How to Stop Losing Marks to Negative Marking in JEE Main

How does negative marking work in JEE Main?

JEE Main awards +4 marks for a correct answer and deducts 1 mark for an incorrect one. A question left unattempted scores zero — no reward, no penalty. With 75 questions across Physics, Chemistry and Mathematics, the paper totals 300 marks.

That asymmetry between +4 and −1 is the entire strategic content of the exam. It means attempting is not automatically good and skipping is not automatically safe: both have a price, and which one is cheaper depends on a number you can actually measure about yourself.

How many marks are you losing to negative marking?

Almost nobody knows their own number, which is why this is the most commonly underestimated leak on a scorecard. Work it out from your last mock: count your wrong answers and multiply by one. That is the raw figure. Then do the more useful version — split your attempts into three groups by how sure you felt, and check each group separately.

Read the last row carefully. Thirteen questions were attempted, thirteen questions’ worth of clock was spent, and the net contribution to the score was below zero. Those thirteen questions were worse than not being on the paper at all. That is what a negative-marking leak looks like, and it is invisible on a scorecard that only reports the total.

When should you guess in JEE Main?

The advice students usually hear is "never guess", and it is wrong. The marking scheme sets a precise break-even: an attempt is worth making whenever your chance of being correct is better than 20%.

So a genuinely random pick between four options is very slightly positive, and eliminating even one option makes attempting clearly worthwhile. If the arithmetic favours guessing, where do the losses come from?

From the last row. In practice most wrong attempts are not random at all. JEE options are written so that the common mistake — the sign you were going to drop, the factor of two you were going to miss — appears as one of the four choices. A student who makes that mistake does not land randomly; they land on the distractor built for them. That pushes real-world accuracy on "guessed" questions below the 20% break-even, which is exactly why measured accuracy matters more than the theoretical number.

How to decide which questions to leave

A rule you can apply under exam pressure has to be simple enough to run in about two seconds. These work because they use signals you already have while reading the question.

1. 1.If you cannot name the concept being tested within about fifteen seconds, leave it. Not knowing what a question is about is different from finding it hard.
2. 2.If you have eliminated at least one option on real grounds — not a feeling — attempt it. The arithmetic is clearly in your favour.
3. 3.If you are choosing between two options and cannot say why one is wrong, treat it as a coin flip and check it against your own half-sure accuracy. If that number is above 50%, attempt it.
4. 4.If you have already spent your per-question cap, leave it regardless of how close you feel. The sunk time is gone either way.

Notice that only one of these is about the question. The rest are about you — your accuracy, your clock, your ability to name the topic. That is deliberate: the exam is identical for everyone in the hall, and the only variable you control is your own decision threshold.

How many questions should you attempt in JEE Main?

There is no universal number, and any article that gives you one is guessing on your behalf. The right count is whatever leaves your guessed group small enough that its negative contribution does not eat into what the confident group earned.

Work it backwards from your own data. If your confident group runs at 90% accuracy and your guessed group at 20%, every question moved from "guess" to "leave" costs you nothing and saves the clock for a question you can actually solve. Aspirants who cut their attempt count often find their net score rises, which feels wrong and is arithmetically obvious.

How to practise leaving questions before the exam

The rule is easy to write down and hard to follow, because leaving a question feels like giving up. It has to be rehearsed in mocks or it will not survive the real hall.

• Mark your confidence on every question as you attempt it — one letter in the margin is enough. You need this to compute the three groups afterwards.
• Set a per-question time cap before the paper starts and hold it even when you are close. The questions you never reached cost more than the one you abandoned.
• After each mock, recompute your accuracy by group. It moves as you improve, so the threshold should move with it.
• Track marks lost to negative marking across papers as a number in its own right. It is the fastest-moving indicator you have.

Common negative marking mistakes

• Treating all wrong answers the same. A wrong confident answer is a concept gap; a wrong guess is a decision problem. They need opposite fixes.
• Deciding your attempt rule from someone else’s accuracy. The break-even is universal; your position relative to it is not.
• Changing an answer late without new reasoning. Second-guessing usually moves you toward the distractor, not away from it.
• Leaving questions you could solve because you are afraid of the penalty. Over-caution costs quietly and does not show up as a wrong answer.