Paper folds: undo one crease at a time
Follow ideal folds, inverse hole positions, sheet bounds and partial unfolding with complete examples. Original untimed practice, without a norm.
This public guide explains one part of the original word-and-coordinate practice. Every complete attempt shows all five topic counts and all five full guides. It is not a private attempt, ranked outcome, percentile or general spatial-ability classification.
The ideal sheet and point hole are explicit
The folding tasks use zero-thickness sheets, fixed original coordinates and ideal point-sized holes punched through every current layer. The wording states the moving side and the fold line, so no imagined crease or physical paper thickness needs to be added. Read the original bounds as well as the final punch position. A point outside the material at an intermediate stage cannot become a hole merely because a reflection formula can generate its coordinate. These are mathematical fold models rather than measurements of manual dexterity or performance with actual materials.
A fresh pair of perpendicular folds
Consider a square with both coordinates between -4 and 4. Fold positive x onto negative x at x=0, then positive y onto negative y at y=0, and punch at (-3,-2). Unfolding the last crease supplies y=-2 and y=2. Unfolding the earlier crease supplies x=-3 and x=3. The four positions are therefore (-3,-2), (-3,2), (3,-2) and (3,2), all inside the original square. The distances three and two stay attached to their own coordinate axes; they are not exchanged by these folds.
Undo parallel folds in reverse order
For another paper model, let x range from -6 to 6. Fold the right half left at x=0, then fold the portion x<-3 right across x=-3. A punch with final x=-2 has x=-2 and x=-4 as preimages of the last fold. Undoing the first fold adds x=2 and x=4. This gives four distinct horizontal locations, with the same unchanged y. Reflecting around zero twice would miss the offset crease and would not reconstruct the stated sequence.
Bounds and coinciding images limit shortcuts
An off-crease punch in a completely covered region may produce two distinct preimages for one fold. A point on the fold line can coincide with its own reflection, and clipped or differently folded sheets may not cover every imagined partner. Thus doubling after every crease is not a universal counting rule. Track the candidate positions, retain only those supported by the material and count each geometric location once. The bank avoids finite hole sizes so overlapping circular areas do not silently change what a point-location question means.
Stop at the unfolding stage requested
Fully unfolding and opening only the last crease are different questions. A still-closed earlier fold can keep multiple punched layers superimposed at one location even after another crease has opened. Count positions in the stated current sheet, not the number of original layers or the positions that would appear after extra unfolding. Draw the sequence as named stages and mark which creases remain closed. This keeps an intermediate-stage answer from being replaced by a correct calculation for a later, unrequested stage.
Use six written items as revision evidence
The folding count records correct choices on six particular original models. A missed choice does not establish difficulty with everyday paper folding, vision, attention or a general spatial trait. It may involve the fold line, moving side, bounds or requested stage, but the count alone cannot identify that cause. Read the full explanation with the original wording. Every completed attempt receives all 30 worked answers and all five full guides, even at zero correct, without a ranked band, percentile or qualification claim.
Make a new model with one clear change
Keep a fresh model small enough to draw: state its rectangle, crease, moving half and punch coordinate, then reverse the fold and check every proposed location against the bounds. Change only the punch coordinate or one fold line on the next example. Paper sketches are allowed, and the test imposes no timer. Repeating its fixed choices can reward memory, so a higher count does not prove transfer, training benefit or faster practical manipulation. The useful learning record is the explicit sequence you can check.