Marking, cut-offs and accuracy in LIC AAO

Most candidates arrive at LIC AAO from bank exams, where a wrong answer costs a quarter of the question's marks. The 2025 LIC AAO Generalist papers worked differently: a wrong answer cost nothing. That one rule changes how you should spend the last minute of every section. It does not change what gets you selected. The cut-offs in 2025 sat close to 80 per cent, and every section had its own minimum. This guide sets out the rules, the arithmetic of guessing and the accuracy habits that matter more than either.

The marking rules, stage by stage

For SC, ST and PwBD candidates the minimums are 16, 16 and 9 in the prelims, 40, 27, 40, 27 and 9 in the mains, and 27 in the interview.

Each objective question has five options with one correct answer. Your score depends only on the number you answer correctly. When a test runs in several sessions, LIC equates scores across sessions using the equi-percentile method, and scores are reported to two decimal places. So a slightly harder shift does not by itself put you at a disadvantage.

Three layers of cut-off

It helps to picture selection as three separate filters.

1. Sectional minimums. You must clear every one. A single section below its minimum ends your attempt, however high your total.
2. The stage cut-off. About 20 times the vacancies in each category go from prelims to mains, ranked on Reasoning plus Quant (70 marks). About three times the vacancies go from mains to interview, ranked on the 300 mains marks.
3. The final merit list. Mains (300) plus interview (60), out of 360. Prelims marks are not carried forward, and the descriptive score is never added.

The 2025 category cut-offs show how far above the minimums the real bar sits:

The UR prelims minimum is 18 of 35 in each scored section, about 51 per cent. The UR cut-off of 58 of 70 is about 83 per cent. Clearing minimums gets you into the ranking; it does not get you shortlisted.

What "no negative marking" means in numbers

With five options and no penalty, a blind guess is right one time in five. Its expected value is one-fifth of the question's marks.

Under these rules, a blank never beats a guess. In the final minute of a section, every question still unanswered should get your best choice.

Two cautions keep this in proportion. First, guessing adds fractions of a mark; it cannot replace solving. Second, the information handout advises candidates against marking answers by random guessing, and LIC analyses response patterns across candidates. The sensible approach is to solve everything you can, eliminate options wherever possible, and mark the rest only in the closing minute.

If your cycle brings back a penalty

Suppose the rule becomes minus a quarter of the question's marks, as in the 2025 specialist prelims. A blind guess on a 1-mark question is then worth 0.2 − (0.8 × 0.25) = 0. After eliminating two options it is worth (1/3) − (2/3 × 0.25) = 0.33 − 0.17, about +0.17. So under a penalty, guess only after elimination, and leave pure guesses blank.

Why accuracy still decides selection

The prelims gives you 20 minutes for 35 questions, about 34 seconds each. The mains gives 80 seconds a question in Reasoning and DI, and only 40 seconds in the awareness sections. A wrong answer costs no marks, but it always costs time.

Worked example. Take two hypothetical candidates in prelims Quant.

Candidate P attempts 34 questions and gets 27 right, then guesses the one left: 27 + 0.2 = 27.2 expected.
Candidate Q attempts 30 questions carefully and gets 28 right, then guesses the five left: 28 + (5 × 0.2) = 29 expected.

Q attempted fewer questions and still finished almost two marks ahead. The four extra wrong answers P made also used about two minutes that could have gone into checking.

Worked example: the sectional trap. A UR candidate in the mains gets Reasoning 24 correct (72), GK 14 correct (28), DI 25 correct (75) and Insurance 26 correct (52). The total is 72 + 28 + 75 + 52 = 227. It does not matter, because GK is 28, below the minimum of 30. One more correct GK answer would have kept the candidate in the ranking.

Setting your attempt target from mocks

Nobody can tell you the right number of attempts. Your mocks can. After each full mock, log four numbers for every section.

Over five or six mocks, a pattern appears. Many candidates find that pushing attempts past a certain number makes accuracy fall sharply. That number is your working limit. Plan to solve up to it with care, then use the last minute for elimination and best guesses.

Practice questions

1. In the 2025 Generalist prelims, you have 20 seconds left in Quant and three questions unread. What should you do? Mark an answer on all three. With no penalty, the expected gain is 3 × 0.2 = 0.6 marks.
2. In mains DI, you have eliminated two options on a question. What is a guess worth? 1 mark on average. One-third of 3 marks.
3. Under a minus-quarter rule, you can eliminate only one of five options on a 1-mark question. Is a guess worth it? Only just. (1/4) − (3/4 × 0.25) = 0.25 − 0.1875 = 0.0625, a very small positive value.
4. A UR candidate scores Reasoning 32, Quant 28 and English 9 in the prelims. The total of 60 is above the 2025 cut-off of 58. Are they shortlisted? No. English is below its minimum of 10.
5. A candidate has 238 in the mains and 29 in the interview. Do they make the final list? No. The interview minimum for UR is 30 of 60.
6. How many more correct GK answers turn a mains total of 226 into 232? Three. Each GK question is worth 2 marks. In DI, two would do it.

What to do next

• Read the marking rule for your post in the latest notification before your next mock.
• Log attempts, accuracy and time left for each section in your next three mocks.
• Practise the last-minute routine: eliminate, then mark every blank.
• See how the stages fit together in the LIC AAO exam pattern explained, and how to run each 20-minute section in the LIC AAO prelims strategy.