---
title: "The Cube-Slicing Puzzle"
id: "2250"
type: "post"
slug: "cutting-puzzles"
published_at: "2026-11-19T05:06:00+00:00"
modified_at: "2026-11-19T23:19:26+00:00"
url: "https://omnimathpuzzles.com/cutting-puzzles/"
markdown_url: "https://omnimathpuzzles.com/cutting-puzzles.md"
excerpt: "Explore two cutting puzzles involving a pie and a butter cube. Discover how geometry and mathematical induction reveal the maximum number of pieces."
taxonomy_category:
  - "Combinatorics"
  - "Geometry"
---

> *Mathematics takes us still further from what is human into the region of absolute**necessity, to which not only the actual world, but every possible world, must conform*. — Bertrand Russell, 1902

Some mathematical problems begin with complicated equations. Others begin with something as familiar as a pie or a cube of butter. Consider the following two puzzles. Their rules are simple, but the results these problems imply are surprisingly beautiful. This may be a good reminder of how complex systems can arise from a simple set of rules![1](#1df1e176-797a-4dd3-b04e-e0ef813de63d)

### Puzzle 1: How Many Pieces Can You Get from a Pie?

Imagine a round pie on a table. You have a knife and can make straight cuts through the pie, one at a time. With one straight cut, you can slice a pie into two pieces. A second cut that crosses the first one will produce four pieces, and a third cut (see the illustration below) can create as many as seven pieces. What is the**greatest number** of pieces that you can get with six straight cuts?[2](#fba6e6cf-4237-434e-af77-2ec1ac34a8eb)

Figure 1. AI-stylized image

Try a few cases and record your results. Can you find a formula that predicts the **maximum number** of pieces for any number of cuts?

*Assume each cut is a straight line

#### Solution

Reveal solution When you start with an uncut pie, it counts as one piece. When the first cut is made, you add one more piece, resulting in a total of two pieces. The second cut adds two more pieces, bringing the total to four. The third cut adds three more pieces, leading to a total of seven, and so on. **Each cut appears to contribute a number of pieces equal to the cut number**, as shown in the table below,

 | Number of cuts | New cut | Additional Pieces | Max number of pieces |
| --- | --- | --- | --- |
| 00 | 0 | 00 | 11 |
| 11 | 1st1st | 11 | 22 |
| 22 | 2nd2nd | 22 | 2+2=42+2=4 |
| 33 | 3rd3rd | 33 | 4+3=74+3=7 |
| 44 | 4th4th | 44 | 7+4=117+4=11 |
| 55 | 5th5th | 55 | 11+5=1611+5=16 |
| 66 | 6th6th | 66 | 16+6=2216+6=22 |

 So far we’ve arrived at a striking numerical pattern; the table in effect tells us that if this pattern holds true after the 3rd cut, then we would have answered our original question. Nevertheless, as tempting as it might be, we have only been able to formulate a conjecture, supported by the evidence to be sure, but a conjecture after all!

 For it is very unlikely that a pattern of this kind is a mere coincidence or chance; we have postulated a reasonable conjecture. What do we do? Well, it is not hard to see why this is the case. Consider the case of the 3rd cut. It intersects each of the two previous lines in two points. Those two points divide the third into three line segments. Each of those three segments cuts a piece of pie into two parts, so each segment adds one new piece, and the three line segments will add three pieces.

 The same is true of the 4th line. It can be rearranged such that it intersects the other three lines in three points, and three points divide a line into four parts, thereby creating four new pieces. And the same is true of the 5th line, and of the 6th line, and so forth.**Each new line intersects the previous ones in a distinct point**.

 In a case like this, a proof by **mathematical induction**[3](#210bd4fb-982a-491d-8646-1865220ec8a3)
gets to the essence of why the pattern holds in general. It reveals the mathematical mechanism that necessitates the truth of each successive case from the previous one. Mathematical induction is a type of deductive reasoning in which you go from particular cases to an infinite number of cases.

 Clearly nn cuts (lines) will divide the pie into a maximum number of pieces if any two of these lines are in **general position** (i.e., mutually intersecting lines, no two of which are **parallel**, and no three of which are **concurrent**). Suppose we have already made kk cuts. To maximize the number of pieces, the next cut must intersect all previous cuts at distinct points inside the pie.

 The (k+1)th (k+1)^{th}line meets each of the previous kk lines; the kk points of intersection divide it into k+1k+1 segments. Consequently, the (k+1)th(k+1)^{th} line cuts exactly k+1k+1 of the regions into which the plane has already been divided, since it divides each region through which it passes into two pieces. Adding the (k+1)th(k+1)^{th} lines increases the number of pieces by k+1k+1. Thus, after adding nn lines, the number of pieces will be

 2+2+3+...+n=1+(1+2+3+...+n)=1+n(n+1)2=n2+n+22by Gauss’ formula and basic algebra2+2+3+…+n=1+(1+2+3+…+n)=1+\frac{n(n+1)}{2}= \frac{n^{2}+n+2}{2} \:\: \text{by Gauss’ formula and basic algebra}

 Hence, the desired answer for n=6n=6 is

 62+6+22=36+6+22=442=22\frac{6^{2}+6+2}{2}=\frac{36+6+2}{2}=\frac{44}{2}=22

 Again.

 **A Note on Mathematical Induction.** Although we derived the formula by summing the additional pieces created by each new cut, the argument also contains the essential idea of mathematical induction. If PkP_{k} denotes the number of pieces determined by k lines, then Pk+1=Pk+k+1 P_{k+1}=P_k+k+1. Starting from P0=1P_0=1, this recurrence leads to the formula above. Writing the argument explicitly in terms of a base case and an inductive step gives a formal proof by mathematical induction.[4](#f8365df7-55f0-4a87-b406-e5061a6c703f)

### Puzzle 2: How Many Pieces Can You Get from a Cube of Butter?

Now replace the pie with a cube of butter.

This time, each cut must be a single flat slice through the cube. As before, the slices can be arranged to maximize the number of resulting pieces, but you cannot move or stack the pieces between cuts. A single slice divides the cube into **two pieces**. Two suitably positioned slices can produce**four pieces**, and three can produce **eight**.

At this point, the pattern may seem familiar. Yet the real challenge comes with the next few slices.

A cube of butter is sliced five times by a butter knife. Into how many pieces **at most** can the cube of butter thereby be divided if each knife stroke is perfectly straight (i.e., planar) and the pieces of butter are never rearranged?[5](#8e35c64a-e35f-4fb2-a13e-93c333bdb049)

Figure 2. AI stylized Image

Can you discover a general formula for the **maximum**number of pieces obtainable with nn slices?

#### Solution

Reveal solution First of all, note that this problem is fully equivalent to asking: *Into how many parts at most can three-dimensional space be divided by*nn *planes*? Since each knife stroke is modeled by a plane and the pieces correspond to the regions formed inside the cube. Secondly, unless one possesses an extraordinary spatial imagination, one cannot visualize all at once the pieces yielded by the 55 knife strokes, and even in such a case, it will not prove that whatever number one happens to get is the true maximum. A better way is to discover a law or a rule that will give the greatest number of pieces that can be obtained with any number of cuts.

 If the 5 planes or any pair is parallel to each other, then there would be fewer parts than if they were in **general position** (For example, if two were parallel, they would not cross each other, potentially dividing an existing region). For the moment, we will define this term intuitively just as we did at first in the case of lines cutting the pie, that is, **each new plane must intersect the previous planes in a distinct line**.

 One plane divides space into 22 regions. Two planes in general position partition space into 44 regions. Three planes divide space into 88 regions. You may be tempted to conclude that four planes will form 1616 regions at most. To understand why it is not the case, let us explain how the planes intersect as we add new planes and make a table to see whether we can spot a pattern.

 Now, the crucial insight is that**a new plane is divided into regions by its intersections with the previous planes.** Each region corresponds to a surface where the new plane intersects each of the regions created by the previous planes. Consequently, to maximize the number of regions, we must, therefore, maximize the number of regions into which the new plane is split, but we know that the lines on this plane will divide the most if they are in general position!

 Fig. 3

Fig. 4

 Then we ask, what does this imply about the way in which the planes must intersect? Well, imagine you’re looking at the new plane head-on. The intersecting previous planes will look like lines; in fact, the lines of intersection. If no two lines are parallel, it means that each line must intersect another in a distinct point. For this to occur, **every three planes must have exactly one point in common**, for otherwise the lines on the surface of the new planes will not be in general position; each line of intersection comes from a previous plane. Thus, when we look at two lines intersecting on the surface, there are three planes involved (the two previous ones and the new plane).

 On the other hand, what does it mean for the planes that no three lines are concurrent? It means that if three lines were to pass through the same point on the surface of the new plane, then at least four planes would have to intersect. Hence, **no four planes can have a point in common**. And that’s it! We’ve deduced from how the planes must meet each other, and we can come back to the question of how many parts at most four planes can produce.

 We add the 4th plane; the three previous planes will intersect the new plane along three lines. If these lines are in general position (we’ll see shortly that they must be by the conditions of the problem), then the plane is divided into seven regions (the result of the first puzzle!). The fourth plane, therefore, adds seven new three-dimensional regions. But the first three planes have already created eight regions, so the new total is 8+7=158+7=15 (See Fig. 3)

 By similar reasoning, we can fill in our region-counting pattern table:

 | Number of planes | New regions added | Total regions |
| --- | --- | --- |
| 0 | 0 | 1 |
| 1 | 1 | 2 |
| 2 | 2 | 4 |
| 3 | 4 | 8 |
| 4 | 7 | 15 |
| 5 | 11 | 26 |

 Note that the above table implies that the relation between the number of planes added and the new regions formed is a function of the lines created by intersection, as explained before (See Fig. 4). Our answer seems to be 26! However, it’s time to turn our earlier insight into a general proof while defining explicitly what it means for a plane to be in general position.

 nn  planes will divide 3-dimensional space into a maximum number of regions if all of them are in **general position**(if no two of them are parallel, any three of them have exactly one point in common, and no four of them share a point).

 Suppose that kk planes have already been arranged in general position; we may explore how adding a (k+1)st(k+1)^{st}plane increases the number of regions into which space is divided. This plane intersects each of the first kk planes in a line; moreover, any two of these lines of intersection have exactly one point in common, since any three planes have exactly one point in common, and no three of these lines are concurrent (if three of them passed through the same point, then at least four planes would pass through the same point, which contradicts our supposition).

 Consequently, these kk lines divide the (k+1)st(k+1)^{st} into k2+k+22\frac{k^{2}+k+2}{2} regions (the result of the first problem), each of which is a surface on which the (k+1)st(k+1)^{st} meets one of the pieces formed by the first kk planes. Thus the (k+1)st(k+1)^{st} planes cuts k2+k+22\frac{k^{2}+k+2}{2} pieces, splitting each in two; accordingly, adding the (k+1)st(k+1)^{st} plane increases the number of pieces by k2+k+22\frac{k^{2}+k+2}{2}. Since a single plane divides space into two regions, it follows from this that nn planes will split space into

 2+12+1+22+22+2+22+32+3+22+...+(n−1)2+(n−1)+222+ \frac{1^{2}+1+2}{2}+\frac{2^{2}+2+2}{2}+\frac{3^{2}+3+2}{2}+…+\frac{(n-1)^{2}+(n-1)+2}{2}

 =2+(12+22+32+...+(n−1)2)+(1+2+3+...(n−1))+2+...+2⏞n−12=2+\frac{\left( 1^{2}+2^{2}+3^{2}+…+(n-1)^{2} \right)+\left( 1+2+3+…(n-1) \right)+\overset{n-1}{\overbrace{2+…+2}}}{2}

 regions. Since

 1+2+3+...+(n−1)=n(n−1)21+2+3+…+(n-1)=\frac{n(n-1)}{2}

 and[6](#c23d47f3-ae45-4e4a-9852-da4ca1412aa4)

 12+22+32+...+(n−1)2=n(n−1)(2n−1)61^{2}+2^{2}+3^{2}+…+(n-1)^{2}=\frac{n(n-1)(2n-1)}{6}

 the total number of regions is

 2+n(n−1)(2n−1)12+n(n−1)4+(n−1)2+\frac{n(n-1)(2n-1)}{12}+\frac{n(n-1)}{4}+(n-1)

 =2+(n−1)(2n2−n+3n+12)12=2+\frac{(n-1)(2n^{2}-n+3n+12)}{12}

 =n3+5n+66=\frac{n^{3}+5n+6}{6}

 Hence, the desired answer for n=5n=5 is

 53+5×5+66=125+25+66=1566=26\frac{5^{3}+5\times5+6}{6}=\frac{125+25+6}{6}=\frac{156}{6}=26

 Again.

#### A Further Challenge: What If We Could Rearrange the Pieces?

In the original puzzle, every cut passes through the cube in its existing position, and the resulting pieces remain in their initial positions. But suppose we change this rule: after each cut, we are allowed to move (translate) and rotate the resulting pieces freely before making the next one.

**What is the greatest number of pieces that can be obtained after**nn**cuts?**

Can you establish an upper bound for the number of pieces and demonstrate a geometric arrangement that actually attains it?

#### Solution

Reveal solution nn planes will divide the cube into a maximum number of parts if **each plane can cut through any and every one piece**. Note that every piece obtained by successively slicing the cube with planes is**convex**[7](#e327efac-a3c4-4494-a67a-e5584c49c902)
, since each piece is the intersection of the original cube with a collection of **half-spaces**[8](#8cf91efe-cf7e-465c-88e9-149115dee49d)
. Therefore, moving or rotating these pieces will not affect their convexity.

 Consequently, a single plane can divide any one piece into at most two smaller pieces. Hence, if we have PkP_{k} pieces after kk cuts, the next cut may produce at most 2Pk2P_{k} pieces. Since initially we have one piece, namely the whole cube, it follows that after the nthn^{th} cut

 Pn≤2nP_{n} \leq 2^{n}

 It remains to show that this upper bound can actually be attained.

 Suppose that after kk cuts we have 2k2^{k} pieces. we may explore how the (k+1)st(k+1)^{st} cut increases the number of regions into which space is divided. Since all of the pieces are bounded solids, we may rearrange the pieces such that they are **pairwise disjoint**[9](#ef117f41-b577-4cb5-b1c6-a406e46f5659)
 and **a single plane**, representing the (k+1)st(k+1)^{st} cut, passes through the interior of each piece.

 To accomplish this, we first choose an interior point of each piece and translate the piece in a direction perpendicular to a fixed plane until that point lies on the plane. We can then separate the pieces by translating them parallel to the plane, sufficiently far apart to prevent any overlap.

 These latter translations preserve the fact that each piece has an interior point on the plane. Accordingly, this plane passes through the interior of every piece, and since the pieces are convex, it divides each into exactly two smaller parts. Hence, the (k+1)st(k+1)^{st} cut produces

 Pk+1=2Pk=2⋅2k=2k+1P_{k+1}=2P_k=2\cdot 2^k=2^{k+1}

 pieces

 Starting with P0=1P_{0}=1, we can repeat this procedure after every cut. Thus, the upper bound is attainable, and we conclude that

 Pn=2n.\boxed{P_{n}=2^n}.

### Notes

Technical Notes on the Problems & their Solutions 1. The proofs of the general cases for Puzzle 1 & 2 are adapted from those given in A. M. Yaglom and I. M. Yaglom’s *Challenging Mathematical Problems with Elementary Solutions*1[↩︎](#1df1e176-797a-4dd3-b04e-e0ef813de63d-link)
2. Adapted from Gardner, M. (1961, October). *Entertaining Mathematical Puzzles*. Dover Publications. p. 26 (PDF version). [↩︎](#fba6e6cf-4237-434e-af77-2ec1ac34a8eb-link)
3. Principle of Mathematical Induction: Let P(n)P(n) be a property that is defined for integers n, and let a be a fixed integer. Suppose the following two statements are true: “P(a)P(a) is true and for all integers” and “fork≥a k \ge a, if P(k)P(k) is true then P(k+1)P(k + 1) is true”. Then the statement for all integers n≥an ≥ a, P(n)P(n) is also true2[↩︎](#210bd4fb-982a-491d-8646-1865220ec8a3-link)
4. This problem is completely equivalent to finding the maximum number of regions that can be created in the plane with nn lines in general position (i.e., mutually intersecting lines, no two of which are parallel, and no three of which are concurrent). If we divide the plane into regions using nn lines and draw a circle that encloses all the intersections of those lines, these regions will meet the circle in a closed area; a portion of the unbounded regions will become bounded by the circle, and the bounded ones will fall within the circle, much like the pieces of cake created by the cuts. [↩︎](#f8365df7-55f0-4a87-b406-e5061a6c703f-link)
5. Adapted from Hoeflin, R. K. (1985, April). *Problem 35 in The Mega Test*. Omni, 7(4), p. 129. [↩︎](#8e35c64a-e35f-4fb2-a13e-93c333bdb049-link)
6. This formula may be proved using Mathematical Induction, a cancellation argument similar to that used to prove Gauss’ trick, or a geometrical argument. It is left as an exercise to prove this result. [↩︎](#c23d47f3-ae45-4e4a-9852-da4ca1412aa4-link)
7. A set in the plane or in three-dimensional space is called *convex* if the line segment joining any two of its points is contained in the set.1[↩︎](#e327efac-a3c4-4494-a67a-e5584c49c902-link)
8. In geometry, a *half-space* is either of the two parts into which a plane divides the three-dimensional Euclidean space.3[↩︎](#8cf91efe-cf7e-465c-88e9-149115dee49d-link)
9. A collection of sets is *pairwise disjoint* if, and only if, no two distinct sets in the collection have any element in common.2[↩︎](#ef117f41-b577-4cb5-b1c6-a406e46f5659-link)

### References

- 1 A. M. Yaglom and I. M. Yaglom, *Challenging Mathematical Problems with Elementary Solutions*, vol. 1, *Combinatorial Analysis and Probability Theory*, translated by James McCawley, Jr., revised and edited by Basil Gordon (New York: Dover Publications, 1987), 12-13, 102-103, 104-105.
- 2Epp, S.S. (2011). *Discrete Mathematics with Applications*. 4th Edition, Brooks/Cole Cengage Learning, Boston, 246, 345.
- Gardner, M. (1961, October). *Entertaining Mathematical Puzzles*. Dover Publications. p. 26 (PDF version).
- 3Half-space (geometry). (2026). In *Wikipedia*. https://en.wikipedia.org/w/index.php?title=Half-space_%28geometry%29&oldid=1351549779
- Hoeflin, R. K. (1985, April). *Problem 35 in The Mega Test*. Omni Magazine, 7(4), p. 129.

## Related Posts

[https://omnimathpuzzles.com/thespacebetweendots/](https://omnimathpuzzles.com/thespacebetweendots/)

[https://omnimathpuzzles.com/the-building-puzzle/](https://omnimathpuzzles.com/the-building-puzzle/)

### Leave a Comment [Cancel Reply](/cutting-puzzles/#respond)
