MCQ, MSQ and NAT in GATE CSE: How to Attempt Each Type

GATE CSE uses three question types, and each needs its own decision rule. MCQs carry negative marking, so a blind guess is worth nothing on average. MSQs and NATs carry no negative marking, but an MSQ gives no partial credit and a NAT gives you no options to check against. Here is how to attempt each type, with the arithmetic behind every rule.

Key takeaways

• MCQ: one correct option of four; a wrong answer costs 1/3 (1-mark) or 2/3 (2-mark). A blind guess has zero expected value, so guess only after eliminating at least one option.
• MSQ: one or more correct of four; no negative marking and no partial credit. Judge each option independently and select only what you can justify.
• NAT: type a number; no options and no negative marking. Carry full precision and round only the final answer, as instructed.
• In the last 10 minutes, give every NAT and MSQ an answer. A blank scores zero; an answer might score.
• Analyst counts suggest roughly half the questions in recent papers are MCQ and the rest MSQ or NAT. These counts are approximate.

The three types at a glance

NAT questions entered GATE in 2014 and MSQ in 2021, so older previous-year papers under-represent both. Third-party analyses put the 2025 CS paper at roughly 34 MCQ, 15 MSQ and 16 NAT. One count of the 2026 Shift 2 paper reported 34 MCQ, 11 MSQ and 19 NAT, with most NATs worth 2 marks. The official pattern and marking scheme are explained in GATE CSE 2027 exam pattern and marking scheme.

MCQ: when a guess pays and when it does not

The MCQ penalty is one third of the question's marks. That number makes a blind guess exactly break-even. For a 2-mark MCQ, a random pick among four options is right one time in four (plus 2) and wrong three times in four (minus 2/3), which averages to zero.

Elimination changes the arithmetic. The table shows the expected marks from a random pick among the options you have not ruled out, assuming your eliminations are correct.

For example, a 2-mark MCQ with one option eliminated gives one third times 2, minus two thirds times 2/3, which is 2/9, about 0.22 marks.

The rules that follow:

1. Never guess with no option eliminated. You take on risk for zero expected gain.
2. Guess when one or more options are genuinely eliminated. The expected value is positive, and it grows quickly with each elimination.
3. Do not change a confident MCQ answer without a definite reason. Change it only when you find a specific error in your working.
4. Read every option to the end. GATE often puts the discriminating clause at the end of an option.
5. Mark negations. Under time pressure, "which is NOT true" is easily read as "which is true".

MSQ: all or nothing, one option at a time

An MSQ has one or more correct options among four, and all four can be correct. You score only if your selection matches the correct set exactly. There is no negative marking, so a wrong selection costs nothing extra, but there is no partial credit either.

That combination creates one clear rule: decide each option independently. Do not look for "the best answer", and do not assume at least one option must be false. For every option, ask whether you can justify it or find a counterexample. Select only what survives.

Here is a short illustration, built on the marking rules themselves:

Which of the following statements about GATE CSE marking are correct?

(A) A wrong answer to a 2-mark MCQ costs 2/3 of a mark.

(B) A wrong MSQ answer costs 1/3 of a mark.

(C) A NAT question has no negative marking.

(D) Selecting one of two correct MSQ options earns half the marks.

The correct set is A and C. Selecting only A scores zero. Selecting A, C and D also scores zero. Only the exact pair earns the marks, which is why each option needs its own check.

Patterns that frequently trip candidates on MSQs:

• A true theorem next to its false converse. MST and shortest-path property sets are a favourite, and one wrong tick costs the whole question.
• Direction of implications. Function composition and injectivity statements test each direction separately.
• Data interpretation in General Aptitude. Base-year and weighted-average statements look obvious until you compute every quantity. Our General Aptitude guide covers these.
• Statements that "sound right". Selecting an option without a counterexample check is a common way to lose a full MSQ.

NAT: no options, so precision decides

A NAT question has no options. You type a number using the on-screen virtual keypad, and a virtual calculator is provided. There is no negative marking, so always enter your best computed value, but nothing on the screen will tell you that you have made a slip.

Five habits prevent the usual NAT losses:

1. Check the unit and the rounding instruction before you enter the answer. Read them at the start, and again before you type.
2. Carry full precision through the working. Round only the final answer, to the decimals the question asks for. Rounding at intermediate steps shifts the final value.
3. Answer the quantity asked. Edges when nodes are asked, faults when hits are asked, bits when bytes are asked, and waiting time when response time is asked are all classic slips.
4. Write the unit next to every number. Milliseconds versus microseconds versus nanoseconds, KB as 1,000 or 1,024, Mbps as 10^6 or 2^20 bits, and bits versus bytes are the usual unit slips.
5. Put every trace on paper. Gantt charts, cache traces, K-maps, congestion-window tables and precedence graphs are too error-prone to hold in your head.

High-priority NAT computations, based on how often past papers have tested them, include paging levels and page-table size, effective access time with a TLB, cache tag bits, pipeline cycles with stalls, the TCP congestion window after a sequence of RTTs, subnet host counts and fragment offsets, B+ tree fan-out, minimum DFA states, and determinants from eigenvalues. This list is a sample; the book's last-minute revision sheet lists the high-priority numerical and conceptual concepts for every subject, alongside formula tables that name the typical trap for each formula.

A time plan that respects all three types

Question types also decide the order in which you spend time. Recent papers have had 30 one-mark and 35 two-mark questions in total. A plan that works:

The arithmetic: 25 one-mark subject questions at 1.5 minutes and 30 two-mark questions at 3 minutes take about 128 minutes. With 18 minutes of GA and 10 minutes kept at the end, roughly 25 minutes remain for the second pass.

Two rules protect that buffer. First, do not spend it on one stubborn 2-mark question, because two 1-mark questions are worth the same. Second, a question marked for review with an answer is evaluated, so there is no need to clear the mark before time runs out.

Practise the rules in mocks, not on exam day

These rules only help if they are automatic. Build them into every full-length mock:

1. Take the full paper in 180 minutes without breaks, General Aptitude first.
2. Enter NAT answers with the stated precision and treat MSQs as all-or-nothing, exactly as the real paper does.
3. Score with the official marking scheme, including negative marks.
4. Read the solutions only after scoring.
5. Log every wrong or skipped question by subtopic, and note the question type. After three tests, that log is your personal revision list.

Add three numbers to each mock review: net marks from MCQ guesses, marks lost to MSQ over-selection, and marks lost to NAT rounding or units. If any of them is large, you have found marks you can recover without learning a new topic. Our guides to GATE CSE mock tests and common mistakes that cost marks go deeper.

The book includes 10 full-length mock tests and 1,120 questions with worked solutions, enough to turn these three sets of rules into habit well before February.