How this test works
Each round shows two flat shapes made of squares. The one on the right is always turned by some angle relative to the one on the left. Your job is to decide whether it is the same shape, just turned, or a mirror image of it, turned. Answer with the buttons, or with the keys ← or S for same and → or M for mirrored.
There are 20 pairs. Exactly half of them are mirror images, shuffled into a random order. The right-hand shape is turned by 45° to 315° in 45° steps, never 0°, and the left-hand shape is also given a random quarter turn so it isn't always upright. The same shape never comes up twice in a row. After each answer you see whether you were right, then the next pair appears.
Why every pair has exactly one right answer
A symmetrical shape, like a T or a plus sign, looks the same in a mirror once you turn it. For a shape like that, "same" and "mirrored" would both be correct. So every one of our 8 shapes (5 to 7 squares each) is checked automatically to make sure its mirror image can't be reached by turning it. The check runs every time the site is built.
How we calculate your result
Your result is the number of pairs you got right out of 20, together with your time. The time for each pair runs from the moment the pair is drawn to your answer; the short pause showing whether you were right is not counted. We show the total time, the average time per pair and your accuracy as a percentage.
Because half the pairs are mirrored, answering at random scores about 10 out of 20 on average. That is the baseline to beat.
Ways to approach it
- Turn the right-hand shape back in your head until it lines up with the left one, then see whether they match.
- Or pick one distinctive feature, such as a short arm or a step, and check which side of the shape it's on in both figures.
- Accuracy first. A quick wrong answer costs more than a slow right one when you compare runs.
Where the idea comes from
Mental rotation tasks go back to experiments published in 1971 by Roger Shepard and Jacqueline Metzler. They found that people took longer to answer the further one figure was rotated from the other, as if they were turning it in their heads at a steady speed. Their figures were three-dimensional block objects; ours are flat, which keeps them clear on a phone screen.
What this test does not measure
It measures how accurately and how quickly you compare rotated flat shapes in this format. It isn't an IQ test, a measure of general intelligence or any kind of clinical assessment, and 20 pairs is a short sample: the same person can easily score a couple of points differently from one run to the next.
For another challenge that is about what you see, try Visual Memory, or plan a sequence of moves in the Tower of Hanoi.