RBI Assistant – Foundation & Exam Strategy
Topic: Marking Scheme
Audience: Advanced students preparing for the RBI Assistant exam (Banking/Financial Services).
Goal: Understand the first‑principles of the marking scheme, derive rigorous strategies, and tackle edge cases.
Output: A set of study notes in Markdown, complete with theory, formulas, and a challenging problem.
1. Overview of the RBI Assistant Exam
| Feature | Detail |
|---|---|
| Exam type | Paper‑based (paper‑and‑pen) |
| Sections | 1. General Awareness (GA) 2. English Language & Logical Reasoning (ELR) 3. Mathematics & Economics (ME) |
| Total marks | 200 (subject to revision; 100 per section) |
| Time | 3 hours |
| Marking | +2 for correct answer, –0.5 for wrong answer, 0 for blank (no penalty for skipping) |
| Answer pattern | Multiple choice (single best answer). |
| Marking scheme | Fixed; no partial credit. |
Why the marking scheme matters
The scheme determines expected value of each attempted question, shapes risk‑management strategies, and influences time allocation.
2. First Principles: How the Scheme Is Derived
-
Total Score & Difficulty Balance
- Let (N) = number of questions per section (≈ 50).
- Let (C) = marks for a correct answer, (W) = marks deducted for a wrong answer.
- The exam designers choose (C) and (W) so that the average expected score per question (under random guessing) is close to zero – discouraging blind guessing.
-
Random Guessing Expectation
For a 4‑option MCQ, probability of a correct guess (p = 1/4).
Expected mark per question:
[ E = pC + (1-p)(-W) ] Setting (E = 0) gives (C = 4W).
RBI uses (C=+2), (W=0.5) → (C = 4W) satisfies the balance. -
Score Distribution
- Suppose a candidate answers (k) questions correctly and (m) incorrectly.
- Total marks: (M = 2k - 0.5m).
- Since (k+m \le N), the maximum achievable marks is (2N) (all correct); minimum is (-0.5N) (all wrong).
-
No Partial Credit
- Simplifies evaluation: binary outcome per question, eliminates ambiguity.
- Encourages confidence‑driven answering rather than “guessing to fill blanks.”
3. Rigorous Analysis of Expected Performance
3.1 Expected Marks for a Single Question
| Ability | Probability Correct | Expected Mark |
|---|---|---|
| Strong (≥ 70 % accuracy) | (p=0.7) | (E = 0.7(2) + 0.3(-0.5)= 1.35) |
| Moderate (≈ 50 % accuracy) | (p=0.5) | (E = 0.5(2)+0.5(-0.5)=0.75) |
| Weak (≈ 30 % accuracy) | (p=0.3) | (E = 0.3(2)+0.7(-0.5)=0.05) |
| Random (1/4) | (p=0.25) | (E = 0.25(2)+0.75(-0.5)=0) |
Interpretation
A candidate with > 70 % accuracy per question gains ~1.35 marks per question on average.
Even a 50 % accurate candidate gains 0.75 marks per question – the scheme favors knowledge over guessing.
3.2 Variance & Risk
For a single question, variance:
[
\sigma^2 = p(C-\mu)^2 + (1-p)(-W-\mu)^2
]
where (\mu = E).
Higher variance indicates risk; a weak candidate sees high variance and should skip.
3.3 Total Expected Score
If a candidate answers (q) questions correctly and (r) incorrectly: [ E_{\text{total}} = 2q - 0.5r ] Given a confidence threshold (t) (probability that the answer is correct), the optimal number of attempted questions can be derived by maximizing (E_{\text{total}}) subject to time constraints.
4. Advanced Strategy: Risk‑Return Optimization
4.1 Problem Formulation
-
Variables
- (x_i \in {0,1}) – whether to attempt question (i).
- (p_i) – candidate’s personal probability of answering question (i) correctly (established through prior practice).
-
Objective
Maximize expected marks:
[ \max_{x} \sum_{i=1}^{N} x_i \bigl( 2p_i - 0.5(1-p_i) \bigr) ] subject to (\sum_{i} x_i \leq T) (time limit), and (x_i \in {0,1}). -
Solution
Sort questions