribution table
3	Continuous distributions	Riemann and improper integrals	MGF and Jacobian problems
4	Joint and conditional distributions, bivariate normal	Several-variable calculus	Marginals and conditional expectation
5	Order statistics; chi-square, t and F	Rank, linear systems	Order statistics from uniform samples
6	Convergence, laws of large numbers, CLT	Eigenvalues, quadratic forms	A timed mixed set
7	Sufficiency, completeness, ancillarity, Basu		Finding sufficient statistics
8	UMVUE, Cramer-Rao, MLE, moments, confidence intervals		UMVUE and MLE until error-free
9	Neyman-Pearson, MLR, UMP and UMPU, LR and large sample tests		Exact size and power
10	Regression analysis	Positive definite matrices, SVD	Coefficient tests, R-squared
11	Multivariate analysis; non-parametric tests		Hotelling's T-squared, rank tests
12	Markov chains, Poisson and birth-death processes, Brownian motion		Stationary distributions
13	Three full mocks, then repair		Error log by section
14	Three mocks, theorem sheets, light revision		Nothing new

One session a week from week 3 goes to revisiting what you covered earlier.

Many ST candidates are in the final year of a B.Sc. or the first year of an M.Sc., or working as analysts. With 2 to 3 weekday hours and fuller weekends, spread the 14 weeks over 24 to 28.

Line GATE up with your semester wherever you can. If inference is being taught this term, do the estimation and testing blocks now and let coursework double as preparation. Weekday sessions can stay on one section; weekends are for mixed past-paper questions. A fixed daily slot of ninety minutes works better than one long weekend day.

The GATE 2027 two-paper list lets ST candidates add DA, CS, MA or XH.

DA (Data Science and Artificial Intelligence) is the most common fit for statisticians. Its probability, statistics, linear algebra and calculus overlap with ST, but it also needs Python programming, algorithms, databases, machine learning and AI. Read the GATE DA preparation guide and the GATE DA syllabus before committing, and try a few GATE DA previous year questions to judge the programming load.

MA (Mathematics) suits students whose real strength is pure mathematics, such as analysis and algebra. Compare its scope in how to prepare for GATE MA.

CS is a large and separate syllabus, sensible only with a computing background. The GATE CSE preparation guide shows what it involves.

XH (Humanities and Social Sciences) may appeal if economics is part of your plans; see the GATE XH preparation guide.

According to the official site, the second paper's centre may differ from the first, within the same city.

Calculus

Convergence tests, uniform continuity, Taylor expansions and Lagrange multipliers. Double integrals turn up inside probability questions, so practise switching the order of integration over triangular regions.

Matrix Theory

Rank, eigenvalues, definiteness and quadratic forms carry over into regression and multivariate work. Know the eigenvalue facts for symmetric, orthogonal and idempotent matrices cold.

Probability

Bayes' theorem questions are short and easy to misread. Practise the Jacobian method for transformations, and know when Jensen's inequality gives a strict bound.

Standard Univariate Distributions

Build one table: support, parameters, mean, variance, MGF, and the links (sum of exponentials is gamma, ratio of gammas gives beta, and so on). Once you recognise a distribution, many questions take two lines.

Joint Distributions

Marginals, conditionals and the bivariate normal. Order statistics from uniform and exponential samples are a favourite. Be able to derive chi-square, t and F from normal samples.

Convergence of Random Variables

Learn which modes imply which, and keep a counter-example for each implication that fails. Slutsky, the laws of large numbers and the CLT feed MSQs directly.

Stochastic Processes

Classify states as transient, recurrent or periodic, solve for stationary distributions, and know Poisson process facts such as exponential inter-arrival times and thinning. Add birth-death chains and the basic properties of Brownian motion.

Estimation

Most demanding section of the paper. Apply the factorisation theorem to every standard family. Use exponential-family completeness, then Rao-Blackwell and Lehmann-Scheffe to reach UMVUEs. Compute MLEs including support-dependent cases like Uniform(0, θ), where calculus does not work. Build confidence intervals from pivots.

Testing of Hypotheses

Most powerful tests from Neyman-Pearson, then UMP one-sided tests via MLR. Students often confuse a test's size with its power at a specific alternative; compute both exactly for binomial and normal examples. Add likelihood ratio and large sample tests.

Non-parametric Statistics

For each test, know its assumptions, what it tests and, for simple ones like the sign test, its null distribution. Kolmogorov-Smirnov rests on the empirical distribution function.

Multivariate Analysis

Conditionals of the multivariate normal, MLEs of the mean vector and dispersion matrix, Hotelling's T-squared and Wishart basics. Strong matrix theory makes this section short.

Regression Analysis

Least squares in matrix form, the hat matrix, R-squared versus adjusted R-squared, Gauss-Markov, Fisher-Cochran, and t and F tests for coefficients.

General Aptitude

Fifteen marks in roughly twenty minutes. Statistics students usually find the quantitative part easy, so spend practice time on verbal and spatial questions.

The official GATE 2026 question paper archive at IIT Guwahati carries recent ST papers with answer keys. Nothing else shows as clearly how ST examiners phrase an MSQ on sufficiency or a NAT on a Markov chain.

Use them section by section while you are still learning the material, and then as whole three-hour papers under exam conditions in the final month. Mark strictly against the official key. Note the cause of each mistake: a theorem condition you forgot, an integration slip, a misread question or arithmetic. Redo those questions a week later.

A few 2027 topics may have little history in past papers. Prepare them from your textbook anyway.

Six to eight full mocks in the last five weeks, sat at your real session time, is a sensible target.

Begin with aptitude, about twenty minutes. Then sweep for one-minute questions: naming a distribution, judging a convergence statement, a rank or an eigenvalue. A second sweep handles longer working you know how to do, such as a UMVUE, a power calculation, a stationary distribution or a regression estimate. Heavy integrals get flagged for the end.

Guess on an MCQ only after eliminating two options. Worked MSQ and NAT answers should always go in, as they carry no penalty. For NATs, carry full precision until the final step and round only as instructed.

One standard book per section, read properly, is enough.

Section	Standard texts
Calculus	Bartle and Sherbert, Introduction to Real Analysis; Tom M. Apostol, Mathematical Analysis
Matrix Theory	Gilbert Strang, Linear Algebra and Its Applications; Hoffman and Kunze, Linear Algebra
Probability and distributions	Sheldon Ross, A First Course in Probability; Rohatgi and Saleh, An Introduction to Probability and Statistics
Estimation and testing	Casella and Berger, Statistical Inference; Hogg, McKean and Craig, Introduction to Mathematical Statistics
Stochastic Processes	Sheldon Ross, Introduction to Probability Models; Karlin and Taylor, A First Course in Stochastic Processes
Non-parametric Statistics	Gibbons and Chakraborti, Nonparametric Statistical Inference
Multivariate Analysis	T. W. Anderson, An Introduction to Multivariate Statistical Analysis; Johnson and Wichern, Applied Multivariate Statistical Analysis
Regression Analysis	Montgomery, Peck and Vining, Introduction to Linear Regression Analysis; Draper and Smith, Applied Regression Analysis

Free lecture courses on probability, inference, linear algebra and real analysis by IIT and IISc faculty are on NPTEL.

Memorising theorems without their hypotheses. MSQs are built from exactly those hypotheses: regularity for Cramer-Rao, completeness for Lehmann-Scheffe, MLR for UMP tests.

Treating calculus and matrix theory as background. Each is a full section of the paper.

Dropping stochastic processes or non-parametric statistics. Each is self-contained and takes about two weeks.

Switching between parameterisations of the gamma and exponential. Pick one convention and keep it.

Guessing MSQs, where one wrong tick zeroes the question.

1. Stop adding topics four weeks out.
2. Sit two or three mocks a week, with equal time spent reviewing.
3. Clear the error log, then clear it again.
4. Keep a one-page distribution table and a one-page list of theorems with their exact conditions.
5. Run quickly over inequalities, convergence implications and the rank tests.
6. Do some aptitude every day.

In the final week, sleep at fixed hours that suit your session and keep revision short.

A valid ST score can be used for M.Tech, MS by research and PhD programmes that accept this paper. IIT M.Tech offers go through COAP; NITs, IIITs and many others use CCMT; IISc and several IITs admit to MS and PhD directly. Not every programme takes ST, so check each one's list of accepted papers. Some public sector organisations recruit using GATE scores and publish their own notifications, which state the papers they accept. A GATE 2027 score is valid for three years from the date the result is announced.

To see where ST sits among the others, browse all 30 GATE papers or the full list of GATE preparation guides.