EROMANGA
HALL OF FAME

All rankings and inductions of the EROMANGA HALL OF FAME (EHOF) are determined entirely by the automatic tallying of public data. The tally consults nothing but the public record: no work and no author is ever added, removed, or moved in the ordering at anyone's discretion. The design of the measure itself — where the gate is set, how the bonus points are apportioned — was laid down by human hands; but it is published here in full as the standing rule, and every amendment to it is noted.

Article 1 (Scope and Materials)

This site treats three bodies of public record. The Doujin Halls (the Works Hall and the Circle Hall) cover works in manga format currently on sale on both stores, DLsite Doujin and FANZA Doujin (CG/illustration format is excluded, since it cannot be measured alike on both stores — the reasoning is set out in Articles 6 and 7; audio, game, video, and novel formats are likewise excluded). SPOTLIGHT alone, a daily feature, covers works in manga and CG/illustration format on sale on DLsite (doujin floor). The Commercial Halls (the Author Hall, the Works Hall, and Publishers) cover the public records of commercial tankobon volumes distributed digitally on DLsite, with FANZA BOOKS ratings used alongside them to compute the combined rating (Article 6). The Paper Era covers the public records of physical bookstore sales, including the years before digital distribution (Article 8). In every case, only public records existing in adult (R18) distribution and its prehistory are treated.

Article 2 (Data)

Sales counts, ratings, and chart positions use the figures published by DLsite, retrieved automatically every morning. The FANZA Doujin sales counts, average ratings, and rating counts used for the combined figures of doujin works (Article 6), and the FANZA BOOKS average ratings and rating counts used for the combined rating of commercial works (Article 6), are retrieved automatically every week from the figures FANZA publishes. Figures are as of the time of retrieval and may include delays. For the publication year of a tankobon, the bibliographic records of the National Diet Library Search and similar sources take precedence; where a work is not listed, its DLsite release year is used.

Article 3 (The Dendo Score)

This site's overall score is called the Dendo Score (dendo: "hall of fame"), displayed on the cards and individual pages of each hall. In every wing, 80 signifies the hall's gate; works and authors not inducted have no Dendo Score. It is not a 100-point scale — the summit exceeds 100. The Dendo Scores of the four wings — doujin works, commercial works, circles and authors — use no relative measure within the hall: each is fixed by an absolute formula that measures reach (sales or rating count) on a logarithmic scale (the revision of August 2026 abolished the former relative formula — a weighted composite of metrics normalized to 0–100 within the hall — and brought all four wings onto this form). The Paper Era (Article 8) alone normalizes a raw score proportional to sales onto 0–100 on a logarithmic scale and converts it as Dendo Score = 80 + 0.25 × raw score — the surviving records there are annual chart positions, from which reach cannot be measured in absolute terms. That conversion is monotonic and does not alter the ordering. Under an absolute formula the score does not move when a hall's membership changes, and the scale carries one meaning in every wing: 80 is the gate, 100 the line of Perpetual Honors (the multiple of reach at which 100 is touched is a calibration particular to each wing, stated in the article concerned). Perpetual Honors is not a selection: it attaches automatically the moment a score exceeds 100. Once attained, Perpetual Honors is never withdrawn — the date of attainment is entered in the ledger, and the title endures even if a later calibration of the gate moves the score. Where the level of a gate is itself amended by recalibration, that too is noted in the article concerned. The formulas are published in full in the articles below.

SPOTLIGHT automatically selects, each day and by pre-set thresholds, older works resurfacing on the charts (Revival), works with sharply rising sales (New Star), and highly rated works with modest sales (Hidden Gem). The thresholds are fixed in advance to exclude editorial discretion, and no human hand intervenes in the daily selection.

Article 4 (Exclusions)

Works withdrawn from sale, free works (previews, trial editions, and the like), and works designated by DLsite as outside its affiliate program are not included.

Article 5 (The Author Hall)

The Author Hall is selected on the foundation of the public records of commercial tankobon distributed digitally on DLsite. It is composed of authors who meet either of the following. (1) A reach R of 1.0 or higher — this is the gate. (2) Work-award points totalling 3.0 or more (induction by award, outside the gate; the same award points as Article 6, summed per author). In addition, the following structural conditions must all be met: five or more tankobon, ten or more years since debut, and a rating-count-weighted average of ★4.3 or higher. Names that are not real individuals (editorial departments, labels, fictitious bylines) are not eligible. Tankobon are counted on a first-edition basis — new editions, complete editions, and the like are treated as the same book as the first edition, and single-chapter releases, split volumes, bundled sets, and anthologies are not counted as tankobon.

Reach R is the author's total rating count across all tankobon divided by the gate level of 5,399, plus a converted measure of their chart record from the paper era. As of the calibration of 27 August 2026, the gate level of 5,399 was the 135th place when the 229 authors who met the structural conditions were ranked by total rating count — that is, the gate is set at "the reach of the 135th author of all time". The figures "229 authors" and "135th place" are the values on the day of that calibration; the size of that population shifts from day to day as the records are updated, while the gate level of 5,399 itself is held fixed until the next calibration. This level is recalibrated once a year (last calibrated 27 August 2026). The paper record uses the author's raw score from Article 8 (the paper pscore), counted toward reach as (paper pscore ÷ 3) × 0.75 — a calibration by which one Toranoana annual first place (3 points) is worth 0.75 times the gate in reach. This keeps authors of the paper era, whose digital record is thin, from being measured by the digital yardstick alone; conversely, paper alone cannot carry an author through the gate, since the structural conditions require five digital tankobon.

Ordering within the hall follows the Dendo Score. The gate (R = 1.0) is exactly 80; every tenfold increase in R adds 20; and R = 10 reaches 100. Authors above 100 hold Perpetual Honors — and that judgement is made on the Dendo Score as a whole, reach together with the bonuses for awards, career, volume and quality. Absent those bonuses, 100 is the line answering to ten times the gate in reach: the level of the 135th author of all time, multiplied tenfold. As in the other halls, no relative measure within the hall (percentile) is used — under a relative measure the same author's score shifts whenever the hall's membership changes, and would carry a different meaning from the scores shown in the Works Hall and the Circle Hall. The revision of August 2026 abolished the former relative formula entirely (four metrics converted to in-hall percentiles and weighted 50:20:20:10, with a capped 6-point honors bonus).

A (awards) sums the work-award points defined in Article 6 — from two prizes only, the reader vote (the Ero-Manga Awards) and the editorial prize (Kono Ero-Manga ga Sugoi!), scored by a table of rank and standing — across that author's works. A cap of 8 points is applied per work before the points are added to the author, and a further cap of 8 applies to the sum. Newcomer prizes do not count as award points — a newcomer prize is an expectation of what is to come, not a yardstick of a lifetime's reach (together with other magazine and publisher prizes, they are recorded on the cards as labels only). Awards are matched to authors by the author name written in the award record itself, rather than only through works with a digital edition, so that awards for works with no digital edition (print-only, out of print) are not lost. An author with 3.0 or more award points is inducted even if their reach falls short of the gate (induction by award). In that case the reach term has a floor of 0, so that a reach below the gate cannot work as a penalty.

The definition of debut — in this hall, an author's debut is the first time manga under that name reached the world in a commercial publication. The primary ground is first appearance in a commercial magazine (adult, general-interest, or web manga magazine alike), and a first contribution to a commercial anthology counts as well. Where the year of first appearance cannot be established, and where a commercial tankobon was published earlier than it, the year of the first commercial tankobon is the starting point. Work that is not manga does not count as a starting point — illustrations for novels, light-novel artwork, game character art, magazine cover illustrations, art books and the like are never counted as a debut here, however early the year. What is measured is not the year an artist began drawing, but the year they emerged as a manga artist. Dojinshi likewise do not count, being outside commercial publishing; where a dojinshi is later reprinted in a commercial venue, the starting point remains the first appearance of newly drawn work.

The starting point of C (career) is the debut year established by sourced research, as defined above. First appearance on DLsite runs several years — sometimes ten — later than the actual debut, and using it as the starting point would make the careers of authors who have been drawing since the paper era look shorter than they are. Debut years are researched individually against the Japanese Wikipedia, the National Diet Library search, mangaseek, publishers' official author pages and the like, with the grounds and the source recorded for each author. Where the research could not establish a year, the DLsite first-appearance year is used as a provisional value and recorded as such — no year is filled in by guesswork. The "career years" that C measures are the years elapsed from the debut year to the present (this year − debut year) — exactly the same number as the "N years active" shown on the author's page. As with V (volume), the number displayed and the number used in the calculation always agree. This is the number an author counts themselves when they say they are in their Nth year: the span of time they have stood in this trade. It does not stretch or shrink with the presence of a new volume, and once an author has debuted the count continues through any hiatus.
The structural condition "ten or more years since debut" is measured in the same elapsed years — every career figure in this hall, whether displayed, scored, or used for eligibility, is counted in elapsed years. The structural condition is measured from whichever is earlier, the researched year or the DLsite first-appearance year, so that a newly discovered later debut year cannot disqualify an author already in the hall.

V (volume) is measured logarithmically, so that the step from 5 to 10 books carries the same weight as the step from 10 to 20: drawing a great deal is credited, without turning the mass production of thin volumes into a structural advantage. Note that the count fed into V is exactly the same number as the "N books" shown on the author's page — the count from the complete timeline, which includes print-only volumes found by research and consolidates new editions, complete editions, split volumes and single-chapter releases; the structural condition "five or more tankobon" is judged on this same count. The number of works registered on the DLsite author page (a mechanical figure that includes split volumes and complete editions) is not used. The number displayed and the number used in the calculation always agree — a reader being able to check the Dendo Score for themselves ranks above convenience of tallying. Only for authors for whom no research record yet exists to build a timeline is the DLsite registration count used provisionally, to be replaced by this count as soon as the research is done. C and V together reach at most 5.5 points, a weight equal to about 1.9 times the reach (10^5.5/20) — that is what this cap means: drawing long and drawing much is worth, at the most, about 1.9 times the reach. In descending order, the weights that decide the ordering run reach > awards > volume + career > quality.

The Dendo Score is displayed to two decimal places, and the ordering is decided on the unrounded value. Authors whose displayed score is identical share a rank.

Note: this hall treats only records that exist as numbers. Authors who worked solely in print, and authors whose digital distribution is exclusive to other stores, therefore fall outside the current tally regardless of the scale of their achievements. Paper-era chart records are counted toward reach only for authors who have five or more digital tankobon. This is not a denial of those achievements but a limit of this site's range of measurement. Once an author's works enter digital distribution, they become eligible automatically, under the same criteria, from that point on. Where all of an author's tankobon are joint story-and-art bylines, the pair is honored as a single unit.

Article 6 (The Works Hall)

The Commercial Works Hall is formed, on a full census of commercial tankobon distributed digitally on DLsite, from works meeting either of the following: (1) 3,279 DLsite ratings or more and a combined average of ★4.3 or higher — this is the gate; or (2) award points totalling 2.0 or more (induction from outside the gate; next paragraphs). Whether the author is a hall author is immaterial. Only tankobon are eligible — single chapters and split volumes are identified mechanically by DLsite's official format classification and excluded (to avoid counting the same content twice). Magazine-format and multi-author joint works are honored separately in the Anthology division, under the same gate and the same formula. Translated editions are treated as the same work as the original and given no independent slot.

The level of the gate and of the whole scale rests on our own estimation — a conversion derived from several sets of real figures — that treats rating counts as a proxy for sales. The hall is thus selected on the basis of estimation, but per-work sales estimates, though computed, are not published. The combined rating is the average ★ of DLsite and FANZA BOOKS, weighted by each store's rating count (the same combining principle as the Doujin Works Hall). Be aware, however, that FANZA BOOKS carries only 1/100 to 1/300 as many ratings as DLsite, so the combined rating is in practice dominated by the DLsite value and almost never overturns the gate — we state this plainly. Where no FANZA BOOKS counterpart can be established, the DLsite value is used as the combined value.

Induction by award record corrects for the fact that reach (sales) depends on the size of the market in each era, which disadvantages older classics. The eligible awards are two: the reader-voted Ero-Manga Awards (2001–) and the third-party-judged Kono Ero-Manga ga Sugoi! (Weekly SPA!, 2015–2023). Award points are set mechanically by the published table below. Where one work holds several awards, the points are summed, capped at 8.0. Because the Ero-Manga Awards are reader-voted and the scale of each edition varies greatly, rank points are multiplied by a year coefficient of min(1, √(that year's total votes ÷ 300)). Kono Ero-Manga ga Sugoi! is a third-party selection with no notion of scale, so no year coefficient applies. Unranked "listed" entries earn no points.

Induction from outside the gate is granted only to tankobon whose award points total 2.0 or more (awards to single chapters or magazines do not qualify a work for induction). For such works the sales term 20 · log₄(R) is floored at 0 — the score is 80 + Q + award points, at most 91.0. This keeps a reach that has not crossed the gate from being carved in as a negative, and at the same time stops works rescued by an award from overtaking, without limit, works that crossed the gate on reach alone.

The hall waits for a reputation to mature; it does not measure a moment of sales. E-book sales accumulate over years after release, and readers' ratings gather over the same span. It is therefore by design that a freshly published work stands outside the gate — not an unfavourable judgement, simply not yet counted. Conversely, a work more than ten years old may still be gathering ratings and cross the gate today. The roster is recomputed automatically every day, and a work is admitted on the morning it crosses. Note too that the act of writing a review is itself the making of a record, and that record decides an induction years later — the effect a work had on readers' culture of reviewing is part of its achievement, as this hall sees it.

The Doujin Works Hall rests on one principle: the sales and ratings published by the digital manga market are combined from the outset, so that a work's demand is stated plainly as a single figure. That market at present comprises two stores, DLsite Doujin and FANZA Doujin; neither is taken as primary and the other as a supplement — both are tallied as equals. On a full census of manga works on the doujin floor — together with FANZA-exclusive works that hold no listing on DLsite — the hall is formed from works with an R-index of 1.0 or higher and a combined average of ★4.3 or higher (CG and illustration works are excluded). The R-index is the sum of each store's sales divided by that store's all-time top-50 level. The combined rating count is the sum of both stores' rating counts, and the combined average is the two stores' average ratings weighted by their rating counts.

Sales are not added as raw numbers but converted first to a value relative to each store's own level, because a unit of sale does not carry identical weight across the two stores — FANZA discounts more aggressively, so one sale there does not mean quite what one sale means on DLsite. Adding raw figures would uniformly favour works that ran up their numbers on the more heavily discounted store. Fixing each store's all-time top-50 level — the summit within that store — as 1, and measuring how near a work comes to it, yields a figure that carries the same meaning across the difference in the stores' characters. Both stores are looked at nonetheless, because attending to a single store drops from the record everything a work was read for on the other. R = 1 is a height that can be cleared by reaching the all-time top-50 level on either store alone, or by the two stores' reach taken together.

A cross-store correspondence is accepted where all three tests pass — agreement of circle, normalized similarity of title (robust to masked characters), and a hash comparison of the cover image. Pairs on which the machine comparison fails are checked one by one by eye, with both stores' covers set side by side, and accepted only where they are confirmed to be the same work — cases genuinely exist in which one work carries a separately drawn cover on each store, and an image hash alone would let them slip (correspondences accepted by eye are recorded so as to remain distinguishable from those confirmed mechanically). Alternate editions (full-color versions, recolored versions and the like) are treated as separate works: the standard edition is matched to the standard edition, the full-color version to the full-color version. Where neither procedure finds a correspondence, the other store's figure is treated as zero. FANZA-exclusive works with no DLsite listing are selected on their FANZA sales alone and always link to FANZA (they are not folded into the Circle Hall or the Author Hall).

Ordering within the hall follows the Dendo Score. For doujin works, that score measures the R-index on a logarithmic scale and adds a quality bonus drawn from the combined average. The gate — R = 1.0 — is exactly 80; every fivefold increase in R adds 20; and R = 5 reaches 100. Unlike the other halls, no relative measure within the hall (percentile or min–max normalization) is used: under a relative measure the same work's score shifts whenever the hall's membership changes, and the same work would carry different scores in a division and in the Doujin Works Hall (the Commercial Works Hall, the Circle Hall and the Author Hall were all brought onto the same absolute formula in the revision of August 2026). An absolute formula, fixed by reach and average alone, leaves the score untouched by the hall's membership and gives the scale one meaning in every wing: 80 the gate, 100 the line of Perpetual Honors. Rating count plays no part in the ordering — it depends on the size of a store's user base and on the era, and counting it would count the same reach, sales, a second time (it remains in the gate condition of a ★4.3 combined average, and in what the cards display). The Dendo Score is displayed to two decimal places, and the ordering is decided on the unrounded value; works whose displayed score is identical share a rank. Since the summit rises above 100, works whose Dendo Score exceeds 100 hold Perpetual Honors — and that judgement is made on the Dendo Score as a whole, reach together with the quality bonus. Absent that bonus, 100 is the line answering to five times the gate in reach: five times beyond the level of the gate itself, the all-time top 50 of both stores. That is where the line of Perpetual Honors is drawn.

Both Works Halls carry a "Search by tag" feature that filters inducted works by DLsite official genre tags. Tags serve display and search only; they play no part in selection or ordering.

Article 7 (The Circle Hall)

The Circle Hall is a hall of manga circles. It is selected on a full census of both stores — DLsite Doujin and FANZA Doujin — and covers manga works only. CG collections are left out of the tally because FANZA Doujin's "comic" category does not contain them, so the two stores cannot measure them alike — were CG collections included, a circle working chiefly in CG would carry its whole DLsite figure with nothing at all from FANZA. The Works Hall (Article 6) likewise measures manga alone, and this hall is brought into line with it. This hall therefore does not measure what a circle working chiefly in CG collections has reached — that no means of measuring it exists is stated here plainly, as the reason it was not attempted.

The criteria are that all of the following be met: ten or more years of activity, five or more manga works (across both stores), and a circle R-index of 1.0 or above. Years of activity are measured as the span of presence on the store, and for that span alone CG collections are also counted (so as not to lose a veteran who began in CG and moved to manga). Translated editions and translation labels are excluded from the tally.

The circle R-index follows the same reasoning as the work-level R-index of Article 6. Raw figures are not added: each store's cumulative circle sales are first taken relative to that store's own level, and only then summed. The level of the gate is the cumulative sales of the all-time No.100 circle on each store — DLsite counts 19,142 manga circles and FANZA 17,239 comic circles, so nearly the same population, and cutting at the same rank is what makes the two of equal standing. The resulting ratio (FANZA ÷ DLsite ≒ 7.4) stays between 6.9 and 7.4 whether the No.100 mark is moved to No.50 or No.150, so the calibration is robust to where the gate is placed. The levels are recalibrated once a year.

Ordering follows the Dendo Score, on the same absolute formula and the same slope as the Doujin and Commercial Works Halls; no relative measure within the hall (percentile or min–max normalization) is used, since under a relative measure the same circle's score would shift whenever the hall's membership changed. Prolificacy as such earns no points — quantity shows in the cumulative sales.

Quality (★) is not made a gate. At the level of a single work, "no work below ★4.3" is coherent; drawn across a circle, the same line becomes a matter of batting average — the more prolific and long-standing the circle, the lower its mean — and would function as a clause shutting out the prolific veterans. ★ enters the ordering only as the quality bonus above (0–3). FANZA carries no ratings, so for circles present on DLsite the weighted average ★ is taken from DLsite ratings.

Identifying a circle across the two stores rests on agreement of work titles, not of names. Two or more works whose normalized titles agree on both stores, or one or more work correspondences under Article 6, or a single title match together with either a name match or a long title (16 characters or more) — any of these establishes the same circle. Where a runner-up stands close, no correspondence is made and the case is held; a circle with no correspondence has the other store's figure treated as zero (erring against overstatement). Separate FANZA IDs carrying the same name are merged as one circle.

FANZA-exclusive circles with no DLsite presence are selected on the same R-index and the same Dendo Score. Since the census carries no release dates, their years of activity are obtained from the release date on each work's detail page, and the same structural conditions are imposed as on circles present on DLsite (ten or more years, five or more works) — it would not be a standing of equals if only one store's members were bound by the span of years. Their quality bonus uses FANZA ratings, their works are identified by cid, and their links always go to FANZA.

Works whose sales figure is not shown in the FANZA Doujin listing (about 8% of the whole) are treated as zero, so the FANZA totals are slightly understated. Among the leading circles the gaps run to a few works at most, too small to move the ordering, but the fact is recorded here.

Article 8 (The Paper Era)

The halls of the Paper Era rest on the public records of physical bookstore sales, including the years before digital distribution (the Toranoana in-store charts, reconstructed from records published on Akiba Blog). Every chart appearance of a work is collected, but only the annual charts are used for scoring (combining half-year and annual charts would count the same sales twice; for years in which no annual chart was held, the first-half chart is counted as an annual equivalent at a factor of 0.5). Ordering follows the overall score — medal points for each year's top three (3, 2, and 1 points), chart-in points for fourth place and below (3 ÷ rank points, following the Zipf-like law by which in-store sales decay roughly in inverse proportion to rank), plus a bonus (up to 3 points) according to a work's position in Toranoana's official all-time sales best 100 (2009–2022). Because the raw score is a quantity proportional to sales, its conversion into the Dendo Score uses a logarithmic scale, just as sales figures do in the other halls. Because the scale of store sales varied greatly across the years, annual rank alone cannot compare across eras; the all-time cumulative sales record supplies a correction of absolute scale. Note that the tally window of the all-time best 100 begins in 2009, so works from 2008 and earlier benefit little from this correction (a limit of the range of measurement). Authors are scored by the sum of the raw scores of their charting works, converted on the same logarithmic scale. Records of reader votes and magazine-sponsored prizes are not used to judge eligibility for the Paper Era halls (their treatment follows Article 6 for the Commercial Works Hall and Article 5 for the Author Hall). An author's raw score here (the paper pscore) is converted and counted toward the reach R of Article 5 — so that achievements recorded only in the paper era are not measured by the digital yardstick alone. The FANZA mail-order annual charts (2006–2022) are collected in the Yearbook as reference material only: the number of entries in their past-year tables shrinks sharply the further back the years go, while the rank numbers run without gaps, so agreement with the original tables cannot be guaranteed — and they therefore play no part in any tally.

Article 9 (Derived Statistics and Annotations)

Derived statistics such as the publisher rankings (commercial and Paper Era) are mechanical tallies of hall data and involve no independent selection judgment. Label names are consolidated under their publisher. The biographies and chronologies on individual author pages rest on public materials — National Diet Library bibliographies, official publisher sources, and the like — and record only adult commercial activity.

The following annotations are records of fact and play no part whatsoever in selection or ordering. VOICES (a record of reception) quotes remarks by peers, critics, and contemporary reviewers, with sources and archive snapshots stated. Own-circle notes record the identity between a commercial hall author and a doujin circle, verified against DLsite official registration data and similar sources. Award labels on work cards are likewise annotations of record (except as provided in Article 6).

Supplementary provision: revisions of these criteria are announced by updating this page. Established 2026.

Supplementary provision (12 September 2026): the Anthology division provided for in Article 6 is abolished as of 12 September 2026. Joint works by multiple authors and magazine-format works are no longer eligible for the Works Hall. The record is kept — the standing of the division as of its abolition is preserved, and the division may be re-established after it has been redesigned. The text of Article 6 is left standing as the record of what was revised.