JEE Main Paper 1 uses one marking rule for every question: +4 for a correct answer, −1 for a wrong one, 0 if you leave it blank. That holds for the 60 multiple-choice questions and for the 15 numerical-value questions. This guide works through what that means for guessing, using arithmetic only. It is based on the JEE Main 2026 information bulletin; the 2027 bulletin is awaited, so check the marking scheme again when it is published.
What is the JEE Main marking scheme?
| Response | Section A (MCQ) | Section B (numerical) |
|---|---|---|
| Correct | +4 | +4 |
| Incorrect | −1 | −1 |
| Unanswered, or marked for review with no answer | 0 | 0 |
There are 75 questions worth 300 marks in 3 hours: 20 MCQs and 5 numerical questions in each of Mathematics, Physics and Chemistry. The full layout is on our JEE Main exam pattern page.
Two details from the bulletin are worth knowing. Section B answers are integers, typed with an on-screen keypad. And if a question is dropped or found to be wrong, every candidate in that shift gets +4 for it, whether or not they attempted it.
How do you work out whether a guess is worth it?
Use expected value: the average result of a decision if you could make it many times.
For a guess, multiply each outcome by its chance and add them up:
- Expected value = (chance right × 4) − (chance wrong × 1)
The arithmetic below assumes a multiple-choice question with four options and one correct answer.
A completely blind guess
If you have no idea, each option has a 1 in 4 chance.
- (0.25 × 4) − (0.75 × 1) = 1 − 0.75 = +0.25 marks
So, strictly on the numbers, even a blind guess has a small positive expected value. That surprises many students. It is small, though, and it comes with a lot of risk, as the next sections show.
After eliminating one option
You are sure one option is wrong, and the other three look equally likely.
- (1/3 × 4) − (2/3 × 1) = 1.33 − 0.67 = about +0.67 marks
After eliminating two options
Two options are out and you are torn between the last two.
- (0.5 × 4) − (0.5 × 1) = 2 − 0.5 = +1.5 marks
The break-even point
A guess breaks even when 4 × (chance right) equals 1 × (chance wrong). That happens at a chance of 1 in 5 (20%). Above that, a guess is worth more than a blank on average. Below it, a blank is better.
| Situation | Chance right | Expected value |
|---|---|---|
| Blind guess, 4 options | 25% | +0.25 |
| One option eliminated | about 33% | about +0.67 |
| Two options eliminated | 50% | +1.5 |
| Break-even | 20% | 0 |
Why is “positive on average” not the whole story?
Expected value describes the long run. On any single question you either gain 4 or lose 1. With only a handful of guesses in a paper, luck matters a great deal.
Take 10 completely blind guesses. The expected total is +2.5 marks. But working out the actual probabilities:
- about 1 in 4 times you would end up with fewer marks than if you had left all 10 blank
- a little over 1 in 4 times you would break exactly even
- a little under half the time you would come out ahead
Now take 10 guesses where you had genuinely ruled out one option each time. You would end up behind only about 1 time in 10, and ahead about 7 times in 10. With two options ruled out, you would end up behind about 1 time in 100.
The lesson: blind guessing is a coin flip with a slight edge. Guessing after real elimination is a good bet.
What can go wrong with elimination?
All of the above assumes your elimination is right. If you cross off an option with false confidence, and it was the correct one, your “50-50” is actually 0%. This is the most common way guessing goes wrong.
So only count an option as eliminated when you can say why it is wrong: a unit that does not match, a limiting case it fails, a sign that cannot be right. “It looks unlikely” is not elimination.
Should you guess on numerical (Section B) questions?
Section B also carries −1 for a wrong answer, so the same break-even rule applies: you need better than a 1 in 5 chance of being right.
A blind integer guess is nowhere near that. There are far too many possible integers, so the chance of hitting the right one by luck is tiny and the expected value is close to −1. Do not type a random number into a numerical question.
A numerical guess can make sense when you have done the working and are down to a small set of candidates. For example, if your method is sound but you are unsure about rounding between two adjacent integers, your chance is roughly 1 in 2, which is well above break-even. The bulletin says the answer should be rounded off to the nearest integer, so read the question’s instruction carefully before you enter a value.
Does guessing affect tie-breaks?
It can. If two candidates have the same total NTA score, the 2026 bulletin breaks the tie in this order:
- NTA score in Mathematics
- NTA score in Physics
- NTA score in Chemistry
- Lower ratio of incorrect to correct answers across all subjects
- The same ratio in Mathematics, then Physics, then Chemistry
If it is still tied after that, candidates share the rank. Steps 4 to 5 mean careless or blind wrong answers are not entirely free even when they do not change your score much. This only comes into play in an exact tie, so it should not stop a well-founded guess. It is one more reason to avoid blind ones. Our guide to how the JEE Main percentile is calculated covers tie-breaks and normalisation in more detail.
What is the time cost of a guess?
The arithmetic ignores time, and in a 3-hour paper time is the real currency. Spending four minutes to eliminate one option on a hard question can be a worse trade than spending those minutes on a question you can actually solve.
A practical approach:
- First pass: answer everything you can solve with confidence.
- Second pass: return to questions where you made partial progress. These are where elimination is quickest and most reliable.
- Last few minutes: for MCQs where you have ruled out at least one option for a clear reason, mark your best remaining choice. Leave Section B questions blank unless you have a worked answer.
Our guide to JEE Main time management goes into this round-by-round approach.
How should you practise this?
Guessing well is a skill, and mock tests are the safe place to build it. After each mock, look at every question you guessed on and sort them:
- guessed after real elimination, and right
- guessed after real elimination, and wrong
- guessed blind
If your “eliminated” guesses are coming out right less than about half the time when you were down to two options, your elimination is less reliable than it feels. That is worth knowing before the real exam, not after.
Previous-year papers are a good way to test your elimination habits under real conditions. You can find them, sorted by chapter, on JEE Notes. For the overall picture of the exam, start with our JEE Main guide.