How Piko Reviews Saved Characters: Spaced Repetition, Grade by Grade

The short answer

Piko reviews saved characters with spaced repetition. Every character you save out of a photographed word comes back on a fixed schedule — one day after the first correct answer, six after the second, then an interval that keeps stretching — and it appears only when actually due. Your child flips the card, taps Forgot or Know it, and the next date is computed from that answer. Review, streaks, badges, radical families and grade-by-grade progress are Piko Pro features; photographing, the gallery, word cards and the four-way character card are free. If you would rather look before you read, Piko is on the App Store.

Why a saved character is not a learned character

Recognition is generous; memory is not. A single photo of a bowl of fruit can put four characters into the gallery in seconds, which is the whole point of the photograph-and-learn flow. It is also where most flashcard apps quietly stop working. Collect three hundred characters in a month of enthusiastic photographing and nothing ever comes back — the gallery becomes a museum. A long scrollable list feels like progress while being, mechanically, only storage.

So the question behind this search is narrow and worth answering precisely: what actually comes back, when, and how often?

In Piko the answer is unglamorous and specific. The moment a recognised word is saved, every character in it is put into the review queue with a due date of today — no opt-in step, no "add to deck" button, nothing for a parent to forget to press. A character is scheduled once: photographing the same character again next week leaves its existing date alone rather than pushing it back out. That single rule matters more than it sounds, because the usual way families lose a deck is by re-saving things constantly and never letting any card mature.

There is one honest gap. Recognition is character-based, and when a photo produces a word Piko cannot split into characters it knows, the photograph is kept as a record but no new card enters the gallery — and therefore nothing is scheduled. It is the one case where you can photograph something all week and never see it come back, and it is a coverage limit of recognition rather than a bug in the scheduler.

The schedule, in actual numbers

The scheduler is a simplified SM-2, and the numbers are worth knowing because "spaced repetition" on its own tells you nothing about how a schedule behaves over months:

• Forgot — repetitions go back to zero, the interval goes back to zero, and the card is due again the same day. It does not slip to tomorrow.
• First Know it — due in 1 day.
• Second Know it — due in 6 days.
• Every correct answer after that — the previous interval multiplied by the card's ease factor, rounded.

Ease starts at 2.5 for every new card and never drops below 1.3. Because Know it is graded as a solid pass, it leaves the ease factor exactly where it was. That makes the all-correct path easy to predict: a card answered correctly without ever slipping comes back at roughly

1 → 6 → 15 → 38 → 95 → 238 days.

By the sixth successful review you are looking at a card due eight months out. That is the whole argument for spaced repetition in one line, and it is also why review cannot be compressed into one long weekend — a card has to actually mature, and the calendar is part of the mechanism.

What a lapse costs is worth understanding too. A Forgot answer knocks the ease factor from 2.5 down to 1.7, and because the repetitions counter also resets, the card starts again at one day and then grows at 1.7x instead of 2.5x. The same card's future path becomes

1 → 6 → 10 → 17 → 29 → 49 days.

Still climbing, still spaced, but visibly slower. Nothing is ever permanently lost — the character comes back the same evening and you have a second run at it — but a child who is confidently tapping Know it without looking is doing more damage than one who is honest about difficulty.

What a review session actually looks like

The screen is deliberately dull, because it has to work on a five-year-old's attention:

• The card shows the character face down, with the hint Tap to see pinyin.
• Tapping flips it to the full card — meaning, pinyin with tones, the Mandarin pronunciation, and the stroke order and shape origin described elsewhere in these docs.
• The child answers with two buttons: Forgot or Know it.
• A running counter shows 7 / 12, and the session ends with a plain "You reviewed 12 today".

The queue is fixed when you enter — every character that was due at that moment, and nothing more. New cards discovered halfway through a session do not appear in it, so a session never grows under you. With nothing due, the app says All caught up and asks you to come back tomorrow, which is the correct answer rather than a low-key way to manufacture a streak.

The uncomfortable edge case is that there is no daily cap. The queue is exactly what is due, so a family that saved two hundred characters over three busy weeks and then opens the Review tab for the first time gets two hundred cards. That is a real session, not a bug, and it is the one place where spaced repetition is worse than nothing: a queue that long will be abandoned. If you are starting a subscription after a collecting binge, clear the first week by saving fewer new characters and answering honestly, not by trying to push through the backlog.

Grade-by-grade progress, and what those characters are

The curriculum view answers the question the gallery cannot: not "how much have we collected" but "how much of what a class would teach have we collected". Piko ships 730 characters organised by primary grade — 291 in grade 1 and 439 in grade 2.

It is worth being straight about what that list is. It is drawn from the mainland China primary-school character lists (the 人教版 / 统编 standard textbook editions), character by character, and every entry has been validated against the app's own catalogue of 6,732 characters. But it is a curated subset of the first-stage characters, not the full official 1,600-character table, and it mixes characters that are read with characters that are written. Progress is counted by membership — characters you have actually collected that belong to that grade's set — not by filling a bar proportionally to your total. A gallery of 300 obscure characters can therefore show almost nothing on the grade-1 row, which is accurate and occasionally humbling.

The framing also follows the device language. On a Simplified-Chinese device the app speaks the mainland primary-school syllabus your child is actually being taught; on other devices it switches to an overseas-learner framing with English glosses on every card. For a heritage family where the home language differs from the school system, that is the difference between a progress bar that means something and one that quietly measures the wrong textbook.

What review is not, and who it is not for

• There are only two answers. The scheduler behind it can distinguish hard, good and easy, but the card a child sees offers Forgot and Know it and nothing in between. You cannot tell Piko that something was almost. Everything correct grows at the same 2.5x rate, which makes the schedule predictable and the self-assessment coarse.
• Forgetting resets rather than backs off. Many mature schedulers drop a lapsed card back one step. Piko sends it to zero. Simple to explain to a child, harsher on a card that was genuinely half-learned.
• Nothing is capped per day, which is the mirror image of that simplicity: no built-in brake when a queue has grown large.
• The rewards lean on collection. Of the six medals, four measure characters collected — first photo, 30, 50 and 100 — and only Tone master at 20 reviews and the 7-day streak measure the behaviour this page is about. The levels (Beginner at 50, Learner at 150, Scholar at 500, Master beyond) are collection thresholds too. Reviewing well is rewarded, but collecting more is rewarded louder.
• It is not a reading assessment. Nothing here tells you what your child can actually read in a book. It tells you which characters came back and which did not.

What Piko does differently

The differentiator is not the algorithm — simplified SM-2 is the same approach any competent flashcard app uses. It is that collection and review are one pipeline rather than two features. A character that enters the gallery is in the review queue in the same operation, on device, with no account and no server round trip. The core learning flow runs fully on the phone, so review keeps working without a connection, and the recognition that produced the card in the first place never sent the photo anywhere: it is processed locally and never uploaded, and saving a card keeps only the extracted subject sticker on the device. No ads, no third-party content, no sign-in, no child data leaving the phone.

The honest limits are the ones above: a binary grade, a full reset on forgetting, no daily cap, and a 730-character curriculum that is a curated subset rather than the complete official list. And review itself is a subscription feature — free Piko gets you the part that builds the deck, and paying gets you the part that empties it.

The decision you can make now

Try the free half first, and be honest about what you are testing. For two weeks, let your child photograph whatever they are already playing with and save maybe five characters a day. If they start bringing objects to you — pointing the camera at the moon because they want to know what it is — the collection pipeline is working and the review layer is what you are missing. If they never volunteer a photo, more review scheduling will not fix that, and no subscription changes it.

If the collecting does happen, you now know exactly what Pro adds and what it does not: a scheduler that returns each character at 1 day, 6 days, then a stretched interval that reaches months, a 730-character grade-by-grade progress view, and streaks and badges that measure review rather than volume — at $4.99 a month or $35.99 a year in the US (CNY 30 / CNY 218 in mainland China), with a seven-day trial for new subscribers.

Piko is free to download and the collection half works without paying anything. Review is where the retention is, and it is the part you decide on.