How a 10-minute review beats a 30-minute drill.

The claim, in one paragraph.

If your child has 30 minutes to practice multiplication this week, the worst use is one 30-minute Saturday session. A better use is three 10-minute sessions on different days — Tuesday, Thursday, Saturday — even if Tuesday's feels too short to matter. The phenomenon is the spacing effect, one of the most reliably replicated findings in learning research. The interesting part for parents isn't the headline; it's the texture: why short and spread works, and where the advice goes sideways when the learner is eight, has homework on Tuesday, and is exhausted by 7pm.

What the research actually shows.

Cepeda and colleagues (2006) ran the meta-analysis people now cite — mostly adults learning verbal material (word lists, vocabulary, prose). Distributed practice consistently outperformed massed practice on delayed tests, with a medium-to-large effect; the under-10 math evidence is sparser, and we'll come back to it. Cepeda et al. (2008) added the part useful to a parent: the optimal gapdepends on how long you need the material to stick. For “know it next Friday,” gaps of about a day work well; for “still know it in three months,” the gaps should stretch toward a week. The often-quoted 10–20% ratio is a rough simplification of a curve that shifts with the retention interval — a starting point, not a formula.

Dunlosky and colleagues (2013) reviewed ten study-technique candidates and rated them by utility / generalizability across materials and learners. Two came out high-utility: distributed practice and practice testing (recalling material rather than re-reading it). Interleaving was rated moderate. Highlighting, summarizing, re-reading, and mnemonics came out low — modest or unreliable.

For middle-school math, Rohrer, Dedrick & Stershic (2015) ran a trial with 7th graders on slope and graph problems and found that interleaved practice — mixing problem types in a single sheet rather than doing them in blocks — produced a large benefit on a delayed test at matched total minutes. Interleaving forces the kid to choose the strategy on each problem, which is the harder, more transferable move. Elementary-arithmetic RCTs are thinner than the secondary-math evidence.

The other strand worth naming is testing. Roediger & Karpicke (2006) and the long tail of work that followed showed that the act of recallingsomething strengthens it more than the act of seeing it again. A 5-minute quiz on yesterday's problems does more for retention than 5 minutes of re-reading worked examples. This is why “a 10-minute review” has to involve the kid actually doing problems, not watching you redo them.

What this looks like at the kitchen table.

Strip the cognitive-psychology vocabulary out and the playbook fits on a sticky note.

• Short beats long. A 10–15 minute session — about the length of a developmentally reasonable focus chunk for a 7-to-10-year-old — is the right unit. If you have 30 minutes this week, budget three of these, not one big block.
• Days apart, not hours apart. The spacing benefit comes from the gap. Two sessions in one afternoon is mostly one session with a snack break. Tuesday + Thursday + Saturday is the spread.
• Mix the types.Don't do ten subtraction-with-regrouping problems in a row. Do three regrouping, two word problems, two place-value, two of yesterday's fractions — even if it feels slower in the moment. The kid's job is to recognize what kind of problem this is, not to apply a method they've already been told to use.
• Recall, don't re-read.If your kid is stuck, the temptation is to walk them through the worked example again. Better: ask “what's the first thing you'd try?” and let them struggle to retrieve a strategy. A struggle that ends in retrieval is the strongest learning event in the session.
• Yesterday before today. Start each review with two minutes on something from a previous session, not on the new material. The cheapest spacing intervention available — and the one most skipped.

Where the textbook advice gets weird with an 8-year-old.

Most of the spacing research was run on college students with vocabulary lists. Elementary kids aren't college students, and arithmetic isn't a vocabulary list. A few places the recommendations bend:

The “optimal gap” assumes the kid can pick up where they left off.A college student looking at a Spanish flashcard on Thursday remembers Tuesday's. A third-grader looking at long division on Thursday may not remember Tuesday at all. If the gap is so long that every session is a cold start, you've crossed from spacing into forgetting. For elementary math, gaps of one to three days seem to work better than the “a week or more” the lab studies recommend for adults — though the developmental literature on this for under-10s is thinner than we'd like.

Interleaving is harder for the kid, and it should be.The benefit of mixing problem types is exactly that the kid has to think about which type they're looking at. So interleaved practice feels worsethan blocked practice — slower, more wrong answers, more forehead-on-table. In the lab this is called a desirable difficulty. At the kitchen table it's “why is this so hard, this used to be easy.” That feeling is the practice working, not failing — but it's easy to mistake for the latter and bail out.

Recall struggle has a lower ceiling in kids than in adults.A retrieval attempt of about 30 seconds of productive effort is gold. Four minutes of escalating frustration trains the kid to expect math to feel bad. The rule we use: if the child's next move isn't available within roughly 30 seconds, the hint ladder kicks in — the same rule in our note on how Koda decides when to interrupt.

How Koda thinks about review.

Carefully — review scheduling is on the roadmap, not a feature we claim as shipping today. What does ship: a skill-gapview that aggregates per-skill accuracy across all of a child's attempts (coarse, all-time, not yet recency-windowed), and a session-level mastery signal that nudges difficulty up or down on rolling success rate and recommends a mode (learning → practice → exam-ready). The mastery layer also defines a 65% relapse threshold; the constant is in place, but the cross-session pathway that would consume it — “this previously mastered skill has slipped, schedule a re-learning pass” — is still being calibrated. The dispatch on top of that — “here's tomorrow's 10-minute set” — is design-queue, not shipping.

One thing to be careful about: algorithmic spaced-repetition schedulers (the Anki / Duolingo class of tools) work well for vocabulary. For multi-step problem-solving, the evidence that an algorithmic scheduler outperforms a teacher or parent with a calendar is thinner — so the review surface should help you decide what to come back to, while the cadence stays with the parent and the child.

The honest counter-arguments.

“My kid finally has momentum — why would I interrupt that?”Don't. The advice isn't “cap every session at 10 minutes.” If a kid is in a productive run, let them ride it. The advice is about the average shape of practice over a week.

“Spacing means slower progress in the short term.”Often, yes. Blocked practice climbs visibly while it's happening, then falls off a cliff on the delayed test — the gains were rented, not owned. If your benchmark is “how well does my kid do on the worksheet tonight,” blocked wins. If it's “how much of this is still true in October,” spaced wins. Pick which question you're actually asking.

“What about a spaced-repetition app?”Kornell (2009) — on vocabulary flashcards, not arithmetic — found spaced flashcards beat crammed ones by a wide margin. By analogy, off-the-shelf apps are fine for pure number-fact fluency. For anything with structure — word problems, multi-step procedures, the conceptual moves under fractions — the flashcard model doesn't carry the load. Flashcards for fact-recall; real problems for everything else.

One last thing.

The hardest part of this advice isn't the science. It's that “ten minutes, three times this week” sounds like much less work than “thirty minutes, Saturday afternoon” and produces a kind of guilt — am I doing enough? — that the longer block doesn't. The short pattern is doing more, not less. A 10-minute session that ends while there's still energy left is the one they come back to. A 30-minute drill that ends in tears is the one that teaches them math is something to avoid. The number of minutes isn't the measure of the work. What sticks at the end of the week is.

If you'd like to know when Koda ships, the waitlist is here. Related notes: the 15-minute focus chunk, math anxiety in elementary kids, and how Koda decides when to interrupt.