🔄 MENTAL ROTATION

Exactly two of the four are the target turned.
The other two are its mirror image.
Pick both. 60 seconds.

Find one corner of the target that sticks out in an odd way, then look for that corner in each option. Turning the whole shape in your head is the slow way.

Mental rotation test — free spatial reasoning test in your browser

A block figure sits at the top of the screen. Below it are four more. Exactly two of the four are that same figure, turned; the other two are its mirror image, also turned. Tap the two that match. You get sixty seconds, and a wrong pair costs you three of them.

The figures are generated fresh every trial, so there is nothing to memorise and no set of answers to learn. Each figure is a bent arm of blocks, a flat version of the cube figures in the original experiments. A run starts with five blocks and two bends. Each level then adds a block, adds a bend, or makes the turns less tidy, one change at a time. Free, no download, and it works the same with a thumb as with a mouse.

How to play

Look at the target at the top. Every option below is either that shape turned, or its mirror.
Tap two options. Tapping a selected one again deselects it, so a slip costs nothing until you commit the second pick.
Both right scores a point and deals a new figure straight away.
A wrong pair costs 3 seconds. All four options are then labelled TURNED or MIRROR for a moment, so you see what you missed.
On a desktop the number keys 1 to 4 select the options.
Your score is correct trials in 60 seconds, and your top 5 is kept on this device. The result screen adds your median time per figure, how many you missed, and a band from Warming Up to Elite Rotator.

Why the answer arrives slower the further the shape is turned

In 1971 Roger Shepard and Jacqueline Metzler ran an experiment that has been repeated in psychology classes ever since. They showed people pairs of block figures and asked a single question: same shape, or mirror image? What mattered was how people answered. The time taken rose in a straight line with the angle between the two figures. Twice the rotation, twice the extra time, right through to 180 degrees.

That linearity is the interesting part, because it is what you would expect if people were actually turning something at a roughly constant rate rather than comparing lists of features. It is one of the more concrete pieces of evidence that visual imagery behaves in some ways like vision, and it is why "mental rotation" is a phrase at all.

The paper-and-pencil version most people have taken is the Vandenberg and Kuse test from 1978, which is where the format on this page comes from: one target and four alternatives, two of them correct. Because you must pick both matches, you cannot get credit for spotting one and guessing; and because the two wrong answers are mirrors rather than different shapes, you cannot answer by counting blocks. Every wrong option has exactly the same number of cells and exactly the same outline lengths as the right ones. Only the handedness differs.

The practical advice: stop rotating the whole figure. Pick one distinctive feature of the target, usually a limb that comes off the body at an odd place, and check only whether that feature turns the same way around the shape in each option. Handedness survives rotation, which is the entire trick. People who are fast at this test are almost never rotating faster than everyone else. They are rotating less.

What your result means

The heading on the result screen is a band. The bands come from your pace, not from anyone else's scores:

⚡ Elite Rotator: 24 or more right, about two seconds a figure.
🦅 Quick: 18 to 23 right, around three seconds a figure.
🎯 Sharp: 12 to 17 right, three to five seconds a figure. Twelve is the target a new player sees.
🔄 Steady: 6 to 11 right.
🧊 Warming Up: fewer than 6.

The two bars under it compare your median time on small turns (up to 90°) with large ones (135° to 180°). A longer second bar suggests you were turning the shape in your head, as Shepard and Metzler's volunteers did. Two bars about level usually mean you were matching a feature instead, which is the faster habit. With fifteen or so trials, a difference of a few tenths of a second can be chance, and practice tends to flatten the gap anyway.

Practice helps. A 2013 review of 217 studies by David Uttal and colleagues found that spatial skills improve with training, that the gains last, and that they carry over to tasks nobody practised.

This is a game, not an assessment. There is no "better than 80% of people" line because nobody has collected norms for a sixty-second, flat version of this test, and the published norms come from a paper test with 3D figures and a time limit of several minutes. Scores also move around with screen size, with whether you are using a thumb, and with how much sleep you had. Race your own best and leave it there.