Counted out of the operators’ own clauses
Sites like BitStarz, ranked on what each one publishes
BitStarz names 74 countries in its own restricted-territory clause. Every alternative here is scored on how much of that list it repeats, and the denominator is printed beside the numerator. Of 50 rows, 29 were measured and 21 carry nothing to compare.

Casinos like BitStarz: what the phrase has to answer
Somebody searching for sites like BitStarz is asking five separate things at once, and a shortlist of ten brand names answers none of them. Which operators run under the same constraints. Which of the things BitStarz is liked for turn up anywhere else. Whose payout clauses bite first. Who would refuse the reader outright. And why the order of the list should be believed at all. Each of those five gets a section below, and each section stops where the evidence stops.
BitStarz names 74 countries in its own restricted-territory clause, read from bitstarz.com/terms on 26 August 2026. That list of 74 is the ruler used here, and it is opened up name by name on the denominator itself. Fifty operators went through the same seven checks, drawn from a pool of 78 records that opened far enough to be scored at all. Twenty-two of the 50 were read from the live page and 28 from an archived copy, and every figure below carries the date it was read.
Seven checks, one hundred points, and fifteen of them unclaimed
The measure sums seven properties, each of them something an operator either publishes or withholds. Weights follow how hard a check is to fake and how many documents have to be opened to settle it.
| check | weight | what scores the points |
|---|---|---|
| restricted-country overlap | 30 | how much of the BitStarz list of 74 this operator repeats |
| same licensing authority | 15 | the authority named on the permit matches, string for string |
| same licence holder | 15 | the company named on the permit is Gareton B.V., which makes it a sister brand |
| runs its own original games | 10 | the operator publishes games built for itself |
| has a VIP programme | 10 | a published tier or loyalty scheme |
| casino only, sportsbook left out | 10 | BitStarz is a casino, so a sportsbook makes an operator less alike |
| same verification posture | 10 | verification described in the terms with the amount left out, as at BitStarz |
The third row is the interesting one. Gareton B.V. is the company on the BitStarz permit, and across all 50 rows that company appears zero times. Those 15 points therefore went unclaimed by everybody, which puts the reachable ceiling at 85 rather than 100. Mirax leading on 73 means 73 of a possible 85, and reading it as 73 per cent would overstate the gap.
Two checks read a published tag rather than a clause, and they are the softest cells here: an operator with no tag for original games scores zero on that check, and this page treats that zero as a reading of a shortlist rather than as proof the operator has none. The full account of that weakness sits on the measure itself.
How to recompute the leading check
Overlap points are 30 multiplied by the matched countries, divided by 74, rounded to the nearest whole point. Three worked examples, each checkable against the operator’s own terms page:
- Oshi Casino repeats 72 of the 74, giving 30 × 72 ÷ 74 = 29.19, printed as 29.
- Mirax repeats 70, giving 28.38, printed as 28.
- Bets.io repeats 38, giving 15.40, printed as 15.
Direction matters more than the arithmetic. The question asked is how much of the BitStarz list an operator repeats, and the answer stays silent about everything else in that operator’s own clause. Oshi Casino excludes 224 countries; 72 of them are BitStarz entries and 152 are extra. A reader who takes 72 of 74 to mean the two lists agree has read the fraction backwards.
The ten closest sites like BitStarz
| # | operator | points | overlap of the 74 | company named on the permit | authority as the operator names it |
|---|---|---|---|---|---|
| 1 | Mirax | 73 | 70 of 74 | Scores55 Tech B.V. | Curaçao Gaming Authority |
| 2 | Metaspins | 68 | 57 of 74 | 3–102–959384 SRL | Curaçao Gaming Authority |
| 3 | Bets.io | 60 | 38 of 74 | Techno Wave Systems S.R.L. | Curaçao Gaming Authority |
| 4 | Bitdreams | 59 | 71 of 74 | Hollycorn N.V. | Antillephone N.V. |
| 5 | Oshi Casino | 59 | 72 of 74 | Novatrix SRL | Anjouan Gaming Board |
| 6 | PlayAmo | 59 | 72 of 74 | Novatrix SRL | Anjouan Gaming Board |
| 7 | Slotum | 58 | 69 of 74 | Novatrix SRL | Anjouan Gaming Board |
| 8 | Wild.io | 58 | 31 of 74 | Stack Gaming Ltd. | Curaçao Gaming Authority |
| 9 | 7Bit Casino | 54 | 71 of 74 | unread | Curaçao Gaming Authority |
| 10 | CryptoLeo | 54 | 48 of 74 | Uno Digital Media B.V. | Curaçao Gaming Authority |
Mirax sits first while ranking fourth on the heaviest check, and the four smaller checks explain it: 28 for overlap, then 15 for naming the Curaçao Gaming Authority exactly as BitStarz does, 10 for a VIP scheme, 10 for leaving sports out, 10 for describing verification while publishing no amount. Forty-five points came from the light end of the formula.
Rows four to seven show the reverse. Oshi Casino and PlayAmo repeat more of the territory list than anything else measured, collect the highest overlap score available at 29, and then forfeit 15 because their permit names the Anjouan Gaming Board. Bitdreams forfeits the same 15 for a subtler reason: it names Antillephone N.V., the older Curaçao master-permit holder, which is a different string from the one BitStarz prints even though the island is the same. That is a real property of the measure, and it is disclosed rather than smoothed over.
What BitStarz is liked for, item by item
Resemblance on paper is one question. Whether the specific things people go to BitStarz for turn up at the alternatives is another, and it splits into five.
The catalogue comes first. The studio list published by BitStarz ran to 62 names when it was counted on 26 August 2026. Sixteen of the 50 rows publish a studio list that could be counted, headed by Instant Casino at 150 names, Betriot at 140 and Bitcasino.io at 120, with Fairspin at 114 and Sportsbet.io at 110 behind them; CryptoLeo sits at the other end with 13. The other 34 went uncounted on the reading date, and their cells stay blank. A studio count is a count of names an operator prints, and operators group and rebrand studios differently, so the column ranks disclosure rather than depth.
Original games are the second. BitStarz publishes games built for itself, and 16 of the 50 carry the same published tag. That check reads a shortlist rather than a clause, which makes its zeros the weakest figures on the page, and the reason is set out on the seven checks in full.
Paying out in coin is the third, and it is where our own record is thinnest. The BitStarz coin list stayed empty when the record was assembled, so this page carries no figure for which currencies BitStarz pays in, and that blank marks a gap on our side instead of a property of the operator. Twenty of the 50 do publish a countable list: Cloudbet names 40 currencies, CryptoGames and Rakebit 15 apiece, Wolf.bet and Rocketpot 13, Bets.io and Wild.io 12, Oshi Casino and Wild Fortune 10. At the short end, KatsuBet names one and Winz.io two, and the other 30 cells are blank.
Speed is the fourth, and the honest answer is the shortest. Across the 50 records a payout processing window with a clause behind it was recorded exactly once, at Vave, whose clause 8.1 gives three working days and was read on 25 August 2026. Every other row, BitStarz among them, is blank on that field. Payout speed is the property this measure handles worst, and saying so costs less than inventing a figure for it.
Loyalty is the fifth and the most widely published: 28 of the 50 print a tier or loyalty scheme on the pages read, BitStarz included, and 22 carry no such tag. That check shares the soft-cell weakness above, so a missing tier here means a tier was absent from the page read rather than absent from the business.
What 72 of 74 buys, and what it withholds
Four of the operators near the top of the overlap column publish territory clauses three times longer than the BitStarz one, so the overlap says their lists swallow the BitStarz list rather than match it.
- Oshi Casino: 224 entries of its own, 72 inside the BitStarz 74. The pair left out are Guadeloupe and Russia.
- PlayAmo: 217 entries, 72 inside. The pair left out are Russia and Slovakia.
- Bitkingz: 218 entries, 71 inside. Brazil, Poland and Russia are absent from it.
- Slotum: 216 entries, 69 inside, with five BitStarz entries absent.
The pair sharing first place on that column get a page each, one for the highest count and its two exceptions and one for the same count with a different pair missing.
Two rows run the other way and are more informative for it. Metaspins publishes 58 entries, of which 57 sit inside the BitStarz list; the single outsider is Malta. Its clause is close to a subset of the BitStarz one, yet the score reads 57 of 74 and pays 23 points, because the denominator is fixed at 74 for everybody. CryptoLeo publishes a clause of exactly 74 entries that agrees with the BitStarz 74 on 48 of them: two lists of identical length, agreeing on roughly two thirds.
One permit number, several brands
The check worth 15 points asks whether the company on the permit is the company on the BitStarz permit, and the answer across 50 rows is a flat zero. What the reading did turn up is clustering among the operators themselves, which is a different fact and a useful one.
Five rows name Novatrix SRL: Oshi Casino, PlayAmo, Slotum and Bitkingz, plus Winz.io further down at 41 points. Four of the five print the identical Anjouan permit number, ALSI-202508056-FI2, read on 26 August 2026. Their territory clauses behave accordingly, running to 216, 217, 218 and 224 entries with 69 to 72 of the BitStarz 74 inside each.
A second cluster is stranger and stays open. Mirax prints Curaçao Gaming Authority number OGL/2024/1307/0748 and names Scores55 Tech B.V.; 7Bit Casino prints the same permit number and names a company we were unable to read. Two brands, one number, and one blank where the counterparty should be. The pairing is recorded as read on 26 August 2026 and left unresolved, because a number appearing twice supports several explanations and our reading picks between none of them. The blank fields behind that row are counted on a strong count beside an unread record.
For a player the consequence of a shared back office lands on the account rather than the shopfront, which is why the company field carries the weight of the authority field despite paying out nothing here.
| Casino | Operator | Licence | Countries barred | Withdrawal limit | Visit |
|---|---|---|---|---|---|
| Vavepaid placement | TechOptions Group B.V. | Curaçao no number | list not read | from 50,000 USDT in instalments cl. 8.8 | Visit |
| BitStarz | Gareton B.V. | Curaçao Gaming Authority | 74 | not published | |
| Oshi Casino | Novatrix SRL | Anjouan Gaming Board | 224 | not published | |
| PlayAmo | Novatrix SRL | Anjouan Gaming Board | 217 | not published | |
| Bitdreams | Hollycorn N.V. | Antillephone N.V. | 216 | not published | |
| 7Bit Casino | not published | Curaçao Gaming Authority | 216 | not published | |
| Bitkingz | Novatrix SRL | Anjouan Gaming Board | 218 | not published | |
| Mirax | Scores55 Tech B.V. | Curaçao Gaming Authority | 220 | not published | |
| Slotum | Novatrix SRL | Anjouan Gaming Board | 216 | not published | |
| Metaspins | 3-102-959384 SRL | Curaçao Gaming Authority | 58 | not published | |
| CryptoLeo | Uno Digital Media B.V. | Curaçao Gaming Authority | 74 | not published | |
| Bets.io | Techno Wave Systems S.R.L. | Curaçao Gaming Authority | 42 | not published |
Twelve operators were read clause by clause for that table, and the fields fill unevenly on purpose: 11 of the 12 print a permit number and 11 publish a territory list. Where a cell is blank, the document stayed shut on the day it was opened.
Whose payout terms bite first
Thirteen of the 50 attach a payout ceiling to a clause number, each of the thirteen read on 26 August 2026.
| operator | ceiling | window | clause |
|---|---|---|---|
| Oshi Casino | 4,000 EUR | a day | 12 |
| Instant Casino | 4,000 EUR | a day | 8.1 |
| Rocketpot | 5,000 USD | a month | 11.5 |
| PlayAmo | 7,500 EUR | a week | 12 |
| Mirax | 10,000 EUR | a week | 8 |
| Bets.io | 10,000 USD | a transaction | 10 |
| Slotum | 15,000 EUR/USD | a week | 13 |
| Wild Fortune | 15,000 EUR | a month | 11.8 |
| Bitdreams | 20,000 EUR | a month | section 11 |
| CryptoLeo | 50,000 USDT | a month | 7.9 |
| Wild.io | 100,000 USD | a week | 9.6 |
| CoinCasino | 500,000 EUR | a month | 9.6 |
| Bitcasino.io | 1,000,000 USDT | a week | 6.10 |
Eight of the thirteen print a figure of 15,000 or under, three print 100,000 or more, and two sit between. Reading that column as one ladder would mislead, because four windows are in play and a daily ceiling constrains the opposite habit to a per-transaction one. Which of them bites depends on how a player cashes out, and the reconciliation, together with the twelve-row detail, sits on the published payout ceilings.
BitStarz sits outside the thirteen. Its terms were read on 26 August 2026 and no ceiling was recorded from them, so the cell stays empty, which on this site means the document was silent on that point at the time of reading. It never becomes a claim that payouts run uncapped.
The identity check is the other half of a cash-out, and here BitStarz has plenty of company. Its terms describe verification and print no amount, and 30 of the 50 describe it the same way, which is exactly what the seventh check pays for. Four rows publish an amount instead: 2,500 EUR at Bitcasino.io, Sportsbet.io and Livecasino.io, each in clause 6.6, and 2,500 USD at Rocketpot in clause 11.4. Sixteen rows leave the field blank, and a published amount is neither a stricter nor a looser posture on its own; it is simply a more legible one.
Who would turn a reader away before the deposit
Twenty-nine of the 50 publish a territory clause that could be read, and crossing those 29 against each other answers the question a brand shortlist tends to skip: whether the reader is wanted at all.
The United States is named in 26 of the 29. The United Kingdom appears in 23, the Netherlands in 22, Spain in 20, France in 19, Lithuania in 18, Ukraine in 17, and Belgium and Israel in 15 apiece. A reader sitting in any of those places is looking at a shortlist where most readable rows have already declined them in writing, which reorders this page for that reader.
Divergence lives at the other end of the tally. Brazil sits inside the BitStarz 74 and appears in only 3 of the 29 other clauses; Liechtenstein appears in 7 and Hungary in 8. Those are the names where operators genuinely differ, and they are where a point or two of overlap changes hands.
Three of the 29 clauses name a single country each: Bitsler names American Samoa, Bitcasino.io names Japan, and Stake names the United States. A clause that short settles who it names and leaves everybody else unaddressed in that document, which is a weaker thing than a welcome. How the 74 group by region, and which entries every measured list repeats, is worked through on the denominator page.
The 21 rows with nothing to compare stay silent on this question. Their territory clause stayed shut on the reading date, so a reader learns where 29 of the 50 stand from this page and learns where the other 21 stand by opening that operator’s own terms.
The fifty, in order
Each row prints the seven cells it was scored on, so any total here can be added up by hand. A shorter cut of the same ranking, ten rows deep, sits on the ten-row version of this table. Where a cell reads unread, the check met a document that stayed shut rather than an answer in the operator’s favour.
Row 13 is a compact worked example. Instant Casino repeats 33 of the 74, so the heaviest check pays 30 × 33 ÷ 74 = 13.38, printed as 13. It names the Curaçao Gaming Authority, which pays 15. Its company, Igloo Ventures SRL, differs from Gareton B.V. and pays zero, as does the absence of original games and of a published VIP tier. Sports betting is left out, worth 10, and verification is described with the amount withheld, worth another 10. The seven cells sum to 48, and the table prints 48.
| # | operator | overlap of 74 | overlap /30 | authority /15 | holder /15 | originals /10 | VIP /10 | casino only /10 | posture /10 | points |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Mirax | 70 of 74 | 28 | 15 | 0 | 0 | 10 | 10 | 10 | 73 |
| 2 | Metaspins | 57 of 74 | 23 | 15 | 0 | 10 | 10 | 0 | 10 | 68 |
| 3 | Bets.io | 38 of 74 | 15 | 15 | 0 | 0 | 10 | 10 | 10 | 60 |
| 4 | Bitdreams | 71 of 74 | 29 | 0 | 0 | 0 | 10 | 10 | 10 | 59 |
| 5 | Oshi Casino | 72 of 74 | 29 | 0 | 0 | 0 | 10 | 10 | 10 | 59 |
| 6 | PlayAmo | 72 of 74 | 29 | 0 | 0 | 0 | 10 | 10 | 10 | 59 |
| 7 | Slotum | 69 of 74 | 28 | 0 | 0 | 0 | 10 | 10 | 10 | 58 |
| 8 | Wild.io | 31 of 74 | 13 | 15 | 0 | 10 | 10 | 0 | 10 | 58 |
| 9 | 7Bit Casino | 71 of 74 | 29 | 15 | unread | 0 | 0 | 10 | unread | 54 |
| 10 | CryptoLeo | 48 of 74 | 19 | 15 | 0 | 0 | 10 | 0 | 10 | 54 |
| 11 | Bitkingz | 71 of 74 | 29 | 0 | 0 | 0 | 0 | 10 | 10 | 49 |
| 12 | Shuffle | 11 of 74 | 4 | 15 | 0 | 10 | 10 | 0 | 10 | 49 |
| 13 | Instant Casino | 33 of 74 | 13 | 15 | 0 | 0 | 0 | 10 | 10 | 48 |
| 14 | Playbet.io | 26 of 74 | 11 | 15 | 0 | 0 | 10 | 0 | 10 | 46 |
| 15 | Bitsler | 0 of 74 | 0 | 15 | 0 | 10 | 10 | 0 | 10 | 45 |
| 16 | CryptoGames | unread | unread | 15 | 0 | 0 | 10 | 10 | 10 | 45 |
| 17 | KatsuBet | unread | unread | 15 | 0 | 0 | 10 | 10 | 10 | 45 |
| 18 | Thunderpick | unread | unread | 15 | 0 | 10 | 10 | 0 | 10 | 45 |
| 19 | Nine Casino | 47 of 74 | 19 | 15 | unread | 0 | 0 | 0 | 10 | 44 |
| 20 | Wild Fortune | 33 of 74 | 13 | 0 | 0 | 0 | 10 | 10 | 10 | 43 |
| 21 | Winz.io | 52 of 74 | 21 | 0 | 0 | 0 | 10 | 10 | unread | 41 |
| 22 | Wolf.bet | 24 of 74 | 10 | 0 | 0 | 10 | 10 | 0 | 10 | 40 |
| 23 | Punt Casino | unread | unread | unread | unread | 10 | 10 | 10 | 10 | 40 |
| 24 | Bitcasino.io | 0 of 74 | 0 | 15 | 0 | 10 | 10 | 0 | 0 | 35 |
| 25 | Stake | 1 of 74 | 0 | 15 | 0 | 0 | 10 | 0 | 10 | 35 |
| 26 | BetFury | unread | unread | 15 | 0 | 10 | 10 | 0 | unread | 35 |
| 27 | Vave | unread | unread | 15 | 0 | 0 | 0 | 10 | 10 | 35 |
| 28 | DuckDice | 11 of 74 | 4 | 0 | 0 | 10 | 10 | 0 | 10 | 34 |
| 29 | JacksClub | 9 of 74 | 4 | 0 | 0 | 10 | 10 | 0 | 10 | 34 |
| 30 | Rakebit | 6 of 74 | 2 | 0 | 0 | 10 | 10 | 0 | 10 | 32 |
| 31 | CoinCasino | 25 of 74 | 10 | 0 | 0 | 0 | 0 | 10 | 10 | 30 |
| 32 | Sportsbet.io | 12 of 74 | 5 | 15 | unread | 0 | 10 | 0 | 0 | 30 |
| 33 | Fairspin | unread | unread | 0 | 0 | 10 | 10 | 0 | 10 | 30 |
| 34 | Rocketpot | 23 of 74 | 9 | 0 | 0 | 10 | 10 | 0 | 0 | 29 |
| 35 | Cloudbet | 1 of 74 | 0 | 15 | 0 | 0 | 0 | 0 | 10 | 25 |
| 36 | Bao Casino | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 37 | BC.Game | unread | unread | 15 | unread | 10 | 0 | 0 | unread | 25 |
| 38 | Betriot | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 39 | Bizzo Casino | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 40 | Casinoly | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 41 | Duelbits | unread | unread | 15 | 0 | 0 | 0 | 10 | unread | 25 |
| 42 | Freshbet | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 43 | Goldenbet | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 44 | Livecasino.io | unread | unread | 15 | 0 | 0 | 0 | 10 | 0 | 25 |
| 45 | MyStake | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 46 | Rolletto | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 47 | TrustDice | unread | unread | 15 | unread | 10 | 0 | 0 | unread | 25 |
| 48 | Velobet | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 49 | Winshark | unread | unread | 15 | unread | 0 | 0 | 10 | unread | 25 |
| 50 | LeeBet | 5 of 74 | 2 | 0 | 0 | 0 | 0 | 10 | 10 | 22 |
Why these BitStarz alternatives are ordered rather than recommended
Twenty-one of the 50 had the territory clause stay shut on 26 August 2026. Those rows carry an empty cell on the heaviest check, and 15 of them settle on the same total of 25 points, which is what the four light checks pay when almost everything else is silent.
That tie is the shape of what we managed to open, and it is no finding about those 15 businesses. An unread clause is capable of being anything, including an exact copy of the BitStarz list, and ranking it low would convert our own gap into a verdict. So the cell says unread, the row keeps its place, and the reader is told which of the two situations produced the number.
Set that against how the phrase is usually served. Three pages ranking for it were read on 2 September 2026, and between them they fill 21 slots with 19 distinct brands.
| page read on 2 September 2026 | brands named | shared with the other two |
|---|---|---|
| psu.com news post | 5 | 0 |
| ordb.co comparison | 10 | 2 |
| casinorankr.com list | 6 | 2 |
Zero brands appear on all three. Two appear twice. The property that put any of those brands on any of those lists is withheld by all three pages, which leaves a reader holding a shortlist and nothing to measure it with. An audience-measurement tool queried the same day named five neighbouring domains for BitStarz by estimated traffic, and the news post carries zero of the five. An estimate of that kind is a model rather than a counter reading, and it is filed here as one more dated claim.
One row makes the difference concrete. MyStake finishes on 25 points with three of its seven cells empty, and MyStake is also one of the five brands recommended by the news post that ranks for this phrase. Two pages, one brand, and opposite impressions: theirs reads as an endorsement, ours reads as a hole in what we managed to open. Neither of the two is a measurement of MyStake, and only one of the two says so.
Across the whole page the arithmetic of that gap comes to 54 empty cells out of 350, leaving exactly half the rows scored on every one of the seven. The check that almost always resolved is the licensing authority: 49 of 50 rows name one, and 34 name the Curaçao Gaming Authority exactly as BitStarz does. The one that almost never resolves in an operator’s favour is the company on the permit, empty on 16 rows and matching on zero.
Where the paid row sits
Vave pays for its placement on this site, its row in the table above carries the tag that says so, and it was scored on the rules that governed the other 49. Those rules were unkind to it. Vave prints Curaçao as its jurisdiction and zero permit numbers, so a reader has nothing to look up; its territory clause was left unread on the reading date, which leaves the strongest of its seven cells blank; and it lands 27th of the 50 on a total of 35.
Its own documents did open on the payout side, read on 25 August 2026: clause 8.1 gives three working days to process a withdrawal, clause 8.7 reserves an identity check at the operator’s discretion before any payout with 14 days to supply documents, and clause 8.8 pays anything above 50,000 USDT in instalments.
One contradiction inside our own files belongs on this page rather than in a private note. The company named on the Vave pages was recorded as TechOptions Group B.V. on 25 August 2026, and the operator record re-read on 2 September 2026 names Latcas B.V. Both readings stand here with their dates, the pair is filed as unresolved, and picking the tidier one would be an editorial act wearing the costume of a fact.
What this measure declines to settle
A territory clause is a term in a contract, and it settles exactly one question: which players that operator turns away. What the law where a reader lives permits is a separate question with a separate source, and this site answers the first and points at the second.
Good standing sits outside the seven checks as well. Third place belongs to a brand whose own permit entry came back marked Revoked on the day it was looked up, and third place is where the arithmetic left it, because these totals rank resemblance and were built for that alone. The register reading travels beside the score instead of inside it, with the number and the date set out on the seven checks in full.
The last omission is the largest. Nobody here holds an account with any operator named on this page, has deposited at one or has been paid by one, so a sentence about how the withdrawals felt is absent by design. What stands in its place is a clause number, a date, and a formula printed where a reader can check it, set out at length on how similarity is counted and on who writes this.