Why You Start Recognizing JLPT Practice Questions (And What To Do Instead)

If you've studied with the same textbook workbook, mobile app, or Anki deck for more than a few weeks, you've probably noticed it: you start getting questions right not because you understood the passage, but because you remember the answer from last time. The reading passage becomes familiar. The distractor you eliminated last week is the same distractor this week.

That's not a discipline problem. It's a structural limitation of any fixed question bank, no matter how well it's built.

Spaced repetition is genuinely great — for vocabulary and kanji

Tools like Anki, WaniKani, and Bunpro use spaced repetition systems (SRS) to resurface a flashcard right before you're about to forget it. For discrete facts — a kanji reading, a vocabulary word, a single grammar pattern — this is one of the most well-evidenced study techniques available, and there's no reason to avoid it.

The problem shows up specifically in reading comprehension. A JLPT reading section isn't testing whether you recognize a passage — it's testing whether you can extract meaning from Japanese you haven't seen before, under time pressure, on the first read. If you've seen the passage or a near-identical one three times, you're no longer practicing that skill. You're practicing pattern-matching against your own memory of the answer key.

The actual JLPT gets this right — practice should too

Every real JLPT sitting uses entirely new passages. The exam doesn't reuse content from previous years for exactly this reason: familiarity with a specific passage is not the skill being measured. Practice materials that recycle the same fixed bank of passages are, structurally, training a different skill than the one you'll be tested on.

What "fresh, level-matched" practice actually requires

Generating a new passage isn't enough on its own — plenty of random Japanese text exists online. The practice has to hold three things constant while varying everything else:

• Script and vocabulary density matched to your level. N5 material should be almost entirely hiragana; N1 material should use the kanji density and formal register of an actual newspaper editorial. Content that doesn't respect this teaches the wrong reading speed.
• Distractor quality matched to your level. At N1 and N2, all four answer choices should sound plausible from a surface reading — the correct one is the only one the passage actually supports. At N5, the wrong answers can be more obviously wrong. Get this backward and you're either memorizing "the weird one is never right" (too easy) or getting discouraged by trick questions above your level (too hard).
• Never the same passage twice. The only way to guarantee this at scale is to generate new material for every session, rather than drawing from any fixed bank — free or paid, big or small.

This is the whole premise behind JLPT-X

Every reading passage, grammar set, and kanji drill is generated fresh when you request it — built to your exact level's script and distractor rules, never pulled from a bank you'll eventually memorize. Pair it with an SRS tool like Anki for your vocabulary and kanji recall, and use fresh generation specifically for the skill that fixed decks can't train: reading something for the first time and actually understanding it.