About 86% of Japanese manga series published between 2010 and 2023 have never received an official English release. That figure comes from a random sample of 361 non-adult series drawn from the MangaUpdates database, with a 95% confidence interval of 82.5% to 89.6%.
Put the other way round: for roughly six series out of seven, there is no English edition to buy, and there is not going to be one.
We looked for this number before measuring it. It does not exist. Searching for it returns lists of "manga you cannot read in English" and a lot of people asserting "most of it" without a source. So this page is the measurement, the method, and the data, in that order.
The number, by year
Each row is a random draw from every series MangaUpdates records for that publication year. Adult titles are excluded here and reported separately below.
| Publication year | Series sampled | Never licensed in English |
|---|---|---|
| 2010 | 28 | 93% |
| 2011 | 22 | 73% |
| 2012 | 28 | 82% |
| 2013 | 31 | 81% |
| 2014 | 30 | 90% |
| 2015 | 20 | 95% |
| 2016 | 28 | 86% |
| 2017 | 24 | 92% |
| 2018 | 25 | 88% |
| 2019 | 24 | 83% |
| 2020 | 24 | 79% |
| 2021 | 23 | 78% |
| 2022 | 23 | 96% |
| 2023 | 31 | 94% |
Two things stand out.
There is no trend. The manga market has grown for most of this period, English publishers have expanded their catalogues, and simulpub has become normal. None of that shows up here. A series published in 2021 is about as likely to be skipped as one published in 2012. Growth in licensing has tracked growth in publication, not outpaced it.
The last two years read high, and you should discount them. A 2023 series can still be licensed in 2027. Treat 96% and 94% as ceilings that will drift down, and treat the older rows as the settled picture.
Adult titles, counted separately
We expected adult manga to be the reason the headline number is high, so we split it out. It turned out to be the opposite of what we assumed.
| Group | Series sampled | Never licensed in English |
|---|---|---|
| All sampled manga | 560 | 84.6% |
| Excluding adult titles | 361 | 86.4% |
| Adult titles only | 199 | 81.4% |
Adult titles are licensed slightly more often than everything else, not less. Removing them pushes the figure up rather than down. We have no strong explanation for this and are not going to invent one: a dedicated English-language adult manga publishing scene exists, which plausibly covers more of that catalogue than mainstream publishers cover theirs.
The honest headline is therefore the higher one, 86%, not the lower.
How this was measured
The point of publishing a number is that someone else can check it, so here is everything needed to do that.
Source. The MangaUpdates public API. Its per-series licensed field is the database's "Licensed (in English)" flag, maintained by contributors since 1998. It is the closest thing to a machine readable census of English manga licensing that exists.
Scope. Series of type Manga, which is the database's Japanese-origin category. Manhwa (Korean), manhua (Chinese), novels and doujinshi are separate types and were excluded. Publication years 2010 to 2023.
Sampling. Stratified by year, then a simple random sample inside each year. This matters: a flat query against the API is capped at 10000 results and ordered by relevance, which would have loaded the sample with famous series and undercounted the long tail, biasing the answer downward. Year by year, every stratum sits under the cap and returns alphabetically, so a uniform draw inside a year is a fair draw. Forty positions were drawn per year and resolved to their exact page and offset.
Statistics. Wilson score intervals, which stay correct near 0 and 1 where the normal approximation does not.
Reproducing it. The sampling script is in the site's repository as scripts/measure_licensing.py, with a fixed random seed, so rerunning it draws the same 560 series. The raw sample, one row per series, is at data/licensing-sample.csv.
What this measurement cannot tell you
Three limits worth stating plainly, because a number without its limits is not worth citing.
MangaUpdates is community maintained, not official. A licence announced last month may not be recorded yet. That pushes the measured figure slightly high.
"Licensed" is not the same as "available". Some series counted as licensed here went out of print years ago, or were cancelled after two volumes. The share of series a reader can actually buy in English today is lower than 14%, not higher. This measurement is generous to the licensed side.
A sample is a sample. 361 non-adult series is enough to say "about 86%" with confidence, and not enough to say "86.4% exactly". The interval, 82.5% to 89.6%, is the honest reading.
Why this matters if you read manga
The practical consequence is simple. If you follow anything outside the top of the charts, the odds are strongly against an English edition ever arriving. Waiting is not a strategy, because for most series there is nothing to wait for.
That leaves reading the Japanese, which is what reading raw manga means in practice. It is also why translating a page well is a different problem from translating a sentence: the text is vertical, hand lettered, and shaped to fit a bubble. We wrote separately about how to translate manga, how the available tools compare, and how this figure compares to scanlation, the other way fans have historically filled this exact gap.
Citing this
If you use these figures, please link back to this page. The full method is above and the raw sample is in the repository, so the number is checkable rather than something you have to take our word for.
Measured 7 August 2026. We plan to rerun the same sample periodically and will note any change here rather than quietly editing the figures.