Two aspirants with the same preparation can score very differently in the same computer-based test. One spends six minutes on a single puzzle, panics, and rushes the last twenty questions. The other attempts the easy questions first, parks the doubtful ones, and walks out with a clean paper. The gap is rarely knowledge. It is exam-day method.
This guide is a complete playbook for CBT-style government exams such as SSC, bank and railway recruitment tests. It covers the night before, the morning routine, the document and centre checklist, the on-screen interface, section order, a three-pass method, time boxes, negative-marking maths with worked numbers, what to do in a bad section, the last ten minutes and the mistakes that cost marks. Every time split in this post is a suggestion for you to test in mocks, not an official rule.
First, know what your exam actually allows
Strategy depends on rules, and rules differ by exam and by year. Before you plan anything, read the latest notification and the instruction page of the official mock or tutorial and note down these points:
- Total questions, total marks and total time.
- Whether sections have their own timers (sectional timing) or one common timer.
- Whether you can move freely between sections or are locked after a section ends.
- Marks per correct answer and the penalty per wrong answer.
- Number of options per question (four in many papers, five in some bank papers).
- Whether an on-screen calculator is provided, and how rough work is done.
As a reference point, the SSC CGL Tier 1 paper has been reported as 100 questions, 200 marks, 60 minutes, four sections of 25 questions each, and a penalty of 0.50 marks per wrong answer. RRB NTPC CBT 1 has been reported as 100 questions in 90 minutes with a penalty of one-third mark per wrong answer. IBPS PO Prelims has been reported as 100 questions in 60 minutes with 20 minutes per section and a 0.25 penalty. Patterns get revised, so treat these as examples and check the latest notification for your exam.
Sectional timing changes everything. If each section has a fixed timer, section order is fixed for you and your time boxes are set by the exam. If there is one common timer, you control both order and timing, and most of this guide applies in full.
The night before: protect your head, not your syllabus
The night before is not the time to learn new topics. New material at this stage mostly creates doubt. Use it for light revision and logistics.
- Pack the bag first. Keep documents, pen and anything else the notice permits in one place so the morning has no decisions.
- Do a short revision pass. Skim your own error notebook, formula sheet and a list of facts you keep forgetting. One hour is enough.
- Do not attempt a full mock. A bad score the night before only damages confidence. A short sectional practice of 15 to 20 minutes is fine if it relaxes you.
- Confirm the route and timing. Check the centre address on your admit card, the travel time and the reporting time written on it. Plan to arrive well before the reporting time.
- Sleep at your normal time. Seven hours at your usual rhythm beats a forced early night that leaves you staring at the ceiling.
- Set two alarms and keep one on a different device.
The morning routine
Keep the morning boring and predictable. Eat what you normally eat, in a normal quantity. A heavy or unfamiliar meal can make a two-hour sitting uncomfortable. Drink water, but not so much that you need a break during the test.
Leave home with a buffer. Traffic, a missed turn or a long queue at the gate should not be able to steal your first question. On the way, do not open books or notes. If you must do something, mentally recall five or six formulas or facts you already know well. The goal is a calm start, not a last-minute cram.
Before entering, keep in mind that candidates are generally checked at the gate and may be photographed or asked for biometric verification. Follow the staff instructions politely. Do not argue about small procedures. Arguments only raise your heart rate before the paper.
Document and centre checklist (general)
The exact list is printed on your admit card and in the notification, and it can change. Use the following as a general memory aid and always follow your own admit card.
- Admit card printout with the photo and details clearly visible. Check the date, shift, reporting time and centre address.
- Valid photo identity proof in the form the notice asks for, usually an original document. The name should match the admit card. If there is a mismatch, read the notice for what to carry and resolve it well before exam day.
- Passport-size photographs if the notice asks for them.
- Pen only if the notice allows it, for example for signing an attendance sheet.
- Anything else the notice names, such as a declaration form or a specific certificate for some categories.
Equally important is what not to carry. Mobile phones, smart watches, earphones, calculators, notes and bags are usually not allowed inside the lab, and many centres have no safe storage. Leave them at home. Do not assume the centre will keep your belongings.
Two more habits help. First, visit or at least locate the centre earlier if it is in an unfamiliar area. Second, do not rely on your phone alone for navigation on the day, because battery and network problems can happen.
How to use the on-screen interface
Most CBT screens look similar. You will see the question in the main panel, a timer at the top, a section name or tabs, buttons at the bottom, and a question palette on the side. Learn this layout through an official mock or a practice test before exam day. On exam day, the tutorial screen is also available; spend a minute reading the instruction page, not skipping it, because it shows the rules for your exam.
The question palette
The palette uses colours or shapes to show the state of each question. Names and designs vary, but the usual states are:
- Not visited: you have not opened it yet.
- Not answered: you opened it and moved on without choosing an option.
- Answered: you chose an option and saved it.
- Marked for review: you flagged it to come back, with no answer chosen.
- Answered and marked for review: you chose an option and also flagged it.
The last state matters. In many interfaces the instructions state whether an answered-and-marked question is counted or not at evaluation. Do not guess this. Read your exam's instruction page, and use that state only the way your exam describes it.
Using mark for review well
Marking is a tool, not a hiding place. Use it with a rule: mark only the questions you plan to revisit and for which you have a reason. A good reason is "I can solve it but it needs more time" or "I am between two options". A bad reason is "I feel uneasy". If you mark forty questions, the flag stops meaning anything and the review phase becomes a second exam you cannot finish.
A practical approach is to use two types of flags in your head, even if the screen has one. If you have a tentative answer, choose it and mark. If you have no idea, leave it unmarked and unanswered, because a question you cannot attack is not worth coming back to unless you have spare time.
Save and next, clear response, and navigation
Know what each button does. The save-and-next button records your answer and moves ahead. A clear-response button removes your selected option. Check whether jumping through the palette saves the option you just clicked, because in some interfaces an option selected but not saved is not recorded. Develop the habit of seeing the palette colour change after each question during mocks, so it becomes automatic.
Calculator and rough work
Most SSC and railway CBTs do not provide a calculator and expect you to work on paper. Some exams, especially certain bank exams, provide an on-screen calculator in some sections. This differs by exam and stage, so check the notification and instructions. If it is offered, practise with it in mocks, because a clumsy calculator can be slower than mental maths.
For rough work, centres generally provide sheets, and you are usually required to return them. Do not bring your own paper. Use the sheet in an organised way: write the question number next to each piece of work. When you come back to a flagged question, your earlier working will be right there.
Choosing your section order
If your exam uses a common timer and lets you move between sections, section order is your first strategic decision. If sections are timed separately and locked, skip this part and use the time boxes set by the exam.
A simple way to decide is to rank your sections on two axes: how many marks you reliably get, and how long those marks take. Then pick the order that builds momentum while protecting time for the slow section.
| Your profile | Suggested order | Reason |
|---|---|---|
| Fast on General Awareness, slower on maths | General Awareness, then Reasoning, then English, then Quant | Quick marks first, so the slow section gets what time remains with a calm mind |
| Strong in maths, average elsewhere | Quant first, then Reasoning, then the rest | You use your fresh mind on your highest-scoring section |
| Nervous at the start | Your most comfortable section first | Early successes settle your nerves for the rest of the paper |
| Weak in one section | Weak section in the middle or last, with a fixed time box | It cannot swallow the time meant for strong sections |
These are suggestions. The only proof is your own mocks. Take at least four or five full-length mocks with different orders, record your scores and time, and keep the order that gives the best results with the least stress. Then do not change the order on exam day.
One caution about leaving a section for last: a section left for the final minutes becomes a rush. If you choose to place a slow section last, make sure you budget its time at the start, not from whatever is left over.
The three-pass method: easy, medium, hard
In a time-limited paper, not every question deserves equal treatment. The three-pass method sorts questions by how quickly you can answer them and handles them in that order. It works inside each section, or across the whole paper if the exam allows free movement.
Pass 1: take the easy marks
Go through the section from the first question. For each question, decide within a few seconds whether it is easy. If you know the method or the answer, solve it and save it. If it looks long, confusing or unfamiliar, leave it unanswered, mark it only if you think you might return, and move on. Do not do "just a little bit" of a hard question. That is how three minutes disappear.
Pass 1 should feel fast. The aim is to collect the questions that cost you under your time box and to see the whole section at least once.
Pass 2: the medium questions
Return to the questions you left because they were longer but doable: a data table that needs three steps, a puzzle you can crack with a diagram, a reading passage you skipped. Because you have already banked easy marks, you can think calmly. Still apply a time cap. If you cross the cap, stop, pick the best-supported option only if you have reasoned elimination (see the negative-marking section), or leave it.
Pass 3: the hard questions and review
Whatever is left is by definition hard for you. With the remaining time, choose the ones where you made some progress. Skip the ones where you have nothing. Use the last few minutes to check flagged answers and confirm that you have not left an obvious easy question blank by mistake.
Why the method works
The method protects you from the most common time trap: spending a long time on a question that is worth the same marks as an easy one. An easy question that takes 20 seconds and a hard one that takes four minutes carry identical marks. Under a common timer, every four-minute question you solve instead of ten easy ones is a loss, unless you had already finished the easy ones.
Time boxes per question type
A time box is the amount of time you allow a question before you decide to move on. The numbers below are starting suggestions for typical question types. Adjust them using your own timed practice.
| Question type | Suggested box | Hard cap before you move on |
|---|---|---|
| General Awareness, direct fact | 10 to 15 seconds | 30 seconds |
| Reasoning: series, coding, odd one out, analogy | 20 to 30 seconds | 60 seconds |
| Reasoning: puzzles, seating, syllogisms | 1 to 2 minutes | 2.5 minutes |
| Quant: simple arithmetic, percentage, averages | 30 to 45 seconds | 90 seconds |
| Quant: algebra, geometry, mensuration, multi-step | 1 to 1.5 minutes | 2.5 minutes |
| Data interpretation set (4 to 5 questions) | 3 to 5 minutes for the set | 6 minutes |
| English: vocabulary, fill in the blank, idioms | 10 to 20 seconds | 40 seconds |
| English: error spotting, sentence improvement | 20 to 30 seconds | 60 seconds |
| English: reading passage (5 questions) | 4 to 6 minutes | 7 minutes |
Use the hard cap as a rule, not a mood. When the cap is reached, you decide to answer by elimination, flag, or skip. You do not "just try one more minute".
Negative marking: expected-value reasoning with real numbers
Negative marking makes blind guessing risky, but it does not make all guessing wrong. The right question is not "will I get this wrong?" but "if I answered many questions like this, would I gain or lose marks on average?" That average is called expected value. It is simple to compute.
The formula for one question is: expected value equals (probability of being right times marks gained) minus (probability of being wrong times marks lost).
Break-even probability
The break-even probability is the chance of being right at which guessing neither gains nor loses on average. If the marks are +M for correct and -P for wrong, then break-even p satisfies M times p equals P times (1 minus p). So p = P divided by (M + P).
Example 1: SSC-style, +2 and -0.5, four options
Break-even p = 0.5 divided by (2 + 0.5) = 0.5 divided by 2.5 = 0.2, that is 20 percent. A blind guess among four options has p = 0.25, which is above 20 percent. Expected value = 0.25 times 2 minus 0.75 times 0.5 = 0.5 minus 0.375 = +0.125 marks per blind guess. It is a small positive.
Example 2: SSC-style, with one option eliminated
You remove one clearly wrong option, leaving three. Now p = 1/3. Expected value = (1/3) times 2 minus (2/3) times 0.5 = 0.667 minus 0.333 = +0.333 marks per question.
Example 3: SSC-style, with two options eliminated
You are down to two options, so p = 0.5. Expected value = 0.5 times 2 minus 0.5 times 0.5 = 1 minus 0.25 = +0.75 marks per question.
Example 4: Railway-style, +1 and -1/3, four options
Break-even p = (1/3) divided by (1 + 1/3) = (1/3) divided by (4/3) = 0.25. A blind guess among four has p = 0.25, so expected value = 0.25 minus 0.75 times (1/3) = 0.25 minus 0.25 = 0. A blind guess neither gains nor loses on average, so there is no reason to take the risk. With one option eliminated, p = 1/3, and expected value = 0.333 minus 0.667 times 0.333 = 0.333 minus 0.222 = +0.111. With two eliminated, p = 0.5, and expected value = 0.5 minus 0.5 times 0.333 = 0.5 minus 0.167 = +0.333.
Example 5: Bank-style, +1 and -0.25, five options
Break-even p = 0.25 divided by 1.25 = 0.2. A blind guess among five has p = 0.2, so expected value = 0.2 minus 0.8 times 0.25 = 0.2 minus 0.2 = 0. Again, neutral. With one option eliminated (four left), p = 0.25, so expected value = 0.25 minus 0.75 times 0.25 = 0.25 minus 0.1875 = +0.0625. With two eliminated (three left), p = 1/3, so expected value = 0.333 minus 0.667 times 0.25 = 0.333 minus 0.167 = +0.167.
Always check how many options your paper has and what the penalty is, because both change the numbers. If the penalty in your notification differs from the examples, redo the arithmetic with the formula above.
What the numbers mean in practice
- When you can honestly eliminate at least one option, guessing is usually a positive bet in all three patterns above.
- With no elimination in the railway and bank examples, guessing has zero expected value. You save time by skipping and spending it on a question you can solve.
- In the SSC-style example, a blind guess has a tiny positive expected value. Even so, it is not worth taking time over, and the risk of a bad streak is real.
Example 6: the risk is real
Take 20 questions in the SSC-style paper, each with one option eliminated (p = 1/3). The expected gain is 20 times 0.333 = about +6.67 marks. But individual results vary. For one question the outcomes are +2 with probability 1/3 and -0.5 with probability 2/3. The variance is 1.5 minus 0.333 squared, which is about 1.39, so the standard deviation per question is about 1.18. For 20 questions it is about 1.18 times the square root of 20, which is about 5.3. So your actual gain could plausibly fall anywhere between about +1 and +12 marks. Guessing helps in the long run, but a single paper is a single draw. This is why a guess should be reasoned, never random.
Example 7: attempts versus accuracy in an SSC-style paper
Take three candidates in a 100-question, +2 and -0.5 paper.
- Candidate A attempts 80, with 70 right and 10 wrong: 70 times 2 = 140, minus 10 times 0.5 = 5, so 135 marks.
- Candidate B attempts 65, with 60 right and 5 wrong: 120 minus 2.5 = 117.5 marks.
- Candidate C attempts 90, with 72 right and 18 wrong: 144 minus 9 = 135 marks.
A and C finish equal even though C attempted ten more questions. B is cautious and loses marks by leaving easy ones. The lesson is that attempts matter, but only when accuracy holds up. Each additional attempt at 80 percent accuracy is clearly worth it. An additional attempt at 50 percent accuracy has an expected value of 0.5 times 2 minus 0.5 times 0.5 = +0.75, which is also positive. The danger is questions below 20 percent confidence, which carry negative value in this pattern.
Honest confidence, not hopeful confidence
Expected-value maths only works if your probability estimate is honest. Test yourself: after a mock, group your guessed questions by how confident you felt and check how many were right. If your "70 percent sure" questions are right only half the time, then adjust your estimates down. Over a few mocks, you will learn your real hit rate and your guess policy will become data-based.
A suggested time table for a 60-minute, 100-question paper
Here is a sample plan for a paper with four sections of 25 questions, a common 60-minute timer and free movement between sections, as in an SSC-style Tier 1 format. If your paper has fixed sectional timers, use the exam's timers instead. The order and splits below are suggestions to test in mocks.
| Section (25 questions each) | Pass 1 | Pass 2 | Total |
|---|---|---|---|
| General Awareness | 4 min | 1 min | 5 min |
| Reasoning | 9 min | 4 min | 13 min |
| English | 9 min | 4 min | 13 min |
| Quant | 12 min | 7 min | 19 min |
| Pass 3, flagged questions and final sweep | Remaining | 10 min | |
Check the arithmetic: pass 1 is 4 + 9 + 9 + 12 = 34 minutes, pass 2 is 1 + 4 + 4 + 7 = 16 minutes, and pass 3 with review is 10 minutes, which gives 34 + 16 + 10 = 60 minutes. The average is 36 seconds per question, but pass 1 deliberately does not attempt everything. In pass 1 you aim to attempt roughly 60 to 65 questions, leaving about 35 for pass 2 and pass 3.
Use clock checkpoints. At 20 minutes, you should be near the end of pass 1 for the second section. At 34 minutes, pass 1 should be finished. At 50 minutes, pass 2 should be finished. These checkpoints tell you early if you are behind, when it is still easy to correct.
A suggested time table for a 60-minute, 75-question paper
Here is a sample for a paper with 75 questions in 60 minutes, split into three sections of 25 and one common timer. This is an illustrative format for planning and for building your own mocks, not a claim about any specific exam, so confirm your exam's own pattern. The average works out to 48 seconds per question, giving you more room per question than the 100-question paper.
| Section (25 questions each) | Pass 1 | Pass 2 | Total |
|---|---|---|---|
| Language or General Awareness | 8 min | 3 min | 11 min |
| Reasoning | 11 min | 5 min | 16 min |
| Quant | 13 min | 6 min | 19 min |
| Pass 3, flagged questions | Remaining | 10 min | |
| Final sweep and readiness to submit | Remaining | 4 min | |
Check: pass 1 is 8 + 11 + 13 = 32 minutes, pass 2 is 3 + 5 + 6 = 14 minutes, then 10 + 4 = 14 minutes for pass 3 and the final sweep. Total = 32 + 14 + 14 = 60 minutes.
If your paper has sectional timers, for example 20 minutes each, use the same logic inside every section: roughly 60 percent of the time on pass 1, 25 to 30 percent on pass 2, and the final minutes for flagged questions. You cannot return to a section after its timer ends, so finish each section cleanly before the timer does.
What to do when you hit a bad section
Almost every aspirant meets a section that feels unfair: a reasoning set of lengthy puzzles, a quant section full of hard geometry, a reading passage on an unfamiliar topic. The danger is not the difficulty. It is the emotional reaction that follows.
- Pause for ten seconds. Take one slow breath. This is cheaper than losing two minutes to panic.
- Remember that difficulty is relative. If a section feels hard to you, it usually feels hard to most candidates. The cut-off adapts to how hard the paper is; check the latest notification for how your exam decides results.
- Cut your target, not your effort. Decide to get the 12 or 15 easiest marks out of 25 rather than chasing 22. Then run pass 1 hard and fast.
- Use the time cap. A bad section turns every question into a time sink. Hold the caps harder than usual.
- Do not take revenge. Spending your next section's time to compensate for the last section usually makes both worse.
- Look for question types you know. Even in a bad paper there are usually a few direct questions: a plain percentage, a simple series, an antonym. Find them first.
- Move on after the section ends. When it is over, it is over. Do not reopen it unless you have spare time at the end.
Example: in the 100-question plan above, suppose quant eats 22 minutes instead of 19 because of a difficult set. You are three minutes behind. Do not rush the remaining sections. Take that deficit from pass 3, and shorten the final review from 10 minutes to 7. You lose only the lowest-value minutes.
The last 10 minutes routine
The last ten minutes should be planned, not improvised. Here is a suggested routine for the 100-question, 60-minute paper. Adapt it to your paper.
- Minute 50 to 52: open the palette summary or scan the palette. Count not-visited questions. A question you never opened is the most wasteful loss, because you may know it.
- Minute 52 to 55: attempt the quickest not-visited and not-answered questions. Use elimination guesses only where the expected-value maths from the earlier section supports it.
- Minute 55 to 58: revisit flagged questions where you already hold a tentative answer and check the one thing most likely to be wrong: a sign, a unit, or a question word such as "not" or "except".
- Minute 58 to 59: confirm that every answer you intend to count is saved and not just selected. Glance over the palette once for answered questions in the wrong state.
- Minute 59 to 60: stop. Do not change an answer in the final seconds unless you have found a clear mistake. Last-moment changes made in a panic tend to lose marks.
Whether to submit early is a separate question. If you finish early, use the spare time to review; there is no prize for leaving first. Also read the on-screen submit instructions carefully, because the submit action may ask for confirmation and cannot be undone.
Common mistakes that cost marks
- Starting a puzzle first. A long puzzle at question 1 can eat 6 minutes of the paper. Open it, judge it, and park it if needed.
- Ignoring the instruction page. You miss rules about calculators, sections and marking states.
- Flagging everything. If everything is flagged, nothing is. Flag with a reason.
- Selecting without saving. If your interface needs a save action, selected is not the same as recorded.
- Guessing without a method. Blind guesses where the penalty cancels the gain are wasted risk.
- Refusing to guess at all. The opposite error. With two options left, the maths in most patterns favours attempting.
- Revenge after a bad section. It makes the next section worse.
- Changing answers late without evidence. Only change an answer if you found a clear error.
- Watching the clock constantly. Use checkpoints, not a stare. Look at the timer at fixed points.
- Skipping mocks in the exam interface. If the interface surprises you on exam day, you lose time and nerve.
- New experiments on exam day. A new section order or a new shortcut method should be tested in mocks first.
Rehearse the plan in mocks
A strategy you have never used is not a strategy. In the last few weeks, sit full-length mocks at the real exam time, follow your time table with checkpoints written on rough paper, and record order, time per section, attempts, wrong answers and flagged questions. Change one variable per mock so you know what worked.
Frequently asked questions
Should I always start with my strongest section? Not always, but it is a good default if you tend to get nervous at the start. If your exam has locked sectional timers, the order is fixed by the exam. Test two or three orders in mocks and keep the one that suits you.
Is it better to attempt all questions or leave some blank? It depends on your accuracy and the penalty. Use expected value: where you can eliminate options and the maths is positive, attempt. Where you have no clue and the guess has zero or negative value, skip.
Can I use the calculator in SSC or railway exams? Generally, these exams expect rough work on paper, and some bank exams provide an on-screen calculator for some sections. Rules can change, so confirm with the latest notification and the instruction page.
What does "answered and marked for review" mean for evaluation? The behaviour depends on the interface and exam. Many instruction pages state whether such answers are counted. Read your exam's instructions and do not assume.
How many questions should I flag? Flag only those you plan to revisit and can finish in the time you expect to have. As a suggestion, a flag list of more than 10 to 12 questions in a 100-question paper is usually too long to review properly.
How early should I reach the centre? Reach before the reporting time on your admit card, with a buffer for travel and entry checks. Late entry may not be allowed at all.
What if I run out of time in the last section? This usually means pass 1 was too slow earlier. In the moment, shift to the easiest questions of that section, use elimination on the rest, and do not panic-guess blindly. After the exam, tighten your time boxes in mocks.
Should I revise on the morning of the exam? A light skim of formulas or facts you know is fine. Avoid learning anything new, and stop revising before you enter the centre.
Is it wise to change an answer during review? Only if you found a specific error, such as a wrong calculation or a misread word. Changing an answer because of a feeling usually does not help.
What is the best use of a mock test? Treat it as a rehearsal of strategy, not just a score check. Review time spent, guesses and flagged questions, then change one variable at a time.
Final thoughts
Exam-day strategy is a skill, and skills improve with practice. Build your plan from the rules of your exam, test it on full-length mocks and sectional tests, and refine it with data from your own attempts. Practise in the exam-style interface on Pareeksha with full-length mocks and sectional tests, so that on the day the screen, the palette and the clock feel familiar and your attention stays on the questions.
Comments (0)
No comments yet. Be the first to share your thoughts.