Spatial construction

Block design test where solved and solved fast stay separate

Eight patterns, built from blocks whose six faces are two solid red, two solid white and two split down the diagonal — three small designs with a minute each and five larger ones with two minutes each. Drag a face into a cell, or pick one with the number keys, and the clock stops the moment the grid matches. The report gives designs finished, of those how many were finished inside a third of their limit, and how many placements each one took, because a design solved in nine moves was seen and the same design solved in thirty was searched for.

100% free
No signup
8 designs, 2 sizes
Six-faced blocks
Placements counted
Point schedule printed

Reproduce a printed design with blocks. Every block has the same six faces — one solid red, one solid white, and four showing the same red-and-white diagonal at each quarter turn — so the whole task is choosing a face and a rotation for each square. 8 designs: three of four blocks with 60 seconds each, five of nine blocks with 120.

The six faces

solid red
solid white
split with red in the top left
split with red in the top right
split with red in the bottom right
split with red in the bottom left

Red against white is a difference in lightness as well as in hue, so the two are distinguishable without any color vision at all — which matters here, because the alternative would be a construction task that quietly measured the retina.

The scoring schedule, in full, before you start

Designs: 8 — three of four blocks, five of nine
Limits: 60 seconds for a four-block design, 120 for a nine-block one
Base: 4 points for finishing a four-block design, 6 for a nine-block one
Bonus: 4 more inside a third of the limit, 2 inside half, nothing after that
Blocks: identical, six faces each, unlimited supply
Also recorded: placements made, against the number the design needs

Those limits and those bonuses are this page’s own and are printed here so that a second run means something. They are not taken from any published administration, and a total from this page converts to no scale used anywhere else.

The practice design has no limit on it and is scored nowhere. Take it if you have not dragged blocks in a browser before — the controls costing you thirty seconds is the one way this version can measure something it did not mean to.

How to take the block design test

Pick a face, put it in a cell, and beat the clock if you can.

Learn the six faces on the untimed practice design
Every block has the same six faces: solid red, solid white, and four diagonals — red over white and white over red, each in two orientations. The practice design is a small one with no clock on it at all, and it exists so the first thing you meet under time pressure is a pattern rather than an interface. Drag from the tray, or press 1 to 6 to pick a face and place it with the arrow keys and Enter if you would rather not drag.
Work the diagonals first, then the corners
Most designs are decided by where the split faces go, because the solid cells are usually obvious and the diagonals are where a pattern is either seen whole or assembled cell by cell. Every design here has half-turn symmetry, so what you work out for one corner tells you the opposite corner for free — noticing that is worth more than moving quickly, and the placement count in the report is there to show you whether you did.
Read three columns, not a total
The result separates designs finished, designs finished fast, and placements used against placements needed. A run that finished six of eight slowly and a run that finished six of eight quickly are different performances, and one number cannot hold both. The point schedule that turns those into a score is printed above the table, in full, so that a total is something you could recompute by hand rather than something the page asserts.

Technical specifications

Designs	8 scored — three 2 by 2 patterns and five 3 by 3 patterns, in that order — plus one untimed practice design that reaches no figure
Time limits	60 seconds for a four-block design, 120 seconds for a nine-block design. The countdown reads the engine's own stimulus onset every frame rather than a timer started in a handler, so it does not drift over two minutes
The blocks	Six faces each: two solid red, two solid white and two split corner to corner. That is the standard cube of the paradigm, and it is what makes a diagonal edge possible at all — no face is half a triangle by itself
Patterns	Generated with half-turn symmetry, so every design is its own 180-degree rotation. Real designs in this tradition are symmetric, and an asymmetric one is a different and harder task that the time limits here would not fit
Points	4 for a completed small design and 6 for a completed large one, plus 4 more inside a third of the limit or 2 more inside half. This schedule is ours, it is printed on the results panel, and it is not any published test's bonus table
Placements	Every face put into a cell is counted, including replacements. Four and nine are the minima; the ratio of what you used to what was needed is reported per design and is the closest thing here to a record of how you solved it
Response	Pointer drag from the tray, or keys 1 to 6 to choose a face with Enter to place it and R to clear a cell. Either route places the same block, and the count does not care which you used
Reference figure	None, and none is possible. The norm tables for the subtest this resembles are sold with a physical kit under a qualification requirement, so this page reports seconds, placements and its own points, and produces no scaled score of any kind

Frequently asked questions

Is this the block design subtest from an IQ battery?

No. It is the same century-old construction task that subtest was built from, rebuilt in a browser with patterns, limits and points of our own. The published subtest is a physical kit of blocks, a printed booklet of designs, an examiner who watches how you work, and a norm table that converts your raw performance into a scaled score — none of which exists here. A number from this page cannot be entered into any battery, compared with a scaled score, or reported to anybody as one.

Why do I get points for speed instead of just for finishing?

Because in this task speed is part of the ability rather than a bonus attached to it. Someone who sees the pattern as four quadrants places nine blocks and stops; someone who has to try faces until the cell looks right will also finish, given time, but has done something different on the way. A completion-only score treats those as identical. The schedule here is deliberately coarse — a large bonus inside a third of the limit, a small one inside half, nothing after — because a finer curve would suggest a precision that eight designs cannot support.

Does dragging with a mouse make this slower than using real blocks?

Yes, and by an amount nobody here can measure for you. Picking up a cube, turning it to the face you want and setting it down is a practiced motor act; dragging a square across a screen is a different one, and a trackpad is different again from a mouse. That cost lands on every design equally, so comparisons within your own run survive it, and it is the main reason the raw seconds on this page should not be lined up against any figure quoted from a table-top administration.

Why is every pattern symmetric?

Because the designs in this tradition are, and because symmetry is what makes a two-minute limit reasonable for nine blocks. A pattern that is its own half-turn gives you the second half of the grid once you have solved the first, which is the insight the task is partly testing for. It also keeps the generator honest: a random arrangement of faces would produce designs of wildly varying difficulty, and eight of those would measure the luck of the draw as much as anything else.

What does the placement count tell me that the time does not?

How the design was solved rather than whether it was. Nine placements on a nine-block design means every block went straight to its cell — the pattern was decomposed before the first move. Thirty means faces were tried and replaced, which can still finish inside the limit and is a completely different route to the same outcome. It is the one figure on this page that no table-top administration records, because nobody watching can count reliably at speed, and it is free here.

I ran out of time on the last few designs. Does that mean anything?

On its own, very little. The large designs come last and there are five of them, so the end of the run is where fatigue, an interface you are still learning and genuine difficulty all arrive together. What is worth looking at is the placement ratio on the ones you did not finish: a design abandoned at nine placements out of nine needed was nearly solved, and a design abandoned at forty was not being solved at all. Those are different failures and the table keeps them apart.

Kohs blocks, and what a browser can and cannot keep of them

The task is a hundred years old and its original purpose was diagnostic breadth: a construction test that needed no language, so it could be given to people whose verbal testing was uninformative or impossible. The design published as Kohs (1920), The block-design tests, Journal of Experimental Psychology used exactly the cube that survives in every version since — six faces, two solid in each color, two split along the diagonal — because that cube is the minimum from which a pattern with diagonal edges can be built at all. What made it durable is that it puts perception and action in the same loop: you are not asked whether two shapes match, you are asked to produce the match, and the sequence of moves is visible while you do it.

Three things transfer cleanly into a browser and one does not. The blocks transfer, the symmetric designs transfer, and the clock transfers. What does not is the norm table, and the reason is worth being plain about: those tables are the commercial product, sold with the physical kit to qualified purchasers, and reproducing one here would be both an infringement and an invitation to read a browser score as a clinical one. So this page prints its own point schedule where a scaled score would go, and prints it in full, so that the total is arithmetic you can check rather than a number handed down. The three columns — finished, finished fast, placements used — are what the run actually observed.

Construction sits slightly apart from the rest of this cluster, and that is why it is worth taking alongside them. The spatial reasoning test asks you to recognize the right answer among four; the paper folding test asks you to imagine a transformation and then recognize its result. Only this one asks you to make the thing, which brings in decomposition, a motor plan and the willingness to abandon a half-built quadrant — none of which a multiple-choice item ever sees. Timed construction under a clock is also close in spirit to the short, heavily speeded employment batteries, so if that is what brought you here, the criteria cognitive aptitude test and the Wonderlic-style test practice the same pressure across a much wider range of item types.

A reaction time here is the interval between the frame that painted the stimulus and the timestamp the browser attached to your key, both read from the same monotonic clock. What neither can see is the display pipeline behind it, so on a 60 Hz screen roughly 16 ms of every figure below is the machine rather than you. That is the timing floor: two numbers closer together than that are the same number, and this page reports no precision it cannot support.

Against a sixty-second limit a frame is nothing, and the figure that matters here is the design time in seconds. The floor is stated because the results panel does print milliseconds in one place — the interval from the design appearing to your first placement, which is short enough for a frame to be a visible share of it.

Not affiliated with, endorsed by, or connected to the owner of Wechsler Adult Intelligence Scale. Wechsler Adult Intelligence Scale is a trademark of its owner and is used here only to name the assessment this page prepares for. The questions on this page are our own: no part of the published test is reproduced, and a score here is not comparable to a score from the real instrument.

This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish a constructional difficulty, a right-hemisphere problem or anything a neuropsychologist tests for. Only a qualified professional, working with more than a browser, can make that judgment.

The subtest this resembles is used clinically precisely because a trained examiner watches the process, not the total. Nobody is watching here, and eight designs on a trackpad support no inference about a person.

Where the eight designs are generated and scored

Every number on this page is worked out by JavaScript running in the tab you are reading it in. Your answers, your reaction times and your score are never uploaded, logged or kept — which is also why the test carries on working after you disconnect from the network, and why nothing here can be held back behind an email address.

The patterns are produced in your browser from a seed drawn at the moment you press start, and every placement, every clock reading and every point is computed in the same tab. No design is fetched from a server and no result is sent to one; the run exists until the tab is closed and then it does not.