The Human Robot Game and Arrow Cards Before the First Screen
The first step in coding for kids happens on the floor at home. In the human robot game, the child calls out commands such as "forward three steps, turn right, forward two steps" to a helper who plays the robot and carries them out exactly as spoken. When the robot walks into a chair because a step was skipped, the child sees for themselves that a machine does only what it is told, in the order it is told. That moment plants the idea of a sequence, a clear order of steps, in children aged 4 to 7 who do not yet need to read.
Paper arrow cards take the game further. The child lays out a row of cards to guide a toy from one floor tile to another, runs the route, and checks whether the toy arrives. A later version adds a "repeat twice" card, which is a first taste of loops. Screen-free robot toys offer another way to practise the same thing: Cubetto is designed for ages 3 and up, and Bee-Bot, a small bee-shaped robot, can store up to 40 direction commands. Homemade cards build the same skills for almost no cost.
All of these games train the four pillars of computational thinking: decomposition, or breaking a big job into parts; pattern recognition; abstraction, or keeping only the information that matters; and algorithms. A child who retells the steps of brushing their teeth, then repairs a version that has been deliberately shuffled, is writing an algorithm without knowing it. The sign that this stage has settled is easy to spot: the child can build a five or six step route alone, and when it goes wrong they can point to the card that needs changing.
A child who can find the wrong arrow card on their own already has the sense of order and repair that carries straight onto a screen.