CETKING CONTENT UNIVERSES

Master Architecture & Memory for Claude

Version: 2.0 — 8 October 2026
System: Cetking Questions Universe + G Strategy Universe + Mocky Universe
Scope: Quantitative Aptitude, Verbal Ability & Reading Comprehension, and Data Interpretation & Logical Reasoning

1. PURPOSE OF THIS MEMORY

You are helping build, maintain and extend Cetking's examination-content ecosystem.

This document is the authoritative working context for three academic areas:

1. Quant Universe
2. Verbal Universe
3. Logic / DILR Universe

These areas operate within the wider Cetking architecture.

The core rule:

Every question belongs to Questions Universe. Every solving shortcut or branded method belongs to G Strategy Universe. Every student attempt and performance result belongs to Mocky Universe.

Do not create disconnected question banks for the same questions.

Wider Cetking architecture

Do not rename database tables merely because the conceptual Universe name has changed.

2. CORE QUESTION ARCHITECTURE

Cetking has three major content layers.

Layer A — Normal / Foundation Questions

These questions primarily develop one concept.

Examples:

• Averages → Find Missing Value
• Geometry → Circles → Chords
• Algebra → Linear Equations
• DILR → Matrix Puzzles

They are taught and tested under the standard taxonomy.

Layer B — G Strategy Questions

These are questions with an identifiable Cetking branded solving method, shortcut or smart technique.

A question can have:

• Normal solution
• G Strategy shortcut solution
• More than one suitable G Strategy

G Strategy does not replace the normal mathematical or logical solution.

Layer C — 3D CAT Questions

These are CAT-level questions, especially historical CAT PYQs, involving multiple concepts.

Example:

Coordinate Geometry + Area + Algebra

They must not be forced into one normal topic classification.

A 3D question has:

• Its own historical/catalog identity
• A primary visible topic
• Multiple linked concepts
• Optional G Strategy references
• Normal and shortcut solutions where applicable

Fundamental distinction

Normal Questions = Concept Mastery

G Strategy Questions = Method Mastery

3D CAT Questions = Integrated Application Mastery

These are overlapping characteristics, not necessarily mutually exclusive database categories.

3. PERMANENT QUESTION IDENTITY

Use separate identifiers for:

Canonical Question ID

The permanent identity of a question in public.question_bank.

Example:

Q-1025

PYQ Historical ID

Identifies an appearance in a particular examination, year, slot and section.

Example:

CAT-2025-S1-QA-001

The original spreadsheet serial number may not equal the original examination question number. Preserve source sequence and verified exam question number separately.

Concept ID

Links a question to one or more approved core concepts.

G Strategy ID

Links a question to one or more approved G Strategy methods.

Test/Assessment ID

Represents a Mocky test or assessment into which the question is placed.

Never duplicate the canonical question merely to show it in another test, book, chapter, PYQ page or mock. Published test records can hold controlled snapshots.

4. STANDARD TAXONOMY

Normal question structure:

Section → Sub Section → Topic → Sub Topic → Core Type

The detailed core-type ID is especially important in Quant.

Example:

Quant → Arithmetic → Averages → Combined Average → Weighted Group Average

Another example:

Logic → Arrangements → Matrix based Puzzles → Single Matrix Puzzles

A question can also carry:

• Difficulty
• Difficulty DNA
• Strategy family
• Source exam/year/slot
• Normal solution
• G Strategy solution
• Media
• Review status
• Publishing status

For 3D questions, use a separate catalog/index and concept-linking table rather than squeezing every linked concept into one sub_topic_name.

5. QUANT UNIVERSE

5.1 Historical CAT Quant inventory

Source:

CAT Master Planning sheet 2.0.xlsx → QAall

Historical coverage:

CAT 2021–2025

330 questions = 66 per year

The original QAall source contains nine columns:

Keep every original field, even where normalized values are also stored.

5.2 Quant CAT sub-sections

These six labels are the frozen broad CAT PYQ analysis categories.

These are historical analysis categories, not necessarily the exact official CAT syllabus taxonomy.

Sample topics within the six sections

Arithmetic

• Profit & Loss
• Averages
• Mixtures
• Percentages
• Ratios

Modern Maths

• Inequality
• Progression
• Logarithm
• Permutation & Combination
• Functions

Algebra

• Linear Equations
• Quadratic Equations
• Polynomials
• Algebraic Expressions

Geo Mensu

• Geometry
• Mensuration
• Coordinate Geometry
• Trigonometry

TS&W

• Time, Speed & Distance
• Time & Work
• Boats, Trains and related applications

Numbers

• Divisibility
• Remainders
• Number of Solutions
• LCM/HCF
• Factors

Do not assume this short list captures every value in the master taxonomy.

6. QUANT G STRATEGY

G Strategy is Cetking's official brand for solving shortcuts and smart methods.

Historical CAT 2021–2025 classification:

Critical rule: Concepts No Shortcut

Concepts No Shortcut is not a G Strategy.

It means:

Use the normal conceptual method to solve the question.

Use:

g_strategy_status = no_shortcut

g_strategy_id = NULL

Do not create a branded shortcut for this category.

Shortcuts topic wise

This is a placeholder/classification indicating topic-specific shortcuts.

It is not yet a final canonical strategy identity.

Use:

g_strategy_status = pending_review

Naming rules

Always use:

• Breakup G Strategy
• Visual G Strategy
• X Maro G Strategy
• Vedic G Strategy

Never use "X Marco." Normalize X maro and X Maro to X Maro.

7. QUANT 3D QUESTION ARCHITECTURE

CAT questions may combine:

• Inequality
• Area
• Coordinate Geometry
• Algebra
• Ratios
• Number Theory
• Functions

One question may use several of these simultaneously.

Recommended concept structure

The source spreadsheet Topic/Sub Topics should initially serve as primary visible tags.

Additional concepts should be attached through relational links and verified against the full question.

CAT PYQ identification

When a new question is received:

1. Normalize formatting.
2. Compare original numerical values and variables.
3. Search historical source stems.
4. Search year/slot and distinctive expressions.
5. Identify candidate matches.
6. Compare full conditions, options, diagrams and answer.
7. If exact, link to the existing PYQ.
8. If similar but modified, mark as a derivative/variant.
9. If uncertain, send to review.
10. If new, create a fresh canonical question identity.

Do not auto-label a merely similar question as a verbatim CAT PYQ.

8. VERBAL UNIVERSE

8.1 Source database

Historical coverage:

CAT 2019–2025

506 question records

Original source columns:

1. Year
2. Slot
3. RC/NON
4. Passage
5. Essay Sub Type
6. Essay Type
7. Reading Ease
8. Full Question
9. Direct / Indirect
10. G Strategy Question
11. RC / CR
12. Except
13. G Strategy Type of Mistake

Preserve original labels and spelling in raw source data.

8.2 Historical RC / Non-RC distribution

Historical annual totals

9. FOUR CETKING VERBAL G STRATEGY SECTIONS

These are the four promoted student-facing categories.

The exact names are:

1. Summary
2. Paragraph
3. Evidence
4. Critical Reasoning

Do not rename Evidence back to Direct RC or Passage Evidence in the UI.

The internal code can remain PASSAGE_EVIDENCE for compatibility.

Section mapping

Do not add Ques or Percentage columns to this clean classification table.

Question counts and percentages belong in separate summary tables.

Summary

Includes:

• Summary
• TSPM

TSPM represents the Cetking main-idea/passage-summary approach. Preserve the exact source label TSPM; do not expand it into an invented acronym.

Paragraph

Includes:

• Odd Man Out
• Jumbled
• Completion

Evidence

Previously called Direct RC / Passage Evidence.

Includes:

• Why Mention This
• Truth Facts

The label highlights directly text-supported reasoning and evidence from the passage.

Critical Reasoning

Includes:

• Inference
• Weakening Strengthening
• Cause Effect Reasoning
• Similarity

10. VERBAL G STRATEGY QUESTION FREQUENCY

Historical CAT 2019–2025:

Reading Ease

Source labels and counts:

Keep Medium and Moderate distinct in original source data. They may be normalized later through an explicit mapping, but must not be silently combined.

Verbal mistake taxonomy

Important mistake types include:

• Opposites
• Opposite of Passage
• Out of Scope
• Distorted Facts
• Partial Reasoning
• Example/Analogy Confusion
• Scope Mismatch
• Extremist Language
• Lack of Context
• Omitting Societal Impact
• Logical Reversal
• Incorrect Cause/Effect
• Hidden Premise
• Overgeneralization

The source has many more detailed labels, some of which are variants, compound labels or question-specific explanations.

Do not automatically treat every unique source string as a new canonical mistake type.

Keep raw mistake strings and map them gradually into stable families.

The mistake taxonomy is important for:

• Option elimination
• Distractor design
• Student error analysis
• Pulse/Mocky weakness reporting
• Rav Singh AI recommendations

11. LOGIC / DILR UNIVERSE

11.1 Historical source

Historical coverage:

CAT 2017–2025

558 question records

Original fields:

1. Year
2. Slot
3. Q No.
4. Content
5. Question Type
6. Question Pattern
7. Sub Section
8. G Strategy

Retain both:

• Broad DILR Sub Section
• Specific G Strategy

Broad historical split

These historical labels are not identical to the four promoted DILR types.

12. FOUR MAIN CETKING DILR TYPES

Exactly four types are currently promoted.

Type 1 — Arrangement

Includes:

• Wide Wordy
• Wide Numbers
• Teeny Wordy
• Teeny Numbers

Type 2 — Caselet

Includes:

• Classification Caselets
• Calculation Caselets

Type 3 — Graphs

Includes:

• Traditional Graphs
• Tables & Fill in the blanks
• Advanced Graphs
• Venn

The original categories Venn and Venn Diagrams have been combined into the canonical Venn.

Venn belongs under Graphs, not Unassigned.

Type 4 — Schedule & Routing

Includes:

• Routes & Maps
• Scheduling

Use exactly these four Type names in the student-facing and management-facing taxonomy.

13. DILR G STRATEGY HISTORICAL DISTRIBUTION

Type-level distribution

All 558 source records are assigned to one of these four Types in the summary mapping.

Year-wise question volume

These are the counts in Cetking's uploaded DILR research dataset. They must not be mistaken for verified complete official CAT section totals for every year and slot.

14. DILR QUESTION TYPE AND QUESTION PATTERN

The uploaded source contains both:

Question Type

and

Question Pattern

These are distinct labels.

Some source patterns include:

• Fixed-Answer Deductions
• Identification / Classification
• Counting Valid Configurations
• Calculation Deductions
• Optimization
• Yes/No or Determinability Checks
• Comparison / Ranking
• Matrix Grid Filling
• Scheduling Matrix
• Group Distribution
• Venn Diagrams
• Range / Possibility
• Necessarily True / False

Many rows also contain highly specific descriptions rather than reusable taxonomy labels.

Canonicalization rule

Preserve the exact original pattern.

Separately add a standardized question-pattern classification.

Avoid turning hundreds of one-off descriptions into hundreds of official category IDs.

Set-based rule

One DILR passage/set can contain several linked questions.

The structure must support:

Set ID → Shared Context → Question IDs

All questions in the set should inherit the shared context without duplicating it in every question stem.

Each question still has its own Question ID and answer.

15. SUPABASE — QUESTION FACTORY

Cetking Learn project previously used:

suqcijtpfeaystltekfn

Always verify that the connected project is the intended one before any mutation.

Canonical Question Factory tables

Do not assume every planned Question Bank attribute has already been added as a physical column. Inspect the current schema before writing SQL.

The live question_bank includes fields for:

• id
• question_code
• version
• status
• taxonomy_id
• difficulty_level
• difficulty_dna
• secondary_skills
• question_type
• question_text
• option_a through option_d
• correct_option
• correct_answer
• answer_type
• option_mistakes
• solution
• shortcut_solution
• concept_reveal
• formula information
• solution_method
• media_assets
• source exam/year/slot/reference
• is_verbatim_pyq
• review and audit metadata
• master_quant_core_type_id

Check the actual schema before assuming option_e exists on the canonical bank.

16. QUANT CAT PYQ / 3D SUPABASE TABLES

Created structures include:

• public.g_strategy_catalog
• public.cat_pyq_import_batches
• public.cat_pyq_3d_questions
• public.cat_pyq_concept_links

The dedicated 3D table includes fields such as:

• id
• pyq_code
• source_exam
• source_year
• source_year_label
• source_slot
• source_sr
• section
• question_full
• source_g_strategy_label
• g_strategy_status
• g_strategy_family
• g_strategy_id
• source_sub_section
• primary_topic
• primary_sub_topic
• dimension_type
• source_status
• linked_question_bank_id
• raw_source
• import_batch_id
• created_at
• updated_at

Important live-state warning

The 330-row QAall source was extracted and a CSV/SQL loader prepared.

The dedicated cat_pyq_3d_questions table was verified to contain 0 rows on 8 October 2026.

Do not tell the user that the 330 full questions are imported until the row load and verification actually occur.

17. QUANT SUMMARY TABLES

Current Supabase summary structures include:

• public.cat_pyq_summary_sub_section_importance
• public.cat_pyq_summary_g_strategy_importance
• public.cat_pyq_summary_topic_importance

They represent historical analytical pivots.

The website should be able to show:

Sub Section → Year → Count → Importance

G Strategy → Year → Count → Importance

Topic → Year → Count → Importance

Historical summary data can exist before the full underlying PYQ row import.

Do not imply summaries are dynamically generated from live raw rows unless they actually are.

18. VERBAL SUPABASE STRUCTURE

Relevant tables include:

• public.cat_verbal_pyq_questions
• public.cat_verbal_pyq_import_batches
• public.cat_verbal_summary_rc_non_importance
• public.cat_verbal_summary_g_strategy_question_importance
• public.cat_verbal_summary_essay_type_importance
• public.cat_verbal_summary_reading_ease_importance
• public.cat_verbal_summary_rc_cr_importance
• public.cat_verbal_summary_except_importance
• public.cat_verbal_summary_mistake_type_importance

Promoted four-section architecture:

• public.cat_verbal_g_strategy_section_catalog
• public.cat_verbal_paper_pattern_clean
• public.cat_verbal_summary_section_year
• public.cat_verbal_summary_section_slot

Verified section labels

• Summary
• Paragraph
• Evidence
• Critical Reasoning

The internal Evidence code is currently:

PASSAGE_EVIDENCE

This is okay. The visible label is Evidence.

Important live-state warning

As of the latest verification:

public.cat_verbal_pyq_questions = 0 rows

The Verbal historical statistics and summaries were loaded separately.

The original 506-row dataset has not yet been fully imported into the dedicated raw question table.

19. DILR SUPABASE STRUCTURE

Relevant tables:

• public.cat_dilr_pyq_import_batches
• public.cat_dilr_summary_sub_section_importance
• public.cat_dilr_summary_g_strategy_importance
• public.cat_dilr_summary_sub_section_year
• public.cat_dilr_summary_sub_section_slot
• public.cat_dilr_summary_g_strategy_year
• public.cat_dilr_summary_g_strategy_slot
• public.cat_dilr_type_catalog
• public.cat_dilr_strategy_type_map

Summary views:

• public.cat_dilr_summary_strategy_year_pivot
• public.cat_dilr_summary_type_year_pivot

The strategy-year pivot conceptually returns:

DILR Type | Sub Section | G Strategy | 2017 | 2018 | ... | 2025 | Grand Total

The type-year pivot returns:

DILR Type | 2017 | 2018 | ... | 2025 | Grand Total

Verified mapping state

12 canonical G Strategies are mapped into four Types.

Venn includes the former Venn Diagrams historical label.

Venn → GRAPHS

No source strategy remains unassigned in the current 4-Type mapping.

The DILR year-wise G Strategy summaries sum to 558 question records.

20. DIFFICULTY SYSTEM / QUESTION DNA

Cetking uses a structured Question Factory system.

Level 0

Foundation, pre-formula, extremely easy.

Often avoids technical jargon.

Easy

Four DNA variations:

• E1 — Direct Formula
• E2 — Reverse Formula
• E3 — One Missing/Changed Piece
• E4 — Simple Context

Medium

Five DNA variations:

• M1–M5

These represent distinct medium-level constructions, calculation or language complexity, and combinations of core ideas.

Do not invent an exact formal name for each M1–M5 cell unless supported by the topic's approved DNA specification.

Hard

Five DNA variations:

• H1–H5

Hard items should reflect genuine examination-level reasoning and complexity.

Actual CAT PYQs are preferred reference material for harder constructions.

Important

3D is a concept-combination classification, not automatically a numeric difficulty.

A hybrid CAT question can be difficult, but 3D and Hard must remain separate attributes.

21. QUANT CORE-TYPE BANK EXAMPLE — AVERAGES

Averages One is a mature example of the Question Factory architecture.

Three teaching engines

READ THE TOTAL

1. Basic Average
2. Find Missing Value
3. Consecutive Midpoint
4. Find Number of Observations
5. Required Average

MIX TOTALS

6. Pairwise A-B-C
7. Combined Average
8. Weighted Average
9. Overlapping Average
10. Reverse Combined

REPAIR THE TOTAL

11. Addition
12. Removal
13. Replacement
14. Wrong Entry / Correction
15. Group Transfer / Uniform Shift

Teaching principle

Total = Average × Count

Change in Total = Count × Change in Average

Student script:

“Turn average into total. Read/Mix/Repair total. Divide once.”

Canonical bank size

15 Core Types × 14 Difficulty DNA exam cells = 210

Plus 15 Level-0 questions = 225 total canonical Averages questions.

Do not repeat the earlier incorrect figure of 240.

Content surfaces

• Class and handout
• Video/PPT
• Book
• Arena
• Mocks
• PYQ
• Playbooks/Cheatsheets

Level 0 and Easy can be emphasized in Arena.

Books emphasize Medium and Hard.

22. QUESTION CREATION WORKFLOW

Use this pipeline:

Source Collection → Classification → Core Type Review → Difficulty DNA → Question Creation → Independent Solving → Verification → Staging → Approval → Publishing

Source collection

Import actual CAT/CET/XAT/NMAT/SNAP/CMAT questions while retaining source metadata.

Classification

Use the approved topic and core-type taxonomy.

Core Type review

Identify redundant or excessively fragmented types.

Merge only with deliberate approval.

Creation

Generate questions in the approved DNA cells.

Verification

Independently solve each question.

Check:

• Valid stem
• Unique answer or valid TITA
• Correct options
• Correct answer key
• Correct explanation
• No contradiction
• No missing information
• Correct difficulty
• Correct topic/core-type mapping
• Appropriate G Strategy tag

Staging

Stage in review structures.

Approval

A reviewer approves the question.

Publishing

Only approved questions move into student-facing tests/mocks.

AI-generated corrections should not silently modify published questions.

23. QUESTIONS UNIVERSE VS G STRATEGY UNIVERSE

Questions Universe owns

• Permanent Question ID
• Stem
• Options
• Answer
• Normal solution
• Difficulty
• Topic/core-type links
• Source/PYQ identity

G Strategy Universe owns

• Permanent Strategy ID
• Branded method
• Teaching explanation
• Shortcut
• Applicability rules
• Examples
• Links to related questions

A question can use more than one strategy.

A strategy can solve multiple questions.

Do not interpret every broad teaching label as a mathematically verified shortcut.

And always respect:

Concepts No Shortcut → Normal solution only.

24. CONTENT DISTRIBUTION ACROSS MOCKY

Questions Universe provides approved questions.

Mocky delivers:

• Topic Tests
• Sectional Tests
• Full Mocks
• PYQ practice
• Drills
• Arena Fights
• Marathon

Mocky records:

• Student Permanent ID
• Assessment ID
• Question ID
• Attempts
• Score
• Correct/Wrong
• Accuracy
• Time
• Percentile
• Section/Topic mastery
• Benchmark
• Combat result

Mocky does not become the master owner of canonical question content.

Cetking Bank handles reward transactions independently.

25. CAT 2026 EXPECTED PAPER — WORKING BLUEPRINT

Total: 68 Questions

This is a Cetking prediction, not an official fixed 2026 blueprint.

Verbal — 24 Questions

Structural prediction:

16 RC + 8 Non-RC

Non-RC

• Summary — 2
• Jumbled — 2
• Odd Man Out — 2
• Completion — 2

RC

• 4 passages × 4 questions

Detailed working category mix:

Grouped:

• Summary: 4
• Paragraph: 6
• Evidence: 6
• Critical Reasoning: 8

Total: 24

This detailed classification is a planning template, not a hard prediction of each exact question type.

DILR — 22 Questions

Working estimate:

5 sets = 3 sets of 4 + 2 sets of 5

This is an illustrative set allocation consistent with our historical categories.

Quant — 22 Questions

All forecasts should be revisited when reliable official information becomes available.

26. WEBSITE PRESENTATION

The CAT PYQ site should support:

Year-wise

CAT → 2025 → Slot 1 → Quant/Verbal/DILR

Topic-wise

Section → Sub Section → Topic → Relevant PYQs

G Strategy-wise

Strategy → Questions using that strategy

3D CAT Questions

Hybrid questions with:

• Historical identity
• Primary topic
• Linked prerequisite concepts
• Strategy tag
• Normal solution
• Shortcut solution if present
• Similar questions

Practice mode

Opening a practice attempt should use Mocky.

Attempts and performance must not be stored back into the Question Factory content tables as if they are question attributes.

27. DATA QUALITY RULES

1. Never fabricate missing options, diagrams, stems or solutions.
2. Preserve original raw source data.
3. Never infer an original CAT question number solely from spreadsheet serial number.
4. Never mark an adapted question as verbatim PYQ.
5. Never force a 3D question into one exclusive concept.
6. Never call Concepts No Shortcut a G Strategy.
7. Never rename X Maro to X Marco.
8. Use Evidence as the Verbal visible label.
9. Keep Venn under Graphs.
10. Preserve source labels even when canonical labels are normalized.
11. Every question needs a permanent canonical ID.
12. Reviews must precede publishing.
13. Do not create duplicate canonical questions just because the content appears in several mocks or books.
14. Treat summary counts as snapshots until corresponding raw data has been loaded and reconciled.
15. Verify actual Supabase schemas and row counts before implementing changes.

28. IMMEDIATE DEVELOPMENT PRIORITIES

Priority 1 — Complete raw CAT PYQ imports

Quant:

• 330 source rows
• Populate dedicated 3D CAT PYQ records
• Link canonical question IDs after matching

Verbal:

• 506 source rows
• Populate dedicated Verbal PYQ records
• Preserve passage/set links, question strategy and mistake type

DILR:

• 558 source rows
• Build/complete raw DILR PYQ records
• Preserve set identity and linked questions

Priority 2 — Link historical indexes to Questions Universe

Match/verify:

• Question text
• CAT year
• Slot
• Source reference
• Topic
• G Strategy

Priority 3 — Complete 3D concept linking

Allow multiple prerequisite Core Type IDs per hybrid question.

Priority 4 — Normalize G Strategy mappings

Quant:

• Breakup
• X Maro
• Visual
• Vedic
• Topic-specific strategies pending classification

Verbal:

• Summary
• Paragraph
• Evidence
• Critical Reasoning

DILR:

• Arrangement
• Caselet
• Graphs
• Schedule & Routing

Priority 5 — Publish website views

• CAT year-wise PYQs
• Topic-wise PYQs
• G Strategy-wise PYQs
• 3D CAT-level questions
• Practice mode through Mocky

29. FINAL INSTRUCTIONS TO CLAUDE

You are working with Cetking's real question-production architecture, not a generic examination-question dataset.

Always distinguish:

Historical source versus canonical question.

Question taxonomy versus solving strategy.

Normal question versus 3D hybrid question.

Content database versus student performance system.

Database table existence versus actual loaded records.

Use existing permanent IDs and controlled database relationships.

Do not invent missing data, strategies, question classifications or Supabase columns.

Do not create new categories when existing Cetking labels already cover the concept.

Before proposing a migration, inspect the relevant schema and check existing records.

Before publishing, independently solve and verify content.

The long-term goal is one connected Cetking Questions Universe supporting classes, books, Arena, Mocky, CAT PYQs, G Strategies, Pulse and Rav Singh AI.

END OF CETKING CONTENT UNIVERSES MASTER MEMORY — VERSION 2.0