I thought that all shapes possess translational symmetry, as simply moving it shouldn't change the shape. Then I thought of a situation where it does, though it contains three criteria.
- The observer is fixed.
- The shape does not possess reflective symmetry.
- The shape is moved on a curved surface.
An easy-to-visualize example of this would be a moon shape being moved on the surface of a sphere. If it was moved halfway around the sphere, it would be flipped from the perspective of the observer (this is assuming the observer can see through the sphere).
However, is this a valid situation? If not, what examples are there of shapes that do not have translational symmetry? I cannot see how there can ever be an example of a shape without translational symmetry on a flat plane.
EDIT: This question is based on a misunderstanding of translational symmetry, my understanding coming from this site: https://www.cuemath.com/geometry/symmetry/
Usually, "translation" refers to Euclidean spaces like $\Bbb R^2$ (plane) or $\Bbb R^3$, the 3-space.
"Translational symmetry" then means you can move an object, and it still looks the same. Moving a moon across a plane or across a sphere$^1$, the moon looks different: it's in a different place and you can tell apart "before" and "after" the move (except the move was trivial, i.e. a no-move).
An example of an "object" that has translational symmetry is the set of intergers in $\Bbb R$: Moving all integers one unit to the right, you still get all integers and the set remains unchanged. The natural numbers, however, are not invariant under translation.
Notes
$^1$Defining translation in other spaces is a bit tricky, because it should be possible to move objects around to start with. For example, you cannot move a piece of flat fabric that's flexible but does not stretch, from a flat part of a space to a place where the space is bumpy: The fabric will wrinkle and the concept of "moving objects around" will fall apart.
You can still extend the concept to other spaces, though, provided the space has constant (non-changing) curvature everywhere. Examples are
Examples of translational invariant objects in hyperbolic space are hyperbolic tilings.
Examples of translational invariant objects on a 2-sphere are implied by the Platonic solids ("blow" them up so the faces lie on a sphere) and also by tilings that have $m$-fold rotational symmetry like a latitudes + longitudes coordinate grid.