Why does random order beat reciting a table in sequence?
Because a child chanting six, twelve, eighteen, twenty-four is recalling a sequence, not a fact. Ask the same child for 6 × 7 on its own and you often watch them start the chant again from the beginning — that restart is the tell. The answer was never stored anywhere the question can reach it.
Mixed order forces the other kind of recall, and it comes with the awkward half of the evidence attached: mixing question types makes the practice session look worse and the test a day later much better. In one study of mixed maths practice, scores the next day were roughly double. Expect the drill to feel harder than the chant — that is what it is for.
Which facts should come up more often than the rest?
The six, seven, eight and nine tables, and inside them a short list that misbehaves everywhere: 6 × 8, 7 × 8, 6 × 7, 6 × 9, 4 × 8 and 7 × 9. In one large sample of nine- and ten-year-olds answering tens of thousands of questions, 6 × 8 was got wrong more often than any other fact on the grid — more often than it was got right.
Turn on weights and give those facts a two or a three. Their segments widen by exactly the same factor, so a child can see that the hard ones come round more often and nobody has to be told the wheel is loaded. Whatever is already secure stays at one; it is on the rim to keep the drill honest, not to fill time.
How do you run this for a whole class rather than one child?
Fullscreen and a name picker, in that order. Put the wheel on the board, spin the fact, then draw the name — because the class thinks about the question for the second it takes to find out who is answering. That second is the whole technique, and it is why cold-calling classes end up with more voluntary answers over time rather than fewer, once being asked stops depending on who put a hand up.
Keep the pace short. Three seconds a spin is set for exactly this: a wheel that takes eight seconds to stop is a wheel a class watches instead of a wheel a class works with. Twelve facts, twelve spins, elimination on, and the drill ends when the rim is empty rather than when the room loses interest.
What else is this wheel good for?
It is the number wheel with facts typed onto it instead of a range, so weights and elimination behave exactly as they do there. When the question is which pupil answers rather than which fact comes next, the name wheel holds the register.
Draw the pupil instead of the question and it is a classroom name picker. Make the segments unequal on purpose and the drill stops being arithmetic practice and becomes a probability lesson.