The Summer-Slide Math Your Curriculum Ignores

First week back, and it's placement-test week. You pull out last spring's assessment, or the curriculum's own diagnostic, to see where your kid stands before picking up where you left off in May. And there it is: 7 × 8, a fact that was instant in May — genuinely instant, drilled to automaticity, checked off — now gets fingers again. Counted out, one group of eight at a time, the way it was solved back before it was "done."

You're not surprised, because you've seen this before, and so has every homeschool parent you know. It's common enough to have a standing slot on the calendar: open almost any curriculum's scope-and-sequence and the first few weeks of fall are penciled in as review. Not new material — old material, again, because last year's facts didn't survive the summer. It's treated like weather: something that happens to everyone, every September, that nobody thinks to question.

Worth questioning, though, because the review unit is not a fact about your kid, or about math being hard to hold onto. It's the predictable output of a specific scheduling choice — and once you see the choice, the fix is smaller than the problem looks.

It's also not a small unit. Scope-and-sequences that budget two or three weeks for "fall review" are budgeting conservatively; plenty of families find themselves well into September still re-covering last spring's fact families before new content can safely start. Multiply that across addition, subtraction, and whichever operation was newest in May, and the "review" isn't a warm-up lap. It's a significant fraction of the fall calendar, spent putting the kid back where she already was.

Forgetting Runs on a Schedule, and So Does the Slide

Start with the part that isn't anyone's fault. Hermann Ebbinghaus charted the forgetting curve in 1885, and every study since has confirmed its shape: newly learned material fades fastest right after learning stops, then more slowly, but it fades — for everyone, every time, with no exceptions for facts your kid "really knew." (We've laid out the mechanics in The Science of Forgetting.)

Now put a homeschool summer against that curve. Roughly ten to twelve weeks between the last worksheet in May and the first placement test in September — ten to twelve weeks in which not a single fact gets retrieved. Zero retrieval, laid across a curve whose steepest losses happen early. The kid who needs fingers for 7 × 8 in September isn't a surprise the curriculum failed to prevent. She's the output the curve predicts, arriving exactly on schedule.

So the slide isn't a kid problem ("she just forgets things"), and it's not a content problem — the facts were taught fine, tested fine, mastered by any reasonable definition in May. It's a scheduling problem. Curricula plan instruction with real care: which fact families in which order, how much practice before the next unit, when to introduce regrouping. They plan retention not at all. The summer gap isn't an accident the design failed to catch — it's designed in, because nothing in a typical scope-and-sequence ever asks the delayed question: will this still be true in September?

"Knew Them in May" Is a Claim About May

Here's the sleight of hand worth catching. An end-of-unit test, a clean worksheet, a "mastered" checkmark in the curriculum's tracker — all of these measure one thing: could she do it that day. That's a real measurement, and it's not nothing. But it is a claim about May. The claim a parent actually cares about — "she has her facts" — is a claim about November, or next spring, or whenever the fact is next needed for something harder. And no May test can establish a November claim, because nothing happened between May and November to check.

This is the same confusion we untangled for language learners in You Don't Have a Memory Problem — mastery is a maintained state, not an event you pass through once and keep. A fact that was mastered and then never checked again doesn't stay mastered by default; it just stops being measured, and the curve does the rest quietly, off the page, where the curriculum can't see it. The checkmark in May didn't lie. It just stopped being true sometime in July, and nothing was watching when it happened.

Why Math Facts Are the Worst Place for This to Happen

Every subject loses something to a long break. Math facts lose more than most, because they aren't a unit that sits beside the others — they're the substrate everything else stands on. Multi-digit multiplication needs instant single-digit multiplication underneath it. Fraction simplification needs instant factoring. Algebra needs both, instantly, buried inside problems that are actually about something else entirely.

This is the automaticity argument we've made for language learning in Why Mastery Learning Beats Moving On — working memory is a small room, and whatever a fact's recall doesn't consume is what's left over for the actual thinking. It applies to arithmetic without modification. A kid who has to derive 7 × 8 mid-problem isn't just slower on that one step — she's spending the room she needed for the regrouping, the place value, the actual new idea the problem was trying to teach. The fact-recall tax gets paid on every single step of every problem that touches it, for as long as the fact stays slow.

Which is why the summer slide in math facts doesn't cost a homeschool family one review unit and then move on cleanly. It degrades everything built on top of the facts until they come back — new material taught at a pace that assumes instant recall, landing on a foundation that's currently unreliable. The review unit isn't catching the family up on arithmetic. It's catching the whole fall's math back up to where it needs the substrate to be.

Picture the actual multi-digit multiplication problem a kid meets a few weeks into fall. It asks for four or five separate fact retrievals stacked inside one procedure, each one feeding the next step's regrouping. If every one of those retrievals is instant, the working-memory room is spent entirely on the procedure — which digit carries where, what the problem is even asking. If even one retrieval is slow, the room gets spent counting instead, and whatever attention the new procedure needed simply isn't there. That's not a motivation problem or a focus problem. It's a room that was already full before the actual math started.

What Keeping May Would Have Cost

Here's the arithmetic that should sting a little: keeping May's facts through September is cheap. Not free, but close to it — and the same forgetting curve that predicts the loss also tells you exactly how to prevent it: spaced retrieval. A fact that's starting to fade needs one successful retrieval to reset its clock — and each successful reset stretches the next safe interval. A fact checked a handful of times across a summer, at growing gaps, barely needs anything at all by the third or fourth check. A few minutes a week, total, for a whole fact family. That's the price of keeping May.

Compare that to what actually happens without it: a multi-week review unit, re-teaching everything to everyone, whether or not it actually faded. And that's the part worth sitting with — the review unit isn't really a review. Without any record of which facts survived the summer and which didn't, the curriculum has no way to target the ones that actually slipped. So it drills the whole fact table again, for every kid, because brute force is the only tool available when nothing was verifying anything. The review unit's real function isn't re-teaching. It's diagnosis — the slowest, most expensive way to find out what a scheduled check would have told you for the cost of a few minutes a week.

A verification schedule already knows. It re-checks exactly the facts due for checking, at exactly the interval each one has earned, and it surfaces exactly the facts that slipped — not the ones that didn't. And this is the part no parent can run by hand, not really: which of two hundred facts is due for its check this week, at what interval, for which of three kids, given when each one was last recalled and how fast. Nobody tracks that on paper across a whole family. It has to be a system's job, because it's a scheduling problem, and scheduling is exactly the kind of thing curricula have never been built to do.

This is the same complaint that shows up, worded almost identically, in every homeschool group thread about drill apps: the countdown-timer kind that leaves a kid in tears before the facts even land, the coin-economy kind where twenty minutes of "math time" turns out to be four minutes of math and sixteen minutes of shopping for a pet. Different failure, same root cause — neither one is actually watching what survives. They measure performance during the drill, which is real but is the wrong quantity, and they have nothing to say about November.

Designed In, Designed Out

The review unit feels like diligence — like the responsible thing, the thorough thing, doing right by a kid who needs the foundation solid before September's new material lands on it. And in a sense it is diligence. It's just diligence spent paying a bill that a design choice created, penciled into the scope-and-sequence every single fall like it's unavoidable. It isn't. The slide was designed in by a curriculum that plans instruction and never plans retention. It can be designed out by one that does.

September doesn't have to be an excavation, digging back down to facts that used to be on the surface. It can be a starting line — the same one May left off at, because something spent the summer quietly checking, a few minutes a week, that it was still there. (If this sounds familiar, it's the same pattern that shows up when kids relearn the staff every fall in piano lessons — the same forgetting curve, the same fix, a different subject; see The September Reset.)