Two plastic bricks. Both made of ABS. Both have eight studs on top and a hollow underside. One clicks, holds, and can be pulled apart ten thousand times without loosening. The other rattles, falls off when the model is lifted, and after a week of play has become a box of loose parts.
The difference is not the plastic. It is not the design, which is essentially identical. It is a number measured in hundredths of a millimetre, and getting that number right is the entire industry.
The famous patent was not the invention
The brick most people picture was patented in 1958, and the patent was not for a studded brick. Studded self-locking blocks already existed: a British toy company had been selling them since the late 1930s, and samples had reached the Danish workshop that would later become famous well before it produced anything similar.
What the 1958 patent covered was a change to the inside. Early studded bricks were hollow underneath, so the walls had to be thick enough to grip the studs of the brick below by friction alone. That works, but it works unevenly: a brick sitting on one stud grips differently from the same brick sitting on four, and a brick pressed onto a baseplate grips differently again. The clutch is inconsistent, and an inconsistent clutch makes large models impossible, because some joints fail while others hold.
The improvement was to put a ring of tubes inside the brick, positioned so that they grip the studs from the inside. Now every connection has the same geometry regardless of how many studs are involved, and the gripping force comes from controlled elastic deflection rather than from the walls squeezing. That is the actual invention: not the brick, but the consistency of the brick.
Holding is not friction, it is a spring
The clutch is usually described as friction, which is misleading. Friction would wear out: plastic sliding against plastic a few thousand times would polish both surfaces smooth, and the brick would go loose.
What actually happens is interference fit. The stud is very slightly larger than the tube or socket it enters. When you press the brick down, the plastic deforms elastically — it bends, within its elastic limit, then holds. The restoring force of that deformation is what keeps the brick on. Because the deformation is elastic, it is fully recovered when the brick is removed, which is why the clutch survives repeated use: nothing is being worn away, only flexed.
This is the same principle as a press-fit bearing, and it is why the fit has to be right within a very narrow band. Too loose and there is no deflection, so no force. Too tight and the deflection exceeds the elastic limit, so the plastic yields, and once it yields the joint is permanently slack. There is no gradual tuning here. The window between "does not hold" and "destroys itself" is small, and it sits inside a few hundredths of a millimetre.
Why the mould is not the shape of the brick
ABS shrinks as it cools, by roughly half a percent. A brick that measures 32 millimetres out of a mould will measure slightly less once it is solid.
The shrinkage is not uniform. It depends on wall thickness, on flow direction, on where the material entered the cavity, and on how the cooling channels are arranged. A brick has thick corners and thin walls, so it shrinks differently in different places, which means it does not simply scale down — it distorts.
So the mould cavity is not cut to the shape of a brick. It is cut to the shape of a brick distorted in the opposite direction, so that after shrinkage it lands on the correct geometry. The toolmaker has to predict the distortion before the first part exists, using simulation and experience, and then correct it by removing steel — because you can always take material out of a mould, and you can almost never put it back.
This is the same reasoning as the tin toy printer pre-distorting a lithograph so that the press can undo it. In both cases, the manufacturing process deforms the object, so the tool must be built wrong on purpose.
What actually separates a good brick from a bad one
Cheap bricks are frequently made of the same resin as expensive ones, from the same suppliers. The difference is almost always in the tool.
A high-volume brick mould is a multi-cavity tool, producing many parts per cycle, cut from hardened steel, with cooling channels designed to keep every cavity at the same temperature. It costs a great deal of money and it has to be maintained: as it wears, the cavities grow, and a cavity that has grown by a few hundredths of a millimetre produces a stud that no longer interferes correctly.
This is why the same factory can produce excellent bricks on one machine and poor ones on another, and why a brand's quality can drift over years without any announced change. Nobody decided to make the brick worse. The tool simply aged.
It is also why the "clutch power" specification exists as a measurable thing rather than a marketing word. Manufacturers test it: a freshly moulded brick must hold a specified mass when suspended, and must still hold a specified fraction of that after a defined number of assembly cycles. A toy that fails the second test is a toy that will be in a bin within a month.
The grid is the product
Everything in the system is a multiple of one dimension. The stud pitch is eight millimetres. A one-stud brick is eight millimetres wide. A two-by-four brick is sixteen by thirty-two. Height works on a different module: a standard brick is 9.6 millimetres tall, which is exactly three times the height of a flat plate at 3.2 millimetres.
That factor of three is not arbitrary. It means that a plate stack can always be converted into a brick stack and back again without breaking the grid, which means any model can be reinforced, and any surface can be levelled. Without it, the system would only be able to build in one direction at one resolution.
This is why the patents expiring did not destroy the business. A competitor can copy the shape, but to be compatible they must hit the same grid, the same tolerances, the same clutch specification, and the same shrinkage compensation — and once they have done all that, they have built the same factory. The moat was never legal. It was dimensional.
What the click is telling you
The sound a brick makes when it seats is not a feature. It is a byproduct of the stud bottoming out in the tube at the same moment the brick face meets the brick below. Two surfaces landing together produce a single transient instead of two, which the ear reads as a click.
A brick that seats with a duller, softer sound is usually one where the two contacts are not simultaneous — the stud is reaching the bottom before the faces meet, or after. It will still hold, briefly. It will also be the brick that comes loose first.
Which is a useful thing to know when buying: the click is a free quality test, and it costs nothing to listen.