*a single puzzle piece*

that can be rotated into 4 positions

that can be rotated into 4 positions

In these puzzles, Truchet tiles like the one above are used to create a specific pattern. All the pieces are the same - it is just a question of rotating them correctly to make the pattern you are aiming for.

*a Truchet puzzle: can you make this pattern*

using only Truchet tiles?

using only Truchet tiles?

We can make these Truchet puzzles a bit more interesting if we have a specific starting arrangement, and add the restriction of using the smallest number of moves (clockwise rotations of individual tiles) to get from the starting arrangement to the target arrangement.

*how many clockwise rotations of tiles will it take to transform*

the square on the left to the one on the right?

the square on the left to the one on the right?

On this page, you can try out some puzzles like this. Because you are only allowed to rotate pieces clockwise, you may have to rotate a given tile 0, 1, 2 or 3 times in order to get it into the desired position.

Please give it a try: https://dmackinnon1.github.io/truchet/match.html

When setting up this puzzle page, I used a restricted set of arrangements for both the starting and target arrangements. For the starting arrangement, I have all the Truchet tiles in the same position. Let's call this type of arrangement a

**uniform**Truchet square. There will be 4 different uniform Truchet squares of a given size; here is one for

*n*= 6:

*a*

are in the same position

**uniform Truchet square**: all tilesare in the same position

Because I like the way they look, I chose the target arrangements to always have

**4-fold rotational symmetry**. An important choice as it turns out. These Truchet squares have sides of even length, and can be divided up into 4 quadrants - as we move around the quadrants in a clockwise direction, the pattern in each quadrant is a 90 degree rotation of the pattern in the preceding quadrant.

*you can make Truchet squares with 4-fold rotational*

symmetry - they tend to look nice.

symmetry - they tend to look nice.

With these restrictions, I found that:

- All 2x2 puzzles are solvable in 6 moves.
- All 4x4 puzzles are solvable in 24 moves.
- All 6x6 puzzles are solvable in 54 moves.

Mysteriously, no matter what pattern was chosen for the target, all puzzles of the same size required the same number of moves to solve.

*symmetric puzzles of the same size always*

have the same number of required moves

have the same number of required moves

In general, it turns out that for any Truchet square

**T**with side length

*n*and 4-fold rotational symmetry,

**T**will always be 6*(n/2)^2 rotations away from any uniform Truchet square.

This was more puzzling than the original puzzles: how could my randomly generated puzzles all require the same number of moves to solve?

To understand why this is the case, we can find a way to count all the rotations required to transform a uniform Truchet square into one that has four-fold rotational symmetry, and see that this does not depend on a particular choice of either the starting arrangement or the target. This turns out to be easier than you might expect.

LetUbe a uniform Truchet square of side lengthn, andTbe a Truchet square with 4-fold rotational symmetry, also of side lengthn. We'll count how many rotations it takes to transformUintoT.

Lett1be a tile in the first quadrant ofU. If we consider its image under rotation into the other quadrants, we have a set of 4 tiles,t1, t2, t3, andt4. One of these will be aligned with the corresponding tile inT(since the tiles inTthat lie in the same positions ast1,t2,t3, andt4are rotated through all 4 positions, while the tiles ofUare all in the same position). It will require 0 moves for this tile be put into the same position as the corresponding tile ofT. As we move around the quadrants to the other images of our selected tile, they will all be in the same position (Uis uniform), but the corresponding tiles ofTwill be rotated. The corresponding tiles ofTwill be 1, 2, and 3 rotations ahead from the tiles ofU. So each tile in the first quadrant will, along with its images under rotation, contribute 0+1+2+3 = 6 rotations to the overall number of rotations required to transformUintoT. There are (n/2)^2 tiles in the first quadrant, hence the total number of rotations required to transformUintoTwill be 6*(n/2)^2.

*for a given n, it always takes the same number of moves*

to transform a uniform Truchet square

to transform a uniform Truchet square

**U**into another*one,*

**T**, if**T**has 4-fold rotational symmetryThe reason for all the puzzles requiring the same number of moves is the 4-fold symmetry of the target (and the uniformity of the starting arrangement). If we allowed non-symmetrical target arrangements, or varied the starting arrangements used, the number of rotations required to transform our starting arrangement into the target arrangement could lie anywhere between 0 and 3n^2.

*a 24 x 24 Truchet square with 4-fold rotational*

symmetry - starting from a uniform square, how

many moves would be required to re-create it?

symmetry - starting from a uniform square, how

many moves would be required to re-create it?

Why do the Truchet squares with rotational symmetry seem particularly nice? The human brain loves symmetry, but in the case of Truchet squares, it seems particularly appropriate to be drawn to arrangements like these. Individual Truchet tiles lack rotational symmetry - this lack of symmetry is what gives them their expressive power. By arranging them into a square pattern that does have rotational symmetry, we are overcoming the asymmetry of the original tile, appealing maybe to our need to unify opposites, or to express a sense of dialectical tension.

*Some earlier posts on Truchet tiles:*

*truchet en plus*

*truchet tiles*