Cube and Dice

What information does a cube problem provide?

A cube has six faces, twelve edges and eight vertices. At each vertex, three faces meet. Each face has four adjacent faces and one opposite face. These relationships are the starting point for cube-and-dice reasoning, but a question may ask about several different tasks: folding a net, comparing views of a labelled die, or counting small cubes after a larger painted cube is cut.

Identify the task before applying a shortcut. The examples here are original exercises supporting the broad spatial and mental-ability skills relevant to Preliminary Paper II, not claims about past-question frequency.

Adjacent and opposite faces are different constraints

Two faces sharing an edge are adjacent. Opposite faces lie across the cube and do not share an edge or a vertex. In the usual three-face corner view, the visible faces are pairwise adjacent. If a view shows labels A, B and C around one corner, none of those pairs can be opposite. However, that single view does not identify which of the remaining three labels is opposite A. Several valid labellings remain possible. Distinguish an excluded relationship from a uniquely established one.

Suppose additional views show A adjacent to B, C, D and E. A has exactly four adjacent faces, so the sixth label F must be opposite A. This is an elimination argument based on complete adjacency evidence.

Seeing A with only three different neighbours would leave two possible candidates for its opposite face. Do not promote a plausible candidate into a necessary answer because a familiar die arrangement comes to mind.

Ordinary numerical dice often use an opposite-sum convention, but an abstract labelled cube need not follow it unless the question supplies that condition.

A change of viewpoint is a rigid movement

When a cube turns, its face labels move together as one rigid object. Opposite pairs and adjacency relationships remain fixed. A picture that preserves those pairs can still have the wrong orientation if it is a mirror reversal rather than a rotation.

For a question about a particular top, front and right arrangement, track all three positions instead of checking only which labels are present. Holding one face fixed leaves a cycle of four side faces, and rotating around that face moves the side labels together through that cycle.

For an original example, let the top be A, bottom F, front B, back E, right C and left D. A roll to the right moves the original right face C to the bottom and the original left face D to the top. A and F move to the right and left positions, respectively, while front B and back E remain front and back. The new top is D. Choosing C confuses the face towards which the cube rolls with the face that rises to the top.

Nets supply a different kind of evidence

A net lays six squares flat before they are folded. Squares touching in the net remain adjacent after folding, but some squares separated in the flat layout become adjacent on the cube. The drawing’s leftmost and rightmost positions do not directly identify opposite faces.

Choose a reference square and imagine each neighbouring square turning through a right angle around its shared edge. Track where the remaining faces point. A valid net must place the six faces on six distinct sides without overlap; merely counting six squares is insufficient.

Consider four squares A, B, C and D in a horizontal row, with E above B and F below B. Fold B as the front face. Then C becomes right, D becomes back, A becomes left, E becomes top and F becomes bottom. The opposite pairs are B–D, A–C and E–F. This explicit mapping is safer than guessing from distance on the paper. It also lets the learner test a proposed corner view: a view containing A and C together cannot be a three-face corner because those faces are opposite.

Painted-cube counting uses position categories

Suppose a large cube is painted on every outer face and divided into n equal parts along each edge. There are n cubed small cubes. For n at least two, the eight corner cubes have three painted faces. Cubes on an edge but away from its corners have two painted faces; there are twelve edges and n − 2 such positions on each. Cubes inside a face but away from its boundary have one painted face, giving six times (n − 2) squared. Completely internal cubes have no paint, giving (n − 2) cubed.

For n = 4, the categories contain 8, 24, 24 and 8 cubes, respectively. Their sum is 64, matching the total. The check is important because counting an entire edge for the two-painted category would include corners already counted elsewhere.

Similarly, counting an entire face for the one-painted category would include its edges and corners. Each small cube must enter exactly one category. The formulas also depend on the painting condition; painting only selected outer faces changes the classification and requires a fresh count.

State assumptions before assigning probabilities

If one small cube is selected uniformly from the 64 pieces in the painted example, the probability that it has exactly two painted faces is 24/64, or 3/8. That conclusion uses a stated random-selection model.

It does not follow from merely saying that someone picks a cube, because an unconstrained selection process could favour visible corner or surface pieces. A geometry count and a probability claim are related but distinct: first establish the favourable and total counts, then verify that equal likelihood is part of the exercise.

A compact solution method

Classify the question as a view, net, roll or painted-cube problem. Record the supplied labels and assumptions, and preserve the distinction between what is possible and what is forced. For a view, build adjacency and opposite-pair constraints.

For a roll, update the complete orientation. For a net, fold from a reference face. For counting, use disjoint position categories and check their sum against n cubed. These steps make the reasoning auditable and expose the specific mistake represented by a close answer option.