Stats

Click on Old Stats if you want to see the old stats, like the Core 6k or Hayashi stats.

TODO: Fix scripts with broken linewraps (affects sentence count, line count, and metrics a/d).

All stats work on deduplicated copies of the script.

If you want to add a script, post it to the Requests page.

See also Changes.

You can get the new analysis tools here. I hope you know how to use rust!

Higher is easier: metric a/b/c/d

Lower is easier: everything else

The "freqlistX% Target" metrics are an estimate of lexical complexity. High measurements mean that the script uses uncommon words more frequently, or common words less frequently. Commonness is based on every long-enough script combined, not on general purpose frequency lists or the frequency list of the single script in question. However, the most common uncovered 20 words in the script are ignored, because they're probably names or setting jargon, or even parsing errors. This entire process ignores grammatical words, proper nouns, and numbers.

The "metric a/b/c/d" lines are derived from linear regression targeting the "freqlist 92.5% Target" metric. An exact description of the processing for these metrics is here. The math is reverse polish notation.

The hours estimate is kanji/18000 + hiragana/31500 + katakana/61400. It assumes a reading speed of 475 characters per minute, in general. It's calibrated based on a bunch of vndb length ratings with "very short" assigned to 1 hour and "very long" assigned to 60 hours. It is completely useless for VNs with nontrivial amounts of gameplay, or where rereading is expected, or where you're expected to miss a lot of content, or where you're only meant to read one or two out of several routes. Example of calibrating with only character count as a source variable: https://i.imgur.com/tqTGVlB.png

Frequency lists, raw stats database: https://github.com/wareya/jpstats/tree/master/workspace

There is a graph at the bottom of the page.

The character counting stats here ignore some punctuation characters. This is because some script ripping methods strip some types of punctuation, such as quotes around lines of dialogue. For example, without this adjustment, hoshimemo would look smaller compared to other VNs than it actually is. If I ever have a chance to revisit old dumps and make sure they all handle punctuation consistently, I will "fix" this, because it's a source of confusion when comparing to the old tlwiki stats. It also does not count newlines, but this won't ever change.

target 90%target 92.5% (default)target 95%
metric a (default)metric dmetric cmetric b
toggle unselected names