Mock Test Analysis and Speed Building
Key Takeaways
- Mock scores matter only when reviewed by section, error type, time cost, and penalty impact.
- Speed should be built through first-pass selection, not by rushing every question from the start.
- A wrong-answer ledger is more useful than a list of total scores because it shows which habits are costing normalized marks.
- CBT 1 mocks should test 100-question screening readiness, while CBT 2 mocks should train 120-question stamina and deeper recall.
Mocks Are a Feedback System
A mock test is useful only after it is reviewed. RRB NTPC has one-third negative marking in CBT 1 and CBT 2, so a raw attempt count can be misleading. A candidate who attempts 95 questions with many guesses may score worse than a candidate who attempts 78 questions with stronger accuracy. The review must show where marks were gained, lost, skipped, and wasted.
Separate CBT 1 and CBT 2 mocks. CBT 1 practice should train 100 questions in 90 minutes, with the 40-30-30 section balance. CBT 2 practice should train 120 questions in 90 minutes, with more General Awareness and more total arithmetic and reasoning load. The second format is not just longer; it punishes weak stamina.
Mock Review Grid
| Review field | What to record | Why it matters |
|---|---|---|
| Section score | GA, Maths, Reasoning | Shows imbalance |
| Error type | knowledge, method, slip, guess | Tells what to fix |
| Time cost | fast, normal, time drain | Finds slow topics |
| Confidence | sure, eliminated, guessed | Controls penalty risk |
| Repeat error | yes or no | Measures repair failure |
Do not record only the correct answer. Write why your answer happened. A wrong history answer caused by a mixed leader-movement pair needs different repair from a wrong percentage answer caused by base selection. A missed syllogism caused by reversing inclusion needs a rule fix, not more random practice.
Score the Mock Like the Exam
Use the real marking logic: plus one for correct, minus one-third for wrong, zero for blank. Then create two totals. The first is the actual mock score. The second is the recoverable score: marks lost to careless slips, copied values, skipped easy questions, or obvious time traps. Recoverable score shows how much improvement is available without learning a new chapter.
Example review:
| Result | Count | Mark effect |
|---|---|---|
| Correct | 72 | +72 |
| Wrong | 18 | -6 |
| Blank | 10 | 0 |
| Net | 100 attempted set | 66 |
If six of the wrong answers were no-elimination guesses, the fix is not more syllabus coverage. The fix is attempt discipline. If five blanks were easy questions left unseen because one puzzle consumed four minutes, the fix is first-pass routing.
Build Speed by Routing
Speed is not the same as reading faster. In NTPC, speed comes from knowing which questions deserve first-pass time. GA sure-shot facts should be answered quickly. Direct arithmetic should be solved if the method is visible. Reasoning items that need a small diagram should be attempted if the setup starts cleanly. Long puzzles, unfamiliar current affairs, and messy calculations should be marked and revisited.
Use a three-pass system:
- Pass one: sure and short questions only.
- Pass two: solvable marked questions requiring calculation or diagrams.
- Pass three: remaining risk decisions after checking time.
This system protects easy marks from being trapped behind one slow problem. It also reduces panic because the candidate has already collected known marks before entering review.
Section-Specific Speed Targets
General Awareness should usually save time. A known fact may take under 20 seconds. A half-known fact should not take one minute. If the anchor does not appear quickly, mark or skip. GA time saved can support Mathematics and Reasoning, but only if GA guessing is controlled.
Mathematics speed comes from topic recognition and neat rough work. Write the formula or relation first: rate, percentage base, average total, ratio parts, or distance. If the first method fails after 40 to 50 seconds, mark the question. Restarting the same arithmetic without a new method is a time leak.
Reasoning speed comes from setup discipline. Draw direction paths, family trees, seating slots, or Venn regions early. A small diagram prevents repeated reading. For series and analogy, name the rule before touching options. For puzzles, leave if no fixed placement appears after one full clue pass.
Error Ledger Categories
Use these categories after every mock:
- Knowledge gap: fact, formula, or rule not known.
- Recognition gap: topic known but not recognized in the stem.
- Method error: wrong setup, wrong base, wrong direction, wrong diagram.
- Calculation slip: arithmetic, sign, unit, or copying error.
- Attempt error: blind guess, overlong question, or unsafe change.
Every category needs a different repair. Knowledge gaps need revision. Recognition gaps need mixed practice. Method errors need worked examples. Calculation slips need slower rough-work habits. Attempt errors need a stricter skip rule.
Mock Frequency and Recovery
More mocks are not automatically better. A full mock without review can reinforce bad habits. In a normal phase, one full mock plus deep review is better than three unreviewed tests. In the final phase, increase frequency only if review quality remains high.
A good weekly cycle is: one full CBT 1 or CBT 2 mock, one section-specific timed test for the weakest area, one mixed set for attempt selection, and two short revision sessions based on the error ledger. After each mock, choose no more than three repair targets. Too many targets become noise.
Final Week Mock Rules
In the final week, reduce experimentation. Use mocks to confirm timing, not to invent a new order. Keep the same first-pass strategy, same rough-work style, and same confidence ladder. Review only high-value errors: repeated GA confusions, formula slips, diagram mistakes, and blind guesses.
Do not chase a perfect mock score. The real goal is stable scoring behavior under a 90-minute CBT with negative marking. A candidate who knows when to stop, mark, skip, and return has a practical advantage over a candidate who simply solves more questions in practice.
After a CBT 1 mock, a candidate has 74 correct, 16 wrong, and 10 blank responses. Six wrong responses were no-elimination guesses. What is the most useful review action?