Abstract Reasoning: Track One Attribute at a Time
What to master
Abstract reasoning tests pattern recognition without language: figure series, figure analogies, odd-one-out, paper folding and cutting, mirror and water images, embedded figures, counting figures and 3×3 matrices. The variables are limited: number of elements, sides, lines, dots or shaded regions; position and rotation (45°/90° steps); size; shading alternation; and addition or removal of elements between frames.
Traps
- Rotation versus reflection. A rotated shape keeps its chirality; a mirrored one does not. Track one asymmetric feature, such as a notch or a dot.
- Counting only small figures. Triangle-counting puzzles usually need composite triangles; count systematically by size.
- Paper folding. n folds create 2^n layers, so one punch gives up to 2^n holes, symmetric about every fold line - and a punch on a fold line merges two holes into one.
- Mirror vs water image. Mirror reverses left–right; water reverses top–bottom. Never apply both.
- Assuming an arithmetic series. The step may be ×2, alternating (+1, +3), or rotation combined with a count change.
Rule of thumb
For any series, list the value of one attribute per frame (say, number of sides) in a row: 3, 4, 6, 9. If the numbers show no rule, the logic is positional or rotational, not numerical.
Worked example
A square sheet is folded in half twice and one hole is punched through all layers, away from every fold. Unfolded, the sheet shows 4 holes, symmetric about both fold lines. If the punch sat on a fold line, two holes would coincide and only 2 would appear.