Shindan: Japanese Level Test

Screenshots

About

Shindan does not teach. It measures, and then it points. You already know the feeling: you study, you make progress, and you still cannot say what your weakest link is. A grammar app knows how much grammar you have. It does not know that particles are what keeps costing you points. Nothing you own sees the whole picture, because nothing you own asks about all of it at once. Shindan asks about all of it at once. One sheet, under a clock, and at the end two answers: where you stand, and the one thing worth studying next. TWO ANSWERS, NOT ONE Your level, on the JLPT scale, from the material that has a level: vocabulary, kanji readings and grammar constructions. Nobody asks "am I at practical level three" – they ask "will I pass N3", and that is the scale this answers in. Your weak spots, as a share of your mistakes: particles, verb forms, register, contractions, mimetics, counters and pitch accent. These have no JLPT level and never will – a particle you meet in week one and use for life. Measuring them on the same scale would make that scale mean nothing, so they are measured apart. NO PRACTICE, NO REPETITION, NO STREAK There is nothing to grind here and that is deliberate. A sheet measures once. Answers are not marked as you go, because seeing three mistakes in a row changes how you answer the fourth – and then the test measures your nerve instead of your Japanese. The result comes at the end, or not at all. WHAT IT WILL NOT DO It will not give you a level from half a sheet. If too much ran out on the clock, it says so instead of guessing – a number from a thin measurement looks exactly like a number from a good one, and someone would build a study plan on it. It does not measure reading longer texts, and it does not measure listening comprehension. Both are missing, both are named on the opening screen, and neither is hidden in the small print. PITCH ACCENT IS MEASURED BY EAR One section plays you a recording of a single word and asks where the pitch drops. It does not ask for the name of the pattern, and it does not show you the word — because then your memory would answer, not your ear. The recordings are real, spoken by a voice actor with a Tokyo accent; speech synthesis does not render the melody convincingly, and a false sound in a diagnostic test is a false result. What gets measured is what you can actually hear in a bare word: whether the pitch drops, and after which mora. Two patterns that sound identical in isolation share one answer instead of being split into two — because the difference between them only falls on the next word, and there is no next word in the recording. BUILT ON THE FAMILY The questions come from the catalogs of the Japanese-learning apps, all by the same author: grammar constructions, vocabulary in context, particles, conjugation, counters, casual speech, honorific register and mimetics. That is why the diagnosis can point somewhere specific – and why it says which app fixes what. The first sheet is free.Read more

Shindan does not teach. It measures, and then it points. You already know the feeling: you study, you make progress, and you still cannot say what your weakest link is. A grammar app knows how much grammar you have. It does not know that particles are what keeps costing you points. Nothing you own sees the whole picture, because nothing you own asks about all of it at once. Shindan asks about all of it at once. One sheet, under a clock, and at the end two answers: where you stand, and the one thing worth studying next. TWO ANSWERS, NOT ONE Your level, on the JLPT scale, from the material that has a level: vocabulary, kanji readings and grammar constructions. Nobody asks "am I at practical level three" – they ask "will I pass N3", and that is the scale this answers in. Your weak spots, as a share of your mistakes: particles, verb forms, register, contractions, mimetics, counters and pitch accent. These have no JLPT level and never will – a particle you meet in week one and use for life. Measuring them on the same scale would make that scale mean nothing, so they are measured apart. NO PRACTICE, NO REPETITION, NO STREAK There is nothing to grind here and that is deliberate. A sheet measures once. Answers are not marked as you go, because seeing three mistakes in a row changes how you answer the fourth – and then the test measures your nerve instead of your Japanese. The result comes at the end, or not at all. WHAT IT WILL NOT DO It will not give you a level from half a sheet. If too much ran out on the clock, it says so instead of guessing – a number from a thin measurement looks exactly like a number from a good one, and someone would build a study plan on it. It does not measure reading longer texts, and it does not measure listening comprehension. Both are missing, both are named on the opening screen, and neither is hidden in the small print. PITCH ACCENT IS MEASURED BY EAR One section plays you a recording of a single word and asks where the pitch drops. It does not ask for the name of the pattern, and it does not show you the word — because then your memory would answer, not your ear. The recordings are real, spoken by a voice actor with a Tokyo accent; speech synthesis does not render the melody convincingly, and a false sound in a diagnostic test is a false result. What gets measured is what you can actually hear in a bare word: whether the pitch drops, and after which mora. Two patterns that sound identical in isolation share one answer instead of being split into two — because the difference between them only falls on the next word, and there is no next word in the recording. BUILT ON THE FAMILY The questions come from the catalogs of the Japanese-learning apps, all by the same author: grammar constructions, vocabulary in context, particles, conjugation, counters, casual speech, honorific register and mimetics. That is why the diagnosis can point somewhere specific – and why it says which app fixes what. The first sheet is free.