Four starts, four finishes
A duration needs two ends, and the claims name neither
Two people can report the same withdrawal as four minutes and as two days without either of them lying, because a payout offers four plausible places to start a clock and four plausible places to stop it. Of the twelve operators advertising instant here, none publishes which pair it is using.
On this page
A duration is a subtraction, and nobody names the operands
Every speed claim in this market is arithmetic with its inputs removed. Something happened at one moment, something else happened at another, and the difference was printed.
Remove the two moments and the number stops meaning anything.
That is not a rhetorical point. It is the whole reason two accounts at the same operator, on the same day, report figures that look irreconcilable: they are subtracting different pairs of events, and the operator has never said which pair its own figure uses. What the word itself is doing in the meantime is set out in what instant means.
Four candidate start points
Press the withdraw button and four separate moments follow, each of which can plausibly be called the start.
The request is submitted. The player's own clock starts here, and this is the only one of the four the player can observe directly.
The request is admitted to the queue. Several operators in this set describe conditions that must be satisfied before a request is processed at all, which means submission and admission are different events.
The request reaches a human or a risk engine. A queue has a length, and the length is not part of the review.
The review begins. This is the latest defensible start, and it is the one that produces the smallest number.
Choose the fourth and a payout looks instant. Choose the first and it looks like whatever it actually was.
Four candidate finish lines
The other end has the same problem, in mirror image.
The transaction is signed. The operator has done everything it intends to do. Nothing is in a wallet.
The transaction is broadcast. It now exists on a network and has an identifier, which is the first moment a player can check anything independently.
It reaches the confirmation count the operator treats as final. None of these operators publishes that count.
The receiving side credits it as spendable. An exchange or a custodial wallet applies its own policy here, and that policy belongs to neither party to the original transaction.
Only the fourth is the moment a reader would call “the money arrived”. It is also the only one no casino can commit to, because it depends on a third party.
Whose stopwatch it is
There are three parties with a view of this, and they see different things.
The operator sees its own internal timestamps and nothing beyond broadcast. Its figure is honest about the segment it can see and silent about the rest.
The network sees a broadcast and a set of confirmations, publicly, for anyone with the transaction identifier.
The player sees a screen state change and then a balance appear somewhere else, with a gap in the middle they cannot decompose.
A comparison table that prints one number is asserting that these three views agree. They do not, and the design of the promise is what keeps them apart, as set out in claim or obligation.
Eleven clauses, one silence and one blank
The measurement question sits on top of a switch that can stop the clock entirely, and the file for this site records three different states of it.
Eleven of the thirteen records carry a readable verification clause: a paragraph reserving the right to demand documents before a payout, at the operator's discretion, with no amount attached.
One — LTC Casino — has terms containing no verification paragraph at all. That is a silence in the document rather than a discretionary policy, and the table records it as such.
One — Flush — could not be read on the point, and its cell says so rather than being filled in from somewhere friendlier.
Any of those three states can produce a stopped clock, and none of them produces a duration. The clauses themselves are collected in the check that stops it, and the reasoning behind keeping three states rather than two is in how the documents are read.
What a checkable figure would have to say
Five sentences would settle the entire subject, and none of the twelve writes any of them.
Name the start event. Name the finish event. State the unit, calendar or business. State whether the figure is a maximum the operator is bound to or an average it observed. State what happens when it is exceeded.
A figure with those five lines beside it is a term. A figure without them is a mood.
The two things this set does publish with a clause number are ceilings, at two operators only, and those are limits on amount rather than on time — worked through in above the ceiling. Everything else on the speed question is a claim about a subtraction whose operands were never disclosed, and the full list of what would make it checkable is in what would count as proof.
Questions people actually type
Why do payout speed figures disagree so much?
Because a duration is a subtraction and the two operands are never stated. A figure measured from approval to broadcast and a figure measured from request to spendable balance describe the same withdrawal and can differ by orders of magnitude. Neither is false; they are answers to different questions, printed as though they were answers to the same one.Which start and finish should a reader assume?
Assume the most favourable pair, because that is the one a marketing page has every reason to choose: the clock starting when the cashier begins working on an already-approved request, and stopping when the transaction is broadcast. That excludes the queue before approval, the confirmations after broadcast, and any pause for documents.Do any of these operators define their terms?
Not in the documents on file here. Twelve carry the instant marker and none of the twelve publishes a processing window, a confirmation count or a definition of what its speed figure measures. Eleven of the thirteen records on this site do carry a readable verification clause, one has no verification paragraph in its terms at all, and one could not be read on the point.Could a reader measure it themselves?
One account can produce one observation of one withdrawal on one day, which is a data point rather than a measurement of an operator. This site holds no account anywhere and reports no timings. What it can do is set out which two moments any honest figure would have to name.