Ask a seven-year-old what six plus seven is and watch their hands. If fingers appear under the table, or the lips move, the answer is being worked out. If it arrives in about a second with nothing moving, it is being remembered. Same answer, two completely different operations, and the difference decides what kind of practice will help.
Recall and calculation are different jobs
A recalled fact comes straight out of memory, the way your own phone number does. A calculated fact is rebuilt each time from something else: six plus seven as double six and one more, or eight sevens as eight fives plus eight twos. Both are legitimate, and the rebuilt version is usually a sign of good thinking rather than a problem to be stamped out.
The reason recall matters is narrower than it is often made to sound. It is not that quick children are cleverer. It is that a child working on 348 divided by 4, or comparing three eighths with two fifths, is holding several things in mind at once, and every fact they have to reconstruct in the middle of that costs some of that space. A ten-year-old who still counts up for six plus seven will usually manage the small step fine and lose the thread of the big one. That is the whole argument for fluency: not speed for its own sake, but keeping the working space clear for the part that is genuinely hard.
Useful fluency is accurate, reasonably quick and flexible: a child who knows that seven eights are fifty-six and can also see why has something more durable than the sound of the words.
Why the timer backfires
Worry and mental arithmetic compete for the same resource. Holding an anxiety in mind — the clock, the number of blanks left, what the score will look like next to everyone else's — occupies part of the same limited attention that retrieval needs. So a timer makes recall worse in exactly the children who are already shaky, and then hands them evidence that they are bad at math.
The loop is easy to see once you know its shape. Timer starts, mind goes blank, score is low, child concludes they cannot do this, child avoids practice, facts stay unlearned, next timed test is worse. Children caught in it will often do fine on the identical questions at the kitchen table on a Saturday with no clock in sight, which is the clearest possible sign that the arithmetic is not the problem.
There is a measurement problem too. Speed is a consequence of fluency, not a cause of it, and pushing for speed before the facts are in memory produces guessing — a child who says fifty-four for seven eights and moves on quickly has just practised an error. If you want a number to track, count how many facts came back with no visible strategy, out of how many you asked. Twelve out of twenty with no counting tells you far more than twenty-two answers in a minute.
A child racing their own quiet count, with nobody watching, is a different thing from a whole-class race. If yours genuinely enjoys beating yesterday, fine. If they have ever cried about it, retire the clock completely.
Which facts, in what order
The most useful thing you can do is stop treating the facts as one flat block of a hundred. They arrive in families of very different difficulty, and most of them fall out of a few.
For addition and subtraction, roughly ages six to eight, a workable order runs: adding zero and one, then doubles (4+4, 7+7), then the pairs that make ten (7+3, 6+4), then near doubles (6+7 is double six and one more), then bridging through ten (8+5 as 8+2+3). Teach subtraction alongside rather than afterwards. Thirteen minus six belongs with six plus seven, and a child who sees them as one family has half as much to hold.
For multiplication, roughly eight to eleven: twos, tens and fives first, because they are patterned and quick. Then fours as doubled twos, then threes. Then sixes as doubled threes, and nines, which have their own pattern and sit conveniently next to the tens. Then the squares, which children tend to enjoy. What is left after all that is a short list, often eight to ten facts clustered around six, seven and eight multiplied by each other.
Write that list down and put it somewhere visible. Seeing that the entire remaining job is 6×7, 6×8, 7×8, 7×9, 8×9 and their reversals changes how a ten-year-old feels about the task, because it turns an endless-looking chore into a countable one. A twelve-by-twelve grid with the known facts crossed off does the same work and is more satisfying to fill in than any book of exercises. Division comes last and mostly comes free: work it back from multiplication, and only practise the ones they cannot reverse quickly.
Spacing and interleaving, in kitchen-table terms
Two ideas do most of the work here. The first is spacing. Three five-minute goes spread across a week beat one twenty-minute go on a Sunday, even though the total time is shorter. Revisit at growing gaps — the next day, three days later, a week, a fortnight — and whatever survives the fortnight tends to stay.
The second is interleaving: mixing the sixes with the threes and the eights in the same handful of cards rather than doing twenty sixes in a row. Mixed practice feels worse at the time and produces better results a week later. Blocked practice feels excellent on the day, because after four questions the child stops reading the numbers and simply adds six each time, which is not retrieval at all.
In practice this means a small pile of cards rather than a page. Keep a working set of about ten: eight the child mostly has and two being learned. When a fact comes back instantly on two separate days, retire it to a weekly pile and bring in a replacement. Five minutes might run like this:
| Time | What happens |
|---|---|
| 0–1 min | Five facts they own, answered fast, deliberately easy |
| 1–3 min | The working set of about ten cards, shuffled and mixed |
| 3–4 min | The two new facts, said out loud with a reason attached |
| 4–5 min | Three cards pulled from the retired pile |
The warm-up is not filler: starting with facts a child owns sets the tone for the four minutes that follow.
Telling memorisation from understanding
A child can recall perfectly and understand nothing, and can understand thoroughly with no recall at all. Both are worth fixing, but the fixes are opposites, so it is worth knowing which you are looking at. Four checks, none of which takes a minute:
- Ask 6×7, then ask 6×70 or 60×7. Bare recall does not transfer. Understanding does.
- Ask them to show you 6×7 as something real: six rows of seven chairs, six bags of seven apples, a rectangle of dots.
- Ask whether 7×6 or 6×7 is easier for them. A child who says they are the same fact has already halved the amount to learn.
- Give them a fact you know they do not have and ask how they would work it out. "I don't know" and "eight sevens is eight fives plus eight twos" are very different states to be in.
If the recall is missing but the understanding is there, spaced retrieval is the answer and it will work. If the answers come out fast but 6×70 defeats them, put the cards away for a fortnight and go back to objects, arrays and drawing. Facts stacked on top of nothing collapse the first time the numbers get bigger, and the collapse tends to arrive around fractions and long division, when it is expensive.
If the fear is already there
Plenty of children come to this with a history, and repair takes a few weeks of mostly doing less.
Take the clock away entirely. Shrink the working set to facts they already have plus one, so that most of every session is success. Say the answers together first, then let them go alone. When something is wrong, do not say "no" — say the fact back correctly in a level voice and come back to it two minutes later. Let them see the answers if they want to; the aim is to get facts into memory, not to catch anyone out. And pick an hour that is not the tail end of a long school day; a bad session costs more than a skipped one.
Then say the thing out loud, because children rarely work it out for themselves: being quick is not what makes someone good at math. Most real mathematical work is slow and involves being stuck. What you are building with a handful of cards is a small set of answers that arrive without effort, so that the slow and interesting thinking has somewhere to happen.