1.3 Mark Calculation, Negative Marking (0.25) & Cut-offs

Mark Calculation, Negative Marking (0.25) & Cut-offs

Mastering the scoring mechanics of the SBI PO examination is essential for building an optimal test-taking strategy. Because competitive exams are decided by narrow margins — often fractions of a single mark — understanding negative marking penalties, multi-shift normalization, accuracy mathematical expectations, and historical cut-off trends separates successful candidates from the rest.

---

1. Objective Scoring Rules & Penalty Structure

In both Phase I (Prelims) and Phase II (Mains) objective papers, marks are awarded or deducted based on the following standard rules:

1. Correct Answer: Full positive mark allocated to the question (e.g., $+1.00$ mark in Prelims; variable $+1.00$, $+1.20$, or $+1.67$ marks in Mains).
2. Incorrect Answer: Penalty of one-fourth (25% or 0.25) of the marks assigned to that specific question.
3. Unattempted / Blank Question: Zero (0.00) marks awarded or deducted.

Penalty Matrix across Marks Weightage

General Net Score Calculation Formula

For any objective test section with total attempted questions $A$, correct answers $C$, and wrong answers $W$ (where $A = C + W$):

$\text{Net Score} = \sum_{i=1}^{C} \text{Marks}_i - \sum_{j=1}^{W} \left( 0.25 \times \text{Marks}_j \right)$

For a standard 100-question Prelims test where every question carries 1 mark: $\text{Net Score} = \text{Correct Answers} - (0.25 \times \text{Wrong Answers})$

---

2. Mathematical Analysis: Blind Guessing vs. Calculated Elimination

In a 5-option multiple-choice framework (standard in SBI PO objective tests), let us analyze the probability and expected value ($E$) of guessing tactics.

+-----------------------------------------------------------------------------------+
|                         EXPECTED VALUE OF GUESSING MATRIX                         |
+-----------------------------------------------------------------------------------+
| Scenario 1: Random Blind Guess (5 Options Available)                              |
| P(Correct) = 1/5 = 0.20  |  P(Wrong) = 4/5 = 0.80                                  |
| E = (0.20 * +1.00) + (0.80 * -0.25) = +0.20 - 0.20 = 0.00 MARKS                   |
+-----------------------------------------------------------------------------------+
| Scenario 2: Eliminate 1 Wrong Option (4 Options Remaining)                       |
| P(Correct) = 1/4 = 0.25  |  P(Wrong) = 3/4 = 0.75                                  |
| E = (0.25 * +1.00) + (0.75 * -0.25) = +0.25 - 0.1875 = +0.0625 MARKS              |
+-----------------------------------------------------------------------------------+
| Scenario 3: Eliminate 2 Wrong Options (3 Options Remaining)                       |
| P(Correct) = 1/3 = 0.333 |  P(Wrong) = 2/3 = 0.667                                  |
| E = (0.333 * +1.00) + (0.667 * -0.25) = +0.333 - 0.1667 = +0.1667 MARKS           |
+-----------------------------------------------------------------------------------+
| Scenario 4: Eliminate 3 Wrong Options (2 Options Remaining)                       |
| P(Correct) = 1/2 = 0.50  |  P(Wrong) = 1/2 = 0.50                                  |
| E = (0.50 * +1.00) + (0.50 * -0.25) = +0.50 - 0.125 = +0.375 MARKS                |
+-----------------------------------------------------------------------------------+

Strategic Takeaway on Guessing

• Pure Blind Guessing (5 options): Expected value is exactly 0.00. Over a large sample (e.g., 20 blind guesses), you gain no net advantage and risk negative variance.
• Calculated Guessing (Eliminating 2 or 3 options): Yields a strongly positive expected value ($+0.1667$ to $+0.375$ marks per question). When you can confidently eliminate 2 or more choices, taking an educated guess is statistically advantageous.

---

3. Score Normalization: The Equipercentile Method

Because SBI PO Prelims and Mains are administered across multiple operational shifts over several days, minor variations in difficulty level between question paper sets are inevitable. To ensure absolute fairness, raw marks are converted into normalized marks using the Equipercentile Normalization Method.

How Equipercentile Normalization Works

1. Percentile ranks are calculated for candidates in each separate shift based on raw scores.
2. A candidate's raw score is mapped to the score that corresponds to the exact same percentile rank in a standard reference shift.
3. The percentile rank formula for candidate $i$ with score $S_i$ in a shift of $N$ candidates is:

$\text{Percentile Rank} = \left( \frac{\text{Number of candidates in shift with raw score } \le S_i}{N} \right) \times 100$

Impact: If Shift A was exceptionally difficult, a raw score of $54/100$ might place a candidate in the 92nd percentile. If Shift B was easier, a raw score of $61/100$ might be needed for the 92nd percentile. After normalization, both candidates receive equal normalized scores.

---

4. Historical Aggregate Cut-Off Trends

Because SBI PO Prelims has no sectional cut-offs, Prelims qualification depends entirely on clearing the overall aggregate cutoff score. In Mains, however, candidates must also clear a minimum qualifying mark in each section (Tests I–IV) and the Descriptive paper individually, in addition to the overall aggregate cutoff.

Representative Historical Cut-off Benchmarks

Note: Pre-2026 Mains was scored out of 250 (200 objective + 50 descriptive). From SBI PO 2026 onward, the Descriptive paper carries 30 marks, so Mains is scored out of 230 (200 objective + 30 descriptive). The pre-2026 Mains cut-offs below are therefore not directly comparable to the 2026 230-mark scale.

Key Cut-off Insights:

• Prelims Cut-off Band: For General candidates, the Prelims cut-off historically hovers between 59 and 67 marks out of 100 (recent cycles: 59.25 in 2023, 61.75 in 2024, 66.75 in 2025).
• Mains Cut-off Band: The pre-2026 Mains aggregate cut-off for General generally fell between 70 and 88 marks out of 250 (e.g., 70.00 in 2023, 87.50 in 2024, 75.00 in 2025). On the 2026 230-mark scale, expect a proportionally lower raw cut-off.
• Final Selection Target: To ensure final appointment, candidates should target a normalized final score of 50+ out of 100 to stay safely above recent General-category cut-offs (44.60 in 2023, 46.24 in 2024).