What this table is for
A reader looking for a comparison usually wants one thing: a shortlist. This table cannot give a shortlist, because a shortlist needs a criterion that has been measured, and the only things we have measured are documents. What it can do is put the document facts side by side and show, in the emptiest columns, where the market has not been measured at all.
16 rows carry a first-party read because those brands publish a document we can cite. The other 9 rest on one or two sources, which is why their label says so. That is a fact about the market, not a gap in the research, and it is the reason this site distinguishes a first-party read from a third-party summary.
The empty columns are the point. A comparison site that fills every cell with a number is telling a reader that all twenty-five brands were assessed on the same scale. Ours says the opposite, and the reader can see exactly which cells are blank.
How to read the columns
Audit is the depth of the reading, not a quality grade. Terms audited means five sections were read and quoted: the state list, the redemption section, the bonus terms, the free entry route and the dispute clause. States checked means only the geo-block notice was verified.
States excluded is a count from the operator's own published list, dated by the terms version beside it. It tells you whether the brand serves your state and nothing about whether it will pay you. A brand with a short list is not better; it is broader.
Age is the operator's own minimum, which can be higher than the state's floor. A 21+ brand in a state whose floor is 18 is not doing anything wrong, and a reader who is 19 should treat that column as the binding one.
Terms is the version and date of the document the row was read from. It is the most important column here, because it is what makes the rest checkable: a reader can open the current terms and see whether the version still matches.
Why the redemption columns are missing
There is no minimum, no window and no method column, and the absence is deliberate. Those three numbers are the ones a reader actually wants, and they are the ones most terms do not publish: a handful state a minimum or a play-through, and the rest leave the cashier to a support ticket.
A table with a minimum and a window for every brand would be a table with invented numbers in most rows. That is the trade this page declines, and it is why the redemption detail lives on the brand pages and on the timing page as a description of the process rather than a league table.
What a reader can do instead is fill those columns themselves for the brand they are considering. One support ticket produces the minimum, the window and the document list, and one small redemption produces the only timeline that has been measured for that account.
What the two terms-audited rows actually say
The two audited brands differ in a way that matters more than their bonuses: one has a narrower footprint and thicker documentation, the other has a broader footprint and a thinner redemption section. A reader in a state both accept is choosing between a brand that documents its process and a brand that documents its states.
That is a real distinction and it is not the one the marketing makes. Neither brand is “better” in the abstract; the difference is which question each one answers on paper. A reader who wants to know whether they can play at all gets more from the broader list, and a reader who wants to know what happens at redemption gets more from the thicker section.
The comparison page between those two brands goes through the rows one by one, including the rows where one side is fuller than the other, and it says so rather than averaging the difference into a verdict.
The rows a reader should fill in themselves
Four columns are missing from the table and all four can be produced by a reader in an afternoon. The first is the redemption minimum, which is in the terms and in the cashier. The second is the processing window, which support will state in writing if asked directly.
The third is the document list, which the verification section names and which support will confirm. The fourth is the one nobody publishes: the time from request to arrival on this reader’s own account, on this reader’s own method. That is a sample of one, and it is more than the table can offer.
A reader who writes those four numbers down has a better comparison than this page, because it is specific to their state, their method and their account. The table exists to tell them which brand is worth that afternoon.
What it would take to publish a ranking
A ranking needs three things this page does not have. The first is a sample: at least four audited brands so that an order means something. The second is a criterion that can be measured rather than read: the honest one in this market is time from request to arrival, recorded by us. The third is a way to keep it current, because a ranking that is six months stale is worse than no ranking.
The second item is the hard one. A payout log needs real redemptions, made by us, on accounts in identifiable states, with dates. That is on the list of things to build and it is not built, which is why every page on this site that touches timing says the same sentence: we describe the process and we do not publish a time we have not observed.
Until then, the closest thing to a recommendation this site offers is the brand comparison, which compares the two audited operators on the dimensions that are documented, and says plainly where the documentation is thinner on one side.
How to use the empty rows
An unaudited row is still useful for one purpose: it tells a reader that the brand exists and claims to serve the US market. From there the reader can do what we could not: open the terms, check the state list, and ask support the three questions that produce the missing columns.
It is also useful as a warning. A brand that appears in an affiliate list with a bonus and no terms page is a brand whose terms have not been written for readers, and the absence of a citable document is itself information about how the operator treats the parts of its business that do not attract customers.
The practical sequence for an unfamiliar brand is the one on the operators page: state list, redemption section, one question to support, one small redemption. This table tells you which of those the brand has published; the rest is the test.