📚 Study Notes – Marking Scheme
Himachal Pradesh Patwari / Lekhpal – Foundation & Exam Strategy
Audience: Advanced students preparing for the HP Patwari / Lekhpal exam
Level: MIT‑style, first principles, rigorous reasoning
Goal: Understand the exact marking scheme, model its statistical properties, and derive an optimal answering strategy.
1. Introduction
The Patwari / Lekhpal Foundation exam is a multiple‑choice, paper‑based test.
The marking scheme is the linchpin of the exam strategy: it determines how you should allocate your time and energy across questions.
Why formal analysis matters
- The scheme is not purely additive; it contains negative marks that create a risk component.
- A rigorous treatment allows you to compute expected score and variance, giving you a confidence interval for your performance.
- It also helps in edge‑case handling (e.g., ambiguous questions, tie‑breaking rules).
2. Overview of the Exam
| Item | Detail |
|---|---|
| Paper format | 100 MCQs, 3 options (A, B, C) each |
| Question type | Mixed: factual, situational, calculation (simple) |
| Time | 1.5 hours (90 minutes) |
| Marking | +3 for correct, –1 for incorrect, 0 for unanswered |
| Scoring range | 0 – 300 (theoretical maximum) |
| Passing criteria | Usually ≥ 33% of total marks (≈100 marks) |
Assumption: No “partial credit” or “negative marking for skipped” in the standard scheme.
3. Marking Scheme Fundamentals
3.1 Basic Formula
For each question (i) (1 ≤ i ≤ 100):
[ \text{Score}_i = \begin{cases} +3, & \text{if answer } A_i \text{ is correct} \ -1, & \text{if answer } A_i \text{ is incorrect} \ 0, & \text{if left unanswered} \end{cases} ]
Total score (S) is the sum of all (\text{Score}_i).
3.2 Expected Value & Variance
Let (p_i = P(\text{student knows the correct answer for } i)).
Assume the student guesses if not sure.
- If answered correctly: probability (p_i).
- If guessed incorrectly: probability ((1-p_i)\times \frac{1}{2}) (since 2 wrong options).
Hence,
[ E[\text{Score}_i] = 3p_i - 1(1-p_i)\frac{1}{2} = 3p_i - \tfrac{1}{2}(1-p_i) ]
[
\text{Var}[\text{Score}_i] = E[\text{Score}_i^2] - (E[\text{Score}_i])^2
]
where
(E[\text{Score}_i^2] = 9p_i + 1(1-p_i)\tfrac{1}{2}).
4. Detailed Breakdown
| Question | Correct | Incorrect | Unanswered |
|---|---|---|---|
| Mark | +3 | -1 | 0 |
| Probability (if guessed) | ( \tfrac{1}{3} ) | ( \tfrac{2}{3} ) | 0 |
| Expected contribution (guessing) | ( +1 ) | ( -\tfrac{2}{3} ) | 0 |
| Net | (+1) | (-\tfrac{2}{3}) | 0 |
4.1 Why Guessing is Generally Beneficial
If you know the answer → +3.
If you guess → expected +1 > 0.
Therefore, always answer a question if you cannot be certain you are wrong.
However, if you know you are wrong, skip.
5. Edge Cases & Advanced Theory
| Edge Case | Handling | Rationale |
|---|---|---|
| Multiple correct answers | None in standard HP Patwari exam. | If existed, scoring would change to +3 for each correct option, -1 for each wrong. |
| Ambiguous questions | Treat as “unknown”; guess. | Standard policy: no partial credit. |
| Time‑pressure | If time < 90 min, adjust strategy: answer all low‑risk questions first. | Risk‑management: maximize expected score per minute. |
| Tie‑breaking | If two answer keys produce same score, the official system uses the official key. | Implication: cannot rely on “average” scoring. |
| Question difficulty distribution | Not uniform; some questions are “hard” (low (p_i)). | Use difficulty indices to adjust answering strategy. |
5.1 Statistical Model: Markov Decision Process
Define state (s_t) as the set of remaining unanswered questions at time (t).
Action (a_t) ∈ {answer, skip}.
Reward (R(s_t, a_t)) = expected score contribution.
The optimal policy maximizes (E\left[\sum_{t=1}^{T} R(s_t, a_t)\right]).
Under the independence assumption, the optimal policy reduces to a greedy rule: answer if (E[\text