Answer-Key Pattern Analysis (all 7 papers, Series A)
🎯 Answer-Key Pattern Analysis — 7 JKSSB Papers (2026)
This analyses the official Series-A answer key of 760 questions across all 7 papers for statistical tendencies of the option letters (A/B/C/D). These are the paper-setters' habits, not laws. With JKSSB's ¼ (0.25) negative marking, guessing purely by option-frequency is a losing strategy on average. Use this only as (1) a last-resort tie-breaker when you have already eliminated to a 50/50 and must pick, and (2) a study lens to notice how certain question formats are built. All 7 papers are set by the same board, so the biases below likely reflect one team's style.
Sample: 760 questions — WO 80 · WG 120 · IN 120 · JA 80 · GD 120 · LA 120 · SF 120. Answers = official provisional Series-A keys.
⭐ The cheat-sheet (headline findings)
- C is the most common answer overall (28.3%); D is the rarest (20.8%). A pure-random key would sit at 25% each — so C is over-represented and D is under-represented by a real margin (~7.5 points).
- D is never the mode in any single paper and never tops 23.3%. If you must blind-guess a whole paper, guess C; never default to D.
- Match-the-following / "correctly matched pairs" → the answer is C or D 71% of the time (C 40%, D 31%), and B only 11.5%. This is the single strongest pattern in the dataset.
- Section matters more than the paper: English & Science lean B, Maths/Reasoning leans A, Computers & JK-GK & National-GK lean C, and Current-Affairs is bimodal A-or-D (B/C almost never).
- "None of the above" is correct only 1 in 15 times (7%) — when you see it, it's almost certainly a trap. "All of the above" is correct only ~23% — the old "pick All of the above" hack does not work here.
1. Overall distribution (760 Q)
| Option | Count | % | vs 25% baseline |
|---|---|---|---|
| A | 191 | 25.1% | ≈ even |
| B | 196 | 25.8% | slightly high |
| C | 215 | 28.3% | ⬆ most common |
| D | 158 | 20.8% | ⬇ least common |
Rank: C > B > A > D. The gap between C (215) and D (158) is 57 questions — far beyond random noise.
2. By paper — where each letter peaks
| Paper | A | B | C | D | Mode | Rarest |
|---|---|---|---|---|---|---|
| WO (80) | 20.0 | 25.0 | 32.5 | 22.5 | C | A |
| WG (120) | 28.3 | 25.0 | 26.7 | 20.0 | A | D |
| IN (120) | 20.8 | 21.7 | 34.2 | 23.3 | C | A |
| JA (80) | 32.5 | 25.0 | 22.5 | 20.0 | A | D |
| GD (120) | 27.5 | 28.3 | 23.3 | 20.8 | B | D |
| LA (120) | 27.5 | 27.5 | 24.2 | 20.8 | A/B | D |
| SF (120) | 20.0 | 27.5 | 34.2 | 18.3 | C | D |
(values are % of that paper)
- C is the mode in 3 papers (WO, IN, SF) and hits 34% in IN & SF.
- A is the mode in 3 (WG, JA, LA-tie).
- D is the rarest in 5 of 7 papers and is bottom-two in all seven.
3. By question TYPE — the most useful cut
| Type | n | A | B | C | D | Takeaway |
|---|---|---|---|---|---|---|
| Match / pairs | 52 | 17.3 | 11.5 | 40.4 | 30.8 | Answer is C or D 71%. Avoid B. |
| Sequence / order | 21 | 23.8 | 19.0 | 28.6 | 28.6 | Lean C/D (57%); B weakest. |
| Statement-based | 94 | 27.7 | 24.5 | 27.7 | 20.2 | Fairly even; slight A/C tilt, D low. |
| Single / direct | 585 | 25.5 | 27.5 | 27.5 | 19.5 | B/C tilt; D clearly lowest. |
| Assertion–Reason* | 5 | 40 | 40 | 0 | 20 | Too few to trust. |
- Match-the-following: if stuck, pick C first, then D. Setters tend to place the "all correct / most correct" mapping in the later options. B is a near-dead option here (11.5%).
- Everything except match/sequence: D is the least likely — break a tie away from D.
- "Correctly matched pairs" with an "All of the above" (D): D wins ~31% here vs 23% globally — slightly more tempting but still not a majority.
4. By SECTION — each subject has its own favourite
| Section | n | A | B | C | D | Section mode |
|---|---|---|---|---|---|---|
| English | 155 | 25.2 | 32.3 | 26.5 | 16.1 | B (D very low 16%) |
| General-Science | 56 | 26.8 | 30.4 | 21.4 | 21.4 | B |
| Maths / Reasoning (unbanked) | 307 | 29.0 | 27.0 | 25.7 | 18.2 | A (D low) |
| Computers | 70 | 20.0 | 25.7 | 37.1 | 17.1 | C |
| JK-GK | 54 | 18.5 | 14.8 | 38.9 | 27.8 | C (B very low 15%) |
| National-World-GK-History-Civics | 96 | 15.6 | 18.8 | 35.4 | 30.2 | C (A low; C+D = 66%) |
| Current-Affairs | 22 | 40.9 | 9.1 | 9.1 | 40.9 | A or D (B/C almost never!) |
Section rules of thumb (last-resort guessing):
- English / Science → lean B, avoid D.
- Maths / Reasoning → lean A, avoid D.
- Computers / JK-GK / National-GK-History → lean C (these are match/statement-heavy, which pulls answers to C).
- Current-Affairs → it's A or D — B and C together were correct only 4 times out of 22. If torn between a middle and an edge option in CA, pick the edge.
Current-Affairs (n=22) and General-Science (n=56) are small — treat their splits as suggestive, not solid. The Maths/Reasoning bucket is everything not yet in a topic bank (≈ numerical + mental-ability + a few stragglers).
5. "All of the above" / "None of the above"
| Trap option | Times it appeared | Times it was correct | Hit-rate |
|---|---|---|---|
| All of the above | 47 | 11 | 23% |
| None of the above | 15 | 1 | 7% |
- The classic "when unsure, choose All of the above" trick fails here — it's right less than a quarter of the time.
- "None of the above" is almost always a decoy — safe to deprioritise unless you're confident.
6. Streaks & sequence behaviour
- Longest identical run seen: 5 in a row (WG had A-A-A-A-A; IN had C-C-C-C-C). Runs of 4 appear in several papers. → Don't second-guess a correct answer just because it repeats your last 3–4. Long same-letter runs are normal.
- Transition tendency is weak (mild alternation): after a B, the next answer is most often C (34%); after C, most often B (30%). This is too soft to bet on — mentioned only for completeness.
7. Practical last-resort decision tree
When you've eliminated what you can and still must guess:
Is it Match-the-following / pairs? → pick C (then D). Never B.
Is it Current-Affairs? → pick A or D (an edge option). Never B/C.
Else, by section:
English / Science → B
Maths / Reasoning → A
Computers/JK/Nat-GK → C
No idea at all? → pick C. And never blind-pick D.
Contains "None of the above"? → it's probably NOT that.
Every one of these edges is worth only a few percent. With −0.25 marking, the break-even for a 4-option blind guess is exactly 25% correct — these tilts (28% for C) barely clear it, and only before accounting for the questions you can partly reason through. The real marks are in preparation, not pattern-gaming. Use this page to shave uncertainty on the last 2–3 genuinely unknown questions, not as a shortcut.
8. Method & limitations
- Source: the 7
*-Answer-Key.mdfiles (Series-A column) + the 7*-Question-Paper.mdfiles, parsed programmatically. - Type classification is regex-based on question wording (keywords: match/list/pairs, arrange/chronological, assertion/reason, statements); 3 questions couldn't be block-matched due to OCR and are excluded from the type cut only.
- Section mapping comes from the six topic banks in
Question-Bank/; everything not yet banked (mainly Maths & Reasoning) is grouped as "unbanked". - Caveats: provisional keys (a handful of later revisions/discards exist and are not netted out — negligible at this scale); single exam board so biases are correlated; sub-sections have small n. The C-high / D-low tendency is the only pattern strong and consistent enough to lean on.
🔗 Related
- Section banks: Computers · JK-GK · Current-Affairs · English · General-Science · National-World-GK-History-Civics
- FAA-Study-Plan · FAA-Prediction-and-Focus-Analysis · Home
