One hundred and ten studios against forty, and the five rows with no count at all

Six of the eleven rows in this table carry a number in the studio column. Put the six in order and the column looks like a ranking. Open the lists and it stops being one.

The six counts, and the five blanks

The Roobet figure is the weakest of the six and is labelled as such. Our record holds the number while the operator's own documents stay unread, which is a softer handle than a count taken beside a clause. Every other figure came off a page the operator publishes for the purpose.

Five blanks are not five small lobbies. Each of those operators runs a catalogue, and each of them will state a figure to any visitor who asks. What is absent is a count from a published page on a stated day, and that absence belongs to this reading. The same eleven rows carry four other columns filled to the same standard, with the source of each cell marked, on the esports comparison.

A drawing, made for this page rather than a picture of anyone's screen. The boxes on the shelf carry no names on purpose: what our record holds is a count of the studios behind a lobby, taken on a dated capture, and not a claim about which titles sit on which shelf. Six rows here carry such a count — 110, 73, 64, 43 and 40 twice — and five rows carry nothing at all, which is what the gaps on this shelf stand for.

The 110 does not contain the 40

Here is the result that undoes the column as a ranking.

Set the Sportsbet.io list of 110 beside the Thunderpick list of 40 and they share 18 studios. Not 40. Eighteen.

Ninety-two of the Sportsbet.io names are absent from the Thunderpick shelf. Twenty-two of the Thunderpick names are absent from the Sportsbet.io one. A reader who moves from the second to the first because 110 is a larger number than 40 gains 92 suppliers and loses 22, which is a swap rather than an upgrade, and no ordering of the column would have shown it.

The other pairings behave the same way. BetFury's 73 covers 20 of the Thunderpick 40. Roobet's 64 covers 14. TrustDice's 43 covers 13. The largest overlap anywhere in the set is half of the smaller list.

Five names in five lists

Take the five counted catalogues together and 182 distinct studios come out of them.

Five of the 182 appear in all five lists: AvatarUX, Endorphina, NetEnt, Popiplay and VoltEnt. Those are the studios a reader can expect to find wherever they land in this table.

A drawing, made for this page. The grid is the shape of the question this section answers: who declares a result, and when the cell stops being able to change. A published rule fixes both in advance and can be read before a wager is placed; an unwritten one is decided after it, by whoever is holding the balance.

One hundred and two of the 182 appear in exactly one list. Nine of them are Thunderpick's alone among the operators counted here, including Apparat, Felix Gaming, MGA Games and Nucleus. A studio at a single operator in a set of five is the case where switching brands genuinely removes a shelf, and it is a much smaller category than the raw counts suggest. What those forty studios sit inside, and the deposit clause that applies to a spin exactly as it applies to a match ticket, is at the casino side.

So the column measures how many wholesale suppliers each lobby happens to have signed, and the answer overlaps far less than the market's reputation for uniformity would imply.

What the counts are counting

An operator publishes a provider index because it wants the names on it found. That makes the page reliable about existence and unreliable about arithmetic in three ways worth stating.

Studios get grouped and rebranded differently by different operators. One catalogue lists a studio's sub-labels separately and another folds them into a parent, and the two decisions can move a count by ten without a single game changing hands.

A studio name says nothing about how many of its titles are on the shelf. A provider with two slots and a provider with two hundred each score one.

And a page captured on one day is a page on one day. Suppliers are added and removed continually, and none of these operators announces either.

Three things a reader usually wants from a catalogue comparison are absent from every document read for this table: the return-to-player configuration each operator has selected where a studio ships several, the actual title count behind each studio name, and any commitment that the shelf will still hold those names next month. None of the three appears in a contract, so none of them appears in a cell here.

Visit

Where the shelf stops mattering

A catalogue decides what there is to play. Two later gates decide whether anything comes back out, and both of them are answered by documents rather than by lobby pages.

The first is the money clause, and this table is thin there by the standards of the trade: one row publishes a processing window with a clause number, and the ceiling column is emptier still. What that silence costs is taken apart at how long a Thunderpick withdrawal takes.

The second is identity, where one row in the table prints an amount and the rest describe a power, laid out at what clause 21.1 leaves to the operator.

And the question of who the company behind each shelf is, and which of the eleven permits a stranger can look up without the operator's help, is at ten permits and one blank.