With the Tier-1 window running from 30 September to 30 October 2026, you are past the stage where new chapters help. What is still fully in your control is the attempt decision: which questions you take on, which you let go, and how fast you decide. Most candidates who miss the cut-off do not miss it because they knew less. They miss it because they spent four minutes on a question worth two marks and then guessed badly on six questions they could have solved. This page works through the actual arithmetic of +2 and -0.5, gives you a defensible attempt target for each section, and tells you what a skip should cost you in seconds. Nothing here needs new syllabus. All of it can be practised in the fortnight you have left.
Tier-1 is a Computer Based Examination: 100 questions, 200 marks, 60 minutes in total. For CGL 2026 the Commission has introduced sectional timing of 15 minutes per section, so each section closes when its own timer ends and unused time does not carry forward. Every correct answer carries +2 and every wrong answer -0.50. Minimum qualifying marks are 30% for UR, 25% for OBC and EWS and 20% for all other categories, but the shortlisting cut-off is far above that. Confirm the exam-day instructions on your admit card and on ssc.gov.in before you sit.
Attempt strategy is a reflex, not a theory. Our free full mocks are 100 questions, four sections of 25, with 15-minute sectional locks and +2/-0.5, so the skip decision is forced on you the same way it will be in the hall. Take four or five of these in the fortnight and the 36-second instinct starts to form on its own.
After a mock, the AI separates your losses into conceptual, calculation, careless and time-pressure errors. That distinction decides your strategy: if most of your losses are time-pressure, you are attempting too many; if they are careless, you are attempting too fast. Free accounts get one AI analysis a week, which is enough for two honest reads before the window opens.
The bank holds roughly 31,000 questions, of which about 29,000 are genuine previous-year SSC questions carrying the Commission's own answer keys. Practising skips on real SSC phrasing matters, because the decision you are training is whether a question is a 30-second question or a three-minute one, and only real papers teach that.
After a mock the system estimates your All-India rank against real 2020-24 cut-off data, so 'should I attempt 80 or 90' stops being a guess. If 82 attempts at your accuracy already clears past cut-offs comfortably, pushing to 95 is adding risk for marks you do not need.
There is no single correct number, because the score depends on attempts multiplied by accuracy. The arithmetic is simple: each attempt is worth 2.5a - 0.5 marks, where a is your accuracy. At 90% accuracy each attempt returns 1.75 marks, at 85% it returns 1.625, and at 80% it returns 1.5. So a candidate at 85% accuracy needs roughly 90 attempts to reach 145, while a candidate at 75% accuracy caps out at 137.5 even if they attempt all 100. Fix your accuracy first; the attempt count follows from it.
No, and this is where most candidates lose marks by being too cautious. With four options, a blind guess has a 25% chance of +2 and a 75% chance of -0.5, which works out to an expected +0.125 marks. The break-even accuracy is exactly 20%, so any question where you can rule out even one option is clearly worth marking. What you should not do is spend time guessing; mark it in two seconds and move on.
For CGL 2026 that choice has largely been taken away from you. The Commission has introduced 15 minutes per section, and a section closes when its timer ends, so you move through the sections in the sequence the screen presents. Plan for that sequence rather than for a favourite-section-first strategy, and confirm the exam-day instructions on your admit card and on ssc.gov.in.
It can be, if your accuracy is high. Eighty attempts at 90% accuracy gives 140 marks, which would have been comfortable against several past years' List-3 cut-offs. The same 80 attempts at 70% accuracy gives 100 marks, which would not. This is why the figure people quote as a 'safe attempt count' is meaningless without the accuracy attached to it.
Careless errors are the most fixable kind in a fortnight, because they are almost never a knowledge problem. They usually come from reading the question stem too fast, or from answering the opposite of what was asked. The practical fix is to re-read only the last line of the question before you click, which costs about two seconds and removes a large share of these losses. Track whether the count falls across your next three mocks; if it does not, you are rushing the wrong section.
SSCIntel — free AI-powered SSC mock tests