NEET PG awards +4 for a correct answer, deducts 1 for a wrong one, and gives nothing for a question left blank. Most candidates translate that into "be careful, do not guess". The arithmetic says otherwise, and it is worth doing once properly.
The expected value of a pure guess
Four options, no information at all. One time in four you are right and gain 4. Three times in four you are wrong and lose 1.
- Gain: 0.25 × (+4) = +1.00
- Loss: 0.75 × (−1) = −0.75
- Net expected value: +0.25 marks per random guess
So even answering at random is very slightly better than leaving it blank. The penalty is not steep enough to make abstention correct.
It improves sharply with elimination
You are rarely guessing from four. Ruling out even one option changes the picture considerably.
- Three plausible options: (1/3 × 4) + (2/3 × −1) = +0.67
- Two plausible options: (1/2 × 4) + (1/2 × −1) = +1.50
Eliminate one option and a guess is worth two-thirds of a mark. Eliminate two and it is worth a mark and a half.
The honest caveats
This arithmetic assumes your guess really is uninformed rather than systematically wrong. If you hold an active misconception, you are not choosing at random — you are being reliably steered to a particular distractor, and the real probability of being correct is below one in four. That is precisely why confident errors matter more than uncertain ones.
It also assumes the time is free. A minute spent agonising over a question you will guess anyway is a minute stolen from questions you could actually answer, and in a 42-minute section that trade is usually bad.
A workable rule
- If you can eliminate anything at all, answer. The maths is clearly in your favour.
- If you can eliminate nothing but the answer costs you no time, answer anyway — it is still positive.
- Never spend more time on a question you cannot narrow. Mark, guess, move.
- Leave a blank only where you suspect your instinct is actively wrong on that topic.
Why the intuition runs the other way
Losing a mark feels considerably worse than gaining one feels good, so the penalty looms larger than the reward even when the arithmetic says otherwise. That asymmetry is a well-known feature of how people weigh losses, and here it produces a systematic, quiet cost across a whole paper.
Deciding the policy in advance is the remedy. Working out the expected value at question 140, tired, with four minutes left in a section, is not a decision you will make well.
The one case for leaving a blank
If you suspect an active misconception on a topic — this is a subject where you have been confidently wrong before — your guess is not random. You are being steered towards a particular distractor, and your real chance of being correct may be below one in four, which flips the arithmetic. This is a narrow case, and it is another reason to know which topics those are before exam day.
Time is the scarcer resource
At roughly 70 seconds a question, a minute spent agonising over a question you will guess anyway is close to a whole other question forgone. The expected value of a guess is a quarter of a mark; the expected value of a question you actually know is four. The arithmetic on time is more decisive than the arithmetic on guessing.
Guess quickly and move. The mark you lose to hesitation is larger than the one you are protecting.
Practise with the same scheme
Practising without negative marking builds a different set of instincts. If your sessions do not penalise wrong answers, you will not develop a feel for when a guess is worth making, and you will arrive on the day using intuition calibrated to a different game.