UPSC Prelims Negative Marking: Score Calculation and Attempt Decisions

Calculate UPSC Prelims negative marking with worked examples, compare attempt decisions, and build a confidence log from your mock tests.

Negative marking changes the value of an attempt: answering more questions does not automatically mean scoring more marks. For UPSC Civil Services Prelims, start with the scoring rule, then use your own mock-test evidence to decide which uncertain questions deserve another look. This guide explains the arithmetic without promising a qualifying score or a universal “safe” number of attempts.

What the official rule says

The UPSC Civil Services 2026 notification specifies a penalty of one-third of the marks assigned to a question for an incorrect answer. Marking more than one answer also attracts that penalty, even if one is correct. Leaving a question blank attracts no penalty. Check the notification and question-booklet instructions for your own exam cycle; do not apply this rule to every exam conducted by UPSC.

For a question worth 2 marks, the mathematical one-third penalty is 2/3 of a mark. Retain the fraction when practising calculations rather than rounding each deduction to 0.66. Official scoring and answer-key treatment remain authoritative.

Calculate a mock score correctly

For a practice paper with 2 marks per correct answer, use: score = 2 × correct answers − (2/3) × wrong answers. Blank answers contribute zero.

Practice resultCorrectWrongBlankCalculated score
Student A602020120 − 13.333… = 106.667…
Student B603010120 − 20 = 100
Student C632710126 − 18 = 108

These are illustrative results, not cutoff predictions. A and B have the same correct-answer count, but B loses more marks through wrong attempts. C attempts as many questions as B yet scores more because accuracy improves. Record correct and wrong counts alongside total attempts in your Scholareum mock-test practice.

When does an uncertain attempt have positive expected value?

Suppose your probability of being correct is p. With +2 for correct and −2/3 for wrong, the expected mark is 2p − (2/3)(1 − p). It becomes positive when p is greater than 1/4.

That is a mathematical model, not an instruction to guess. You cannot reliably assign a probability merely because an option feels familiar. The model also ignores time cost: spending several minutes on a marginal attempt can prevent you from answering an easier question.

With four equally plausible options, a random guess has zero expected value under this model. Eliminating one genuinely incorrect option leaves a theoretical one-in-three chance and positive expected value, but only if the remaining options are equally plausible and the elimination is sound. False elimination can make the decision worse. Use the evidence-based option-elimination guide to separate a reason from a hunch.

Build a confidence log over three mocks

Before looking at the key, label every attempted question: certain, eliminated with evidence, or guess. After checking the answer, count correct and wrong answers within each group. A practical row might read: “Eliminated with evidence: 12 correct, 8 wrong; accuracy 60%; reason for each rejected option recorded.”

Keep the labels consistent across three tests. If your “certain” group contains many wrong answers, review conceptual gaps or careless reading first. If elimination errors repeat, write down the incorrect assumption behind each rejected option. Follow the mock analysis routine rather than simply increasing your attempt target.

A repeatable review sequence

  1. Complete questions for which you have a clear route to the answer.
  2. Revisit uncertain questions only when you can state supporting or eliminating evidence.
  3. Check words such as “only”, “not” and “incorrect” before finalising an answer.
  4. Keep your answer-sheet entries consistent with question numbers.
  5. Review the result by confidence group, not just by total score.

Use the test-series catalogue to find relevant practice and consult the exam calendar alongside official announcements. For comprehension practice, pair this method with the CSAT inference and assumption guide.

Frequently asked questions

Is a blank answer penalised?

Not under the cited Civil Services Prelims rule. An unanswered question contributes zero.

Is there a fixed number of attempts that guarantees qualification?

No. Correct answers, wrong answers, paper difficulty and the Commission’s qualifying requirements matter. An attempt count alone cannot establish your score or qualification.

Should I round each penalty before adding it?

For practice calculations, retain one-third until the final result. Rounding each deduction separately introduces avoidable arithmetic differences. Follow official result calculations when interpreting an actual exam score.

Does this guide replace the official notification?

No. It is a practice framework based on the cited 2026 rule. Use UPSC’s official website for notifications, papers and answer-key updates.