Reconnaissance de Motifs: Guide

Comprendre les puzzles de matrices 3x3 et comment identifier les règles qui les gouvernent.

Matrix reasoning puzzles (like Raven's Progressive Matrices) present a 3×3 grid (or 4×4 at expert difficulty) where each cell contains shapes with various attributes. One cell is missing, and you must identify the pattern rules to determine what belongs there. Rules can apply to rows, columns, diagonals, or the entire grid. Each puzzle may have multiple simultaneous rules affecting different attributes.

Grid[row][col] = f(position, rules)

• Row rule: Same shape repeats across each row
• Column rule: Size increases going down each column
• Diagonal rule: Corner cells all share the same attribute

Comment le reconnaître

• Scan each row for consistent patterns
• Scan each column for consistent patterns
• Check diagonals and corners for matching elements
• Consider shape, size, and fill as independent attributes

Erreurs fréquentes

• Focusing on only one attribute when multiple rules exist
• Missing column patterns while focusing on rows
• Not considering diagonal relationships

Comment les formes changent ou restent constantes dans les lignes et colonnes d'une matrice.

Shape rules determine which geometric form appears in each cell. The simplest rule is 'constant shape' where all cells contain the same form. Progression rules change the shape systematically—either the same order repeats in each row, each column has its own shape, or each row/column contains three different shapes. Understanding shape rules helps isolate this attribute from size and fill.

shape(cell) = f(row, col) or constant

• Constant: All cells contain circles
• Row pattern: Each row shows circle → square → triangle
• All different: Each row has 3 unique shapes (order may vary)

Comment le reconnaître

• Check if all shapes are identical (constant rule)
• Compare shapes across each row—do they follow a pattern?
• Compare shapes down each column—do they follow a pattern?
• Note: shape is independent of size and fill

Erreurs fréquentes

• Confusing shape order with size or fill patterns
• Assuming shapes must change when they might be constant
• Missing that different rows can have different shape sets

Comment la taille change systématiquement dans les lignes ou colonnes d'une matrice.

Size rules govern whether shapes are small, medium, or large. Common patterns include constant size (all same), increasing size (small→medium→large across rows or down columns), and decreasing size (large→medium→small). Size progression is often the most visually obvious pattern but can be subtle when combined with other rules.

size(cell) = small | medium | large, following position rule

• Constant: All cells have medium-sized shapes
• Row increase: Each row goes small → medium → large
• Column decrease: Each column goes large → medium → small

Comment le reconnaître

• Compare sizes within each row—are they equal or changing?
• Compare sizes within each column—are they equal or changing?
• Note direction: increasing (growth) or decreasing (shrinking)
• Size is independent of shape and fill

Erreurs fréquentes

• Confusing size direction (increasing vs decreasing)
• Missing size patterns when distracted by shape changes
• Assuming size must vary when it might be constant

Comment les remplissages des formes (vide, mi-rempli, plein) changent dans la matrice.

Fill rules determine whether shapes are empty (outline only), half-filled (partial shading), or fully filled (solid). Like other attributes, fill can be constant across the grid or progress systematically across rows or columns. Fill patterns often create visual rhythm that helps identify the correct answer.

fill(cell) = empty | half | filled, following position rule

• Row pattern: Each row shows empty → half → filled
• Column pattern: Each column shows empty → half → filled
• All different per column: Each column has one of each fill

Comment le reconnaître

• Examine fill states across each row
• Examine fill states down each column
• Look for the empty→half→filled progression
• Fill is independent of shape and size

Erreurs fréquentes

• Confusing 'half' with 'empty' or 'filled'
• Missing fill patterns while focused on shape
• Not checking both row and column directions

Règles s'appliquant aux diagonales ou cellules d'angle de la matrice.

Beyond rows and columns, matrices can have diagonal rules. The main diagonal runs from top-left to bottom-right; the anti-diagonal runs from top-right to bottom-left. Corner rules specifically match the four corner cells. These patterns are typically found in harder puzzles and require looking beyond the obvious row/column structure.

Diagonal: cells[0,0], cells[1,1], cells[2,2] share attribute. Corners: cells at (0,0), (0,2), (2,0), (2,2) share attribute.

• Diagonal shape: Top-left, center, bottom-right all have same shape
• Anti-diagonal fill: Top-right, center, bottom-left all have same fill
• Corners same: All four corner cells have matching shapes

Comment le reconnaître

• After checking rows and columns, examine the diagonals
• Compare the four corner cells for matching attributes
• Diagonal rules often combine with row/column rules
• These patterns appear in harder difficulty levels

Erreurs fréquentes

• Only checking rows and columns, missing diagonal patterns
• Confusing main diagonal with anti-diagonal
• Forgetting that diagonal rules can apply to any attribute

Stratégies pour les puzzles avec plusieurs règles de motifs simultanées.

Harder matrix puzzles combine multiple rules—for example, shapes might follow a row pattern while sizes follow a column pattern and fill follows a diagonal. Solving these requires systematically checking each attribute independently, then combining the constraints to find the unique answer that satisfies all rules simultaneously.

Answer = intersection of all rule constraints

• 2 rules: Row shapes (circle, square, triangle) + Column sizes (small, medium, large)
• 3 rules: Constant shape + Row fill + Column size
• 4 rules: Diagonal shape + Anti-diagonal fill + Row size + All-different shapes per row

Comment le reconnaître

• Analyze each attribute (shape, size, fill) separately
• Write down what each attribute should be based on its rule
• Combine constraints: the answer must satisfy ALL rules
• Eliminate options that violate any single rule

Erreurs fréquentes

• Finding one rule and stopping the search
• Not checking all three attributes independently
• Choosing an answer that matches some but not all rules

Les formes tournent d'angles fixes dans les lignes ou colonnes.

Rotation patterns involve shapes turning by consistent angles (90°, 180°, 270°) as you move across rows or down columns. The shape type remains the same, but its orientation changes predictably.

Each cell's rotation = base_rotation + (position × rotation_step)

• Triangle pointing up → right → down → left across a row
• Square rotated 45° in column 1, 90° in column 2, 135° in column 3
• All shapes rotated by the same constant angle

Comment le reconnaître

• Same shape appears in different orientations
• Rotation follows a regular pattern (90° increments)
• Check both rows and columns for rotation progressions

Erreurs fréquentes

• Confusing rotation with different shapes
• Missing subtle rotation differences
• Not considering all four rotation states

Le nombre de formes dans chaque cellule augmente ou diminue systématiquement.

Count patterns involve the quantity of shapes changing across rows or down columns. A cell might contain 1, 2, or 3 identical shapes, with the count following a predictable progression.

Cell count = base_count + (position × count_step)

• 1 circle → 2 circles → 3 circles across a row
• Single shape in row 1, pairs in row 2, triplets in row 3
• Count increases from left to right in each row

Comment le reconnaître

• Count the shapes in each cell
• Look for consistent increases (1→2→3)
• Check if the pattern applies to rows or columns

Erreurs fréquentes

• Focusing only on shape type, ignoring quantity
• Assuming all cells have one shape
• Missing the count pattern when shapes overlap

La grille présente une symétrie horizontale, verticale ou ponctuelle.

Symmetry patterns require the grid to mirror itself along an axis or around a central point. Horizontal symmetry means left and right columns are reflections. Vertical symmetry means top and bottom rows mirror each other. Point symmetry (180° rotational) means opposite corners match.

Horizontal: cell[r][0] = cell[r][2], Vertical: cell[0][c] = cell[2][c], Point: cell[r][c] = cell[2-r][2-c]

• Left column shapes identical to right column shapes
• Top row mirrored in bottom row
• Corners with matching shapes (top-left = bottom-right)

Comment le reconnaître

• Compare opposite sides of the grid
• Check if folding the grid would align shapes
• Look for matching pairs across the center

Erreurs fréquentes

• Confusing horizontal and vertical symmetry
• Not checking all symmetric pairs
• Missing point symmetry around the center

Les couleurs des formes changent systématiquement dans les lignes ou colonnes.

Color patterns involve the hue of shapes changing as you move through the grid. Unlike fill (empty/half/filled), color refers to the actual hue (primary, secondary, tertiary). Colors may progress across rows, down columns, or share the same color along diagonals.

Color follows sequence: primary → secondary → tertiary

• Purple → Blue → Green across each row
• Same color in all diagonal cells
• Colors change down columns but stay consistent in rows

Comment le reconnaître

• Look for hue changes (not just brightness/fill)
• Check if colors follow a sequence
• Compare diagonal cells for matching colors

Erreurs fréquentes

• Confusing color with fill pattern
• Not noticing subtle color differences
• Assuming color is constant when it varies

Les formes se déplacent dans leurs cellules à travers les lignes ou colonnes.

Position patterns involve shapes moving within their cell boundaries. A shape might be positioned left, center, or right within its cell, with the position changing systematically across rows or down columns.

Position sequence: left → center → right (or top → center → bottom)

• Shape in left of cell → center → right across row
• Shapes shift from top to bottom down each column
• All cells in a row have shapes at the same position

Comment le reconnaître

• Notice where shapes sit within their cells
• Look for consistent position shifts
• Check if position changes row-wise or column-wise

Erreurs fréquentes

• Ignoring position and focusing only on shape type
• Confusing size changes with position shifts
• Not noticing subtle position differences

Les styles de contour des formes changent dans les lignes ou colonnes.

Border patterns involve the outline style of shapes changing systematically. Borders can be none (invisible outline), dashed, or solid. This attribute changes across rows or down columns following a predictable pattern.

Border sequence: none → dashed → solid

• No border → dashed border → solid border across row
• Column 1 has no borders, column 2 has dashed, column 3 has solid
• All shapes in a row share the same border style

Comment le reconnaître

• Examine the outline of each shape
• Look for dashed vs solid vs invisible borders
• Check if border style follows row or column pattern

Erreurs fréquentes

• Confusing border style with fill pattern
• Not noticing subtle border differences
• Assuming all shapes have the same border

Les formes suivent une séquence ordonnée définie dans les lignes ou colonnes.

Shape sequence patterns involve shapes appearing in a specific order. Unlike random shape distribution, these follow predictable progressions like circle→square→triangle. The sequence may also appear in reverse order.

Shapes follow ordered progression: shape[n] → shape[n+1] → shape[n+2]

• Circle → Square → Triangle in each row
• Same three-shape sequence repeating down columns
• Reverse sequence: Triangle → Square → Circle across row

Comment le reconnaître

• Identify the specific shape sequence
• Check if sequence repeats in every row/column
• Look for reversed sequences in harder puzzles

Erreurs fréquentes

• Missing the ordered nature of the sequence
• Confusing sequence with random shape variation
• Not recognizing reverse sequences

A est à B ce que C est à ? — appliquez la même transformation pour prédire D.

Les puzzles d'analogie géométrique présentent une grille 2×2 : la ligne supérieure A → B (transformation à un seul attribut), la ligne inférieure C → ?. La transformation qui mappe A sur B doit être la même que celle qui mappe C sur ?. Concept de Spearman ~1923.

• taille : petit cercle → grand cercle :: petit triangle → grand triangle

Une grille 3×3 où huit cellules suivent une règle et une la viole — cliquez sur l'intruse.

Les puzzles d'intrus présentent une grille 3×3 complète où une seule règle de ligne ou de colonne gouverne huit des neuf cellules. Exactement une cellule — l'intruse — a une valeur différente. Test Cattell Culture Fair (1949).

• Règle borderRow : chaque ligne a des bordures aucune/tiretée/pleine. L'intruse a 'pleine' où 'tiretée' est attendue.

Une bande horizontale montre les trois premières images d'une progression — prédisez la quatrième.

Les puzzles de continuation de séquence visuelle montrent une bande horizontale 1×4 avec trois cellules visibles et une quatrième marquée '?'. Un seul attribut (taille, remplissage, bordure, nombre, rotation) progresse à travers un ensemble fixe de valeurs ; prédisez la quatrième.

• taille : petit → moyen → grand → ? (prédire très grand)
• remplissage : vide → quart → demi → ? (prédire plein)