Andaman & Nicobar Islands Forest Guard – Marking Scheme
Advanced, MIT‑Level Treatment
Audience – Aspirants preparing for the Forest Guard recruitment exam, need a rigorous understanding of how marks are awarded, why the scheme is designed that way, and how to optimally answer questions.
Goal – Provide a principled, first‑principles view of the marking scheme, explore edge cases, present mathematical tools to evaluate strategies, and challenge you with a problem that blends probability, decision theory, and time management.
1. Exam Overview (Quick Refresher)
| Section | Total Marks | Question Type | Time (approx.) | Negative Marking |
|---|---|---|---|---|
| General Knowledge | 80 | 20 MCQs (4 options) | 30 min | –0.25 per wrong |
| English | 80 | 20 MCQs (4 options) | 30 min | –0.25 per wrong |
| Reasoning | 80 | 20 MCQs (4 options) | 30 min | –0.25 per wrong |
| Total | 240 | 60 MCQs | 90 min | –0.25 per wrong |
Key Assumptions
- All questions are single‑answer multiple choice.
- Answers are closed‑book; no partial credit.
- No “blank” marks for unanswered questions.
2. First‑Principles Derivation of the Marking Scheme
2.1 Why Negative Marking?
- Eliminate Guessing: Without penalty, candidates will answer all questions, inflating scores through random chance.
- Fairness: Penalizes wrong answers, rewarding knowledge over luck.
- Statistical Effect: For a 4‑option MCQ, random guessing yields a 25 % success rate → expected marks per question = (0.25 \times 1 + 0.75 \times (-0.25) = 0). Thus, the expected score of a purely random guess is zero.
Mathematical Formulation:
Let (p) be the probability a candidate knows the answer. Then:
[ E[\text{marks per Q}] = p \cdot 1 + (1-p) \cdot \Big(-\frac{1}{4}\Big) = p - \frac{1-p}{4} ]
Setting (E=0) gives (p = \frac{1}{4}). Hence, only when a candidate’s confidence > 25 % does it become rational to answer.
2.2 Score Calculation
For a section of 20 questions:
[ \text{Score} = N_{\text{correct}} \times 1 + N_{\text{wrong}} \times \Big(-\frac{1}{4}\Big) ]
where (N_{\text{correct}} + N_{\text{wrong}} = 20).
The maximum score is 20; the minimum (if all answered wrong) is (-5). Since no negative total scores are reported, the exam likely reports the raw score; the candidate’s net score is what matters for cut‑offs.
3. Edge Cases & Practical Nuances
| Edge Case | What Happens | Strategy Implication |
|---|---|---|
| Unanswered Q | 0 marks (neither positive nor negative) | Avoid if (p \le 0.25). |
| Ambiguous Question | 0 marks if left blank (no penalty) | If confidence < 25 %, skip. |
| Low-Stakes Section | Some sections may carry more weight in cut‑offs. | Allocate more time to high‑weight sections. |
| Time‑pressure | Unfinished sections may be auto‑graded as wrong. | Prioritize finishing all questions. |
| Wrong Answer vs. Blank | Wrong → -0.25; Blank → 0 | Prefer blank over wrong if (p < 0.25). |
4. Decision‑Theoretic Treatment
4.1 Bayesian Confidence & Expected Value
Let (C_i) denote the candidate’s confidence (probability that the answer is correct) for question (i).
- Expected Mark for Question (i):
[ E_i = C_i \cdot 1 + (1-C_i)\Big(-\frac{1}{4}\Big) = C_i - \frac{1-C_i}{4} ]
- Optimal Decision Rule:
[ \text{Answer if } E_i \ge 0 ;; \Longleftrightarrow ;; C_i \ge 0.25 ]
Thus, candidate should answer only if their confidence exceeds 25 %. This rule is independent of the question's actual difficulty; it depends only on personal confidence.
4.2 Time Allocation Model
Assume total available time (T = 90) min, and each question requires on average (t_i) minutes to answer. We can formulate a knapsack problem:
Maximise total expected marks:
[ \max_{\text{subset }S \subseteq {1,\dots,60}} \sum_{i \in S} E_i