Read Your Level

How the easiness score is calculated

The easiness algorithm has recently been updated, but this article does not yet reflect the updates. An update will be coming soon!

Easiness can be thought of as the absence of conscious effort. Read Your Level is based on this philosophy - our easiness score aims to rank books by the number of mendokusai moments which students of a given level may encounter while reading. More mendokusai moments correspond to more conscious effort, and thus less easiness.

The current incarnation of our easiness score is based purely on those mendokusai moments caused by unknown kanji. There are plans to improve the algorithm to take sentence length and vocabulary into account.

Kanji frequency

When a kanji is encountered which is frequently used within the same book and many other books, there is a high probability that the reader will be comfortable with it due to having seen it a number of times before. Conversely, when the reader encounters a kanji which is rarely used within the same book and other books, there is a safe bet that the reader will need to look it up. Looking up kanji is an absolute pain in the arse.

Read Your Level counts occurrences of each of the Joyo Kanji across all of the books in its database, and then assigns an initial "painfulness" score to each based on its rank. The painfulness scores look something like the following:

Painfulness graph

Of course, if a kanji on the right side of this graph is encountered 10 times in the first 6000 characters, it is likely that the 10th encounter will be less painful than the first, particularly if the kanji are spaced at increasing intervals. As such, I decrease the painfulness for each subsequent occurrence of a kanji.

Notable behaviors

Ranking easiness this way causes some unexpected behaviors. One example is that self-help books and get rich quick schemes are often ranked as easier than light novels. While this is surprising at first, an inspection of the kanji involved shows that the first category has significant repetition of a number of the more difficult kanji, while light novels generally use a more varied selection of rare kanji.

"Easiness" is being constantly improved

While the current easiness score is not perfect, it is still a useful tool for finding material which matches your level. Any feedback on scores which you feel are incorrect would be very much appreciated, as I aim to constantly improve the easiness score to make it as accurate and useful as possible. Improved easiness scores will be initially announced to the newsletter, so make sure to sign up so you don't miss out!

Make sure you don’t miss out on improved easiness scores! Drop your e-mail in the box.