Spatial & Visual Reasoning — Fold it, turn it, see it from the other side

Cut a cereal box along its edges and flatten it. What you get is a net — a flat shape that folds back into the solid. A cube has 6 square faces, so a cube net is always 6 squares joined edge to edge. But here is the catch that traps everyone: not every arrangement of 6 squares folds into a cube.

Fold one in your head and you feel why. As the squares swing up, each one has to land on its own face — floor, ceiling, or one of the four walls. If two squares both try to cover the same wall, the fold jams. Six squares is necessary. It is not enough.

The faster question: which faces end up opposite?

Most net problems do not ask ‘is this a cube?’ They ask ‘after folding, which face sits across from which?’ You could fold every time. There is a rule that saves you.

Look at any straight row of squares in the net. Walk along the row. Faces pair up by skipping one: the 1st with the 3rd, the 2nd with the 4th. Two squares that touch can never be opposite — on a real cube, neighbours share an edge, and opposite faces share nothing.

When four faces sit in a row, they wrap around the cube like a belt, and that belt hands you two opposite pairs at once (1–3 and 2–4). The two squares poking out of the belt are the last pair, top and bottom. All six faces sorted, no folding needed.

THE MOVE: opposite faces are the two with exactly one square between them in a straight line. Skip one to find the partner.

A drawing of a 3D pile is a liar. It shows you the front and the top and hides everything behind and underneath. Try to count the whole pile at once and you will always miss the cubes in the back. So do not count the pile. Count slices.

Cut the shape into flat layers — floor layer, next layer up, next — and count one easy 2D slice at a time. Write each layer’s number down. A solid n×n×n cube is n identical layers of n×n cubes, which is why it holds n³ small cubes: a 3×3×3 is three layers of 9, so 27.

See why the back cubes hide? On the page you only meet the front edge of each floor — the other eight cubes of every floor sit behind it. Counting floor-by-floor catches all nine each time; eyeballing the drawing never does.

Front, top & side views

Kangaroo loves showing a stack and asking what it looks like from the front, the top, or the side — or running it backwards: here are the three flat views, how many cubes are there?

One idea unlocks all of them. A view records the tallest stack along each line you look down. From the front you see the highest cube in each column. From the top you see the footprint — which floor squares are covered at all. From the side, the highest cube in each row.

To rebuild a cube count from views, lay the footprint out and raise each column to the tallest height the front and side views allow. If the problem asks for the maximum, push every column as high as it is permitted; for the minimum, raise only as much as the views force.

THE MOVE: slice into layers and count flat. Each view is the tallest cube along your line of sight.

Look at a brick wall, a bathroom floor, a honeycomb. Copies of one shape, fitted edge to edge, with no gaps and no overlaps. That is a tiling, and Kangaroo turns it into two questions: can this shape cover the region, and what is the fewest (or most) pieces that do it.

Reach for area first — it is free. If the region has area A and each tile covers area t, then you need at least A ÷ t tiles, and never more than that with a single tile size. That one division kills half the answer choices before you place a single piece.

Area gives the ceiling. Then you must check the shape actually reaches it, because thin arms, odd corners, or a parity clash can stop you one short. A quick checkerboard colouring is the classic tie-breaker: if every tile must cover an equal number of black and white squares but the region has unequal black and white, a perfect cover is impossible.

THE MOVE: divide areas for the count, then confirm a real arrangement (or a colouring argument) reaches it.

Hold a word up to a bathroom mirror. AMBULANCE on the front of a van comes out backwards — which is why real ambulances print it reversed, so it reads right in your rear mirror. That is a reflection: left and right swap across a line. A rotation is different — the whole figure turns around a point, like a clock hand.

Both go wrong the same way: you try to flip or spin the entire shape in your head and lose your place. Do not. Track one feature. Pick a single corner, one arrow, one beam, and ask only where it lands. Across a mirror, a point hops to the other side, the same distance from the line. The rest of the figure obeys.

Two facts handle most of these. The back view of something is its front view flipped left-to-right — so reading a wall from behind reverses it. And a beam that hits a mirror bounces off at an equal angle: trace it one straight segment at a time.

Reflect to straighten a bent path

Here is reflection used as a tool, not a question. A frog at A must hop down to touch a straight bank, then up to B. What route is shortest? The bent two-leg path is hard to measure — so get rid of the bend. Mirror B across the bank to a copy B′ on the far side. Every down-then-up route to B is exactly as long as the matching straight-across route to B′ (the second leg is just flipped over the line). The shortest of those is the dead-straight line A→B′, and where it crosses the bank is the touch point.

The same straighten-the-bend idea reads a bouncing beam or a fold: replace the reflected leg with its mirror image and the whole path snaps into one straight line you can measure.

Strategy framing inspired by Posamentier, The Art of Problem Solving (teacher resource).

Why a turn is not a flip

This is the mix-up that costs the most points. A rotation keeps the shape itself unchanged — same handedness, only pointed a new way, the way your right hand is still a right hand after you wave it around. A reflection flips handedness — your right hand’s mirror image is a left hand, and no amount of turning will ever make them match. Watch one letter prove it:

Same letter, two fates: turn the b and you get q; mirror it and you get d. They land in different places because a turn and a flip are genuinely different moves. When a problem says turn / rotate, spin it; when it says flip / mirror / back view, reverse it.

For symmetry questions (does this picture have a mirror line?), test the pieces, not the whole. A stacked word is vertically symmetric only if every single letter is. One bad letter kills it.

THE MOVE: follow one point, arrow, or beam through the flip or turn — never the whole figure.

Pick up a real die and add the dots on any two opposite faces. You get 7 every time: 1 across from 6, 2 across from 5, 3 across from 4. Three pairs, all summing to 7. Memorise them and a whole family of problems cracks open.

Two facts ride along with it. Faces that share an edge are neighbours, never opposite — so to find a face’s partner, list everything that touches it and the one number left over is the opposite. And all six faces add to 1+2+3+4+5+6 = 21, which is your budget for any ‘total of the visible faces’ problem: start from 21 per die and subtract whatever is hidden or glued away.

Beyond dice: numbered cubes

Some problems put your own numbers on a cube and ask about corner sums (add the three faces meeting at a corner). The quiet gift here: two corners at opposite ends of a space diagonal together touch all six faces once, so their corner sums always add to the same total. That lets you pair a known corner with an unknown one and solve in one line.

THE MOVE: opposite faces sum to 7, all six sum to 21, neighbours are never opposite. For corner sums, pair opposite corners.

Hand a kid a few odd pieces and ask which ones snap together into a clean shape, and the instinct is to shove pieces around at random until something looks right. Resist it. There is a calmer way in, and it starts with counting before fitting.

Whether you are joining pieces or slicing a solid, ask first: what number must stay fixed, and what changes? If three blocks combine into a cube building, the building must have exactly as many little cubes as the three blocks together — count cubes and cross out every choice with the wrong total before you fit anything. If you cut or drill a solid, count how each cut adds or removes faces, edges, and corners.

When you do fit pieces, match a bump to a notch: a piece’s stick-out has to drop into another piece’s matching dent, no gap and no overlap. Find the bump and the dent it fills, slot them, and trace the outside edge — if it comes out clean, that is your pair.

For face counts, do not forget the surfaces a hole or tunnel creates inside. A block with a tube drilled through it gains the inner walls of that tube as brand-new faces.

THE MOVE: count the fixed quantity first (cubes, area, corners), then fit by matching bump to notch.

One more ‘count what changes’ trap, away from solids: a straight log gets cuts, and each cut adds one more piece — so pieces are always one ahead of cuts. Sketch it and you see why the count is off by one.

Framing inspired by AoPS Prealgebra.

Grid problems look like easy counting, and that is the trap. Every single time, the danger is the same: you miss a cell or count one twice because you trusted your eye. The cure is to count in an order you cannot lose track of.

If a grid is filled with numbers in a steady march — 1 to 8 across the first row, 9 to 16 across the next, and so on — you never have to read every cell. Pin down one cell you are sure of, then step to the others by the grid’s rhythm: the cell to the right is ‘plus 1’, and the cell directly below is ‘plus one whole row’.

With eight per row, the cell under any number is exactly 8 more. So once you place a single anchor, you know precisely which numbers a puzzle piece would sit on — and you match the piece to those positions, not to a fuzzy picture.

For ‘how many little squares / rectangles’ problems, count by size class: all the 1×1s, then all the 2×2s, and so on, so nothing slips past. For path problems, count by which cell you step through.

THE MOVE: anchor one cell you are sure of, then reach every other by ‘plus a row, plus a column’.

The hardest spatial problems are not about seeing — they are about counting without missing or doubling. And the single most powerful counting move sounds backwards: count too much on purpose, then take back the extra.

Picture marking every little cube that touches a drawn line, going face by face. A cube tucked on an edge or corner of the big cube belongs to two faces, so you count it twice. That is fine — as long as you then subtract those doubles back out. Counting loose and correcting beats trying to count each thing exactly once while juggling overlaps in your head.

This is the same idea as the overlapping circles above: add the two counts, then remove the shared middle once so it is not counted twice. On cubes it is the cubes on shared edges; on tunnels it is the cubes where two tunnels cross; on grids it is a cell two shapes both claim.

The other half of clean counting is casework: split the situation into cases that do not overlap and do not leave a gap, and count each. Keep a written tally so a case never silently disappears.

Counting figures hidden inside a figure: sort by size, then by direction

‘How many triangles in this picture?’ is the classic miss-or-double trap. Pointing at random never works. The fix is a two-key sort: group every triangle first by its size, then within each size by which way it points (up or down). A cut-out cardboard triangle you slide around the figure makes sure none escapes.

Take a triangle split into a 3-row grid of nine small triangles. Sort and tally:

Add the column entries: 6 + 3 + 3 + 1 = 13 triangles. The down-pointing ones are exactly the ones beginners forget — sorting by direction forces you to look for them. When a figure is symmetric (a six-point star, say), up-count equals down-count, so you tally one and double, halving the work.

Strategy framing inspired by Posamentier, The Art of Problem Solving (teacher resource).

THE MOVE: count everything loosely, then subtract exactly the things you counted more than once.