Attention at seven is short, and the hour-long app session is mostly a design choice, not a child's limit. Here is what the research supports.
How long
Sixty to a hundred and twenty seconds of work per round, five to eight items, then a change of activity. Ten to fifteen minutes across three or four different games is a good session. Short sessions spread over days beat one long sitting — spacing is the strongest single lever there is for anything that must stick.
Mix the subjects
Interleaved practice — mixing addition, missing signs and brackets in one sitting — looks worse while it is happening and tests better weeks later. Block first when a skill is brand new, then mix.
How strict, and who decides
Everything above describes the setting everyone starts on, Kiddo: a wrong answer never passes, the task waits, and nothing is taken away. That is right for a child who is still finding it hard, and wrong for one who already knows the material and is tapping through it. So it is a setting, on the parent screen, and it changes every game at once.
- Kiddo — no limit. The question waits until it is right and the help grows. Nothing costs points.
- Learner — three mistakes are forgiven on one question; on the fourth the answer is shown and the round moves on. Each mistake costs 2 points.
- Brainiac — one mistake is forgiven. Each costs 4 points, and nothing is hinted at: wrong answers do not fade out and no key lights up.
- Mastermind — the first answer is the answer. Each mistake costs 6 points and there are no hints at all.
Points are what stop a child guessing. A right answer first time is worth ten and a corrected one four, and on the stricter settings a wrong answer takes points off as well — so tapping through four options is worth less than thinking about one. A round never scores below zero, and the cost is the same whether the mistake comes in the first ten seconds or the last, because a penalty that only bites once there is something to lose is no penalty at the start of a round.
Two things do not change. A question the game moves past is never marked right, and it comes back in the daily review the same as any other miss. And on the three stricter settings, a round where the game had to move past anything earns no sticker — otherwise running out of tries would be worth the same as knowing the answer. On Kiddo the sticker always comes, because nothing is ever skipped there: the question waits until it is right, so a child on the gentlest setting can never lose a sticker for finding something hard.
Why there are no streaks here
A streak converts a choice into an obligation, and losing one stings more than keeping it pleases. For a seven-year-old that is a bad trade. Rewards that are not announced in advance — a sticker after a round, a small animal that wanders across the screen — do not damage the wish to play, while promised rewards measurably do.
Praise the work, not the child
"You worked that out yourself" holds up after a failure. "You are so clever" does not: in controlled studies, children praised for cleverness gave up sooner on the next hard thing. Every line of encouragement in this game names the effort.
When your child gets stuck
Wait about eight seconds before stepping in. Then narrow the field rather than giving the answer — "it is not that one" — and only then show how. Aim for roughly four out of five right: that is where a child keeps going. Much higher is boring, much lower and they stop believing they can.
Sit with them for the mistakes
The parent screen lists the facts your child keeps missing. Going through three of those together is worth more than another ten minutes of play, because an error that gets explained is the one that gets learned.