Released and received are two different hours, and a weekend can sit between them

1

Two events a cashier screen treats as one

Authorisation is a decision and posting is a bookkeeping entry, and nothing requires the two to happen in the same hour. A card payment is approved at one moment and turns up against a balance at another, and the distance between them is where this page lives. On a good afternoon it is half an hour. Across a weekend it is three days.

Eighteen hours and thirty minutes is the figure this note gets for a Tuesday afternoon request, and 65 hours 30 for the identical request made at the identical hour on a Friday. The difference is 47 hours. Nothing in the account changed between the two, nothing in the sum changed, and nobody behaved differently. A calendar was consulted and it answered differently.

Three suppositions produce every figure on this page and each is named here once so that the rest of the note can lean on them. A request is looked at for 18 hours before anybody decides anything. Decisions are handed out during working hours on working days, which this page puts at nine in the morning to five in the afternoon, Monday to Friday. A push against a card posts 30 minutes after it is released, on any day of the week including both weekend days.

None of those three was observed, measured or timed. They are working figures chosen to let the arithmetic run, and a reader who prefers different ones can substitute them and rework every sum below, which is rather the point of writing the model out instead of printing a band.

Somebody searching for a visa casino is usually after the part that is genuinely quick, and the quick part is real: half an hour from release to a balance moving is about as fast as money travels. What the half hour cannot do is start before somebody releases it, and release is the part nobody prints.

2

Hours during which a release can happen at all

Forty hours a week is the size of the window this model allows for releases, being eight hours across five days. A week holds 168 hours. So releases are possible across 23.8 per cent of the week and impossible across the other 76.2 per cent, and a request whose review finishes inside the closed three quarters waits for the window to reopen.

Eighty hours of weekday darkness is the cheap kind of closure, because the window reopens within fifteen hours at worst and the delay is measured in a single night. Forty-eight hours of weekend is the expensive kind, because it can only discharge into one morning and everything that accumulates across it lands in the same queue.

Four parts make up any payout wait and this page is unusual in how they divide. The look at the request is 18 hours. A document round is nought, for a reader whose papers went in before playing, and is dealt with separately below. The rail itself is half an hour. And the calendar contributes between nothing and 47 hours, which makes it larger than the other three put together on the day it bites.

Half an hour of rail against 47 hours of calendar is a ratio of 94 to one, and it is the sentence this page exists to produce. A reader comparing payment routes on speed is comparing the half hour. The thing that actually decides the answer is not on any comparison a cashier publishes, because it is a property of the week rather than of the payment.

3

One request, lodged on each day of a week

Taking a single request at four in the afternoon and moving it across seven consecutive days gives seven answers from one model, and five of the seven are identical. That flatness is worth as much as the two exceptions, because it shows the delay is not a smooth function of anything.

Eighteen and a half hours is the answer on five days out of seven, which means the ordinary experience of this rail is a wait of under a day. Sixty-five and a half is the Friday answer and 41.5 is the Saturday one. A Sunday afternoon request lands in the same 18.5 as a Tuesday, because Monday morning is fewer than 18 hours away from Sunday at four once the review has run.

Friday is the worst day by 24 hours over Saturday and by 47 over everything else, and the asymmetry has a simple cause. A Friday afternoon review finishes on Saturday morning, so the request misses the Friday window by hours and waits the whole weekend. A Saturday afternoon review finishes on Sunday morning, which has only one non-banking day left to cross rather than two.

Nought, nought, nought, nought, two, one and nought is the count of non-banking days each of the seven requests crosses, in order from Monday. Multiplying that count by 24 and adding 18.5 reproduces the first six answers exactly, and the seventh needs a correction of minus 6.5 for the Saturday case because its Monday release is at nine rather than at ten. The calendar is doing nearly all of the arithmetic.

4

Four o’clock Friday set beside four o’clock Tuesday

Laying the two out hour by hour makes it obvious where the 47 hours went, and the answer is that they went nowhere, because nothing was happening during them. Both requests are reviewed for the same 18 hours. Both are released in the first available window. Both post half an hour later. One of them simply waits.

Row three is the whole of the difference and rows one, two, four and five contribute nothing to it. Forty-seven hours appear between the moment a review finishes and the moment anybody is in a position to act on it, and that interval has no owner. It is not the operator being slow, and it is not the rail being slow. It is Sunday.

Two is the count that explains it, and counting is all that is needed. Saturday and Sunday both fall between a Saturday morning review and a Monday morning release, so the arithmetic is 24 plus 24, less the hour between nine and ten in the morning, which gives 47. A reader who can count weekend days between two dates can work out the whole delay without knowing anything about payments.

Three and a half times is the Friday wait expressed against the Tuesday one, 65.5 divided by 18.5. Put the other way, a Tuesday reader is paid before a Friday reader has got past a single Saturday. There is no fee anywhere that buys a reduction of that size, and no rail that delivers it either.

5

Asking 26 hours sooner buys 46 hours back

Thursday at two in the afternoon is the interesting comparison, because it is 26 hours earlier than Friday at four and it answers in 19.5 hours rather than 65.5. A review that begins at two on Thursday finishes at eight on Friday morning, an hour before the window opens, so it is released at nine and posted at half past.

26 h earlier at the request buys 46 h earlier at the balance, a return of 1.77 to one

One point seven seven is that return, 46 divided by 26, and it is the only lever on this page that pays more than it costs. Every other saving available here is at best one for one. Moving a request from Friday afternoon to Thursday afternoon hands back nearly twice the time it gives up, and it is a decision taken with a calendar and no money at all.

Eight in the morning is the hinge, and the hinge is sharp rather than gentle. A review ending at 08:00 on Friday is released at 09:00 the same morning. A review ending at 17:30 on the same Friday is released at 09:00 on Monday, 63.5 hours later. Nine and a half hours of difference at one end of the chain becomes 63.5 at the other, and the amplification is the window doing its work.

Ten in the evening on a Thursday is the knife edge worth knowing about. Reviewed for 18 hours it finishes at four on Friday afternoon, inside the window by an hour, and posts at half past four for a total of 18.5 hours. An hour of extra review and the same request waits until Monday.

Nothing here is advice about when to press a button and it should not be read as any. It is a description of a model whose inputs are named, and the only claim being made is that under those inputs the hour of a request is worth more than the choice of rail. A reader who disagrees with the inputs gets different numbers and the same conclusion.

6

On the slower route the week turns upside down

Not every payment back to a card goes by a push. The older route, where the money travels back as a reversal of the original payment, takes several working days to post rather than half an hour, and this page puts it at three working days from release. Running the same seven requests through it produces a result that is the opposite shape.

One hundred and thirty-eight hours is the worst answer in the third column and it falls on Tuesday, Wednesday and Thursday, which are three of the five best days in the second. Ninety hours is the best answer and it falls on Monday and Sunday. Friday, the disaster of the fast route, comes in at 137, which is an hour better than midweek.

Two things cause the inversion and both are counting. A three-working-day settlement released on a Wednesday morning crosses a weekend before it lands; one released on a Monday morning does not. So the slow route pushes its own weekend problem three days downstream from the release, which moves it onto a completely different set of request days.

Forty-eight hours is the whole of the spread on the slower route, from 90 to 138, against 47 hours of spread on the faster one, from 18.5 to 65.5. Almost the same spread, sitting on top of a base that is five times larger, and pointing at opposite ends of the week. No single rule of thumb about which day to ask on survives both rows.

Which route is in use is therefore a thing worth establishing before working out anything else, and a cashier page that says only that payments go back to the card has not said it. The difference between the two columns is 71.5 hours on a Tuesday and 71.5 on a Friday, and that is larger than any other quantity on this page.

7

Two hours of paperwork costing sixty-five

Papers, where identity was left hanging until after a win, behave strangely in a model with a window in it. Its cost is not its length. Its cost is whether it pushes the end of a review past five in the afternoon on a Friday, and a very short round can do that while a very long one does not.

Forty-eight hours of document checking added to a Tuesday afternoon request moves the answer from 18.5 hours to 66.5, which is 48 hours exactly. The round cost what it took, no more, because the review still ended inside a weekday window. That is the well-behaved case and it is what anybody would predict.

Two hours added to a Thursday evening request behaves nothing like that. Lodged at ten at night, reviewed for 18 hours, it finishes at four on Friday afternoon and posts at half past for 18.5 hours. Add two hours of paperwork and the review ends at six on Friday evening, an hour after the window shut, so it releases on Monday at nine and posts at half past: 83.5 hours.

2 h of document checking costing 65 h, against 48 h of it costing exactly 48

Thirty-two and a half to one is what those two hours bought, which is the sharpest single figure on this page and the least intuitive. A delay is not expensive in proportion to its length. It is expensive in proportion to what it makes the request miss, and a window turns a two-hour delay and a sixty-five-hour one into the same thing.

On the British side this scenario is not supposed to arise, because the check belongs before play rather than after a win. A name, where somebody lives and the date they were born go on file and get verified before anybody gambles, and the regulator has said that this holds where an incentive paid for the play and holds for games carrying no prize. Nought hours is the right entry for anybody who sent documents in before a first press.

Where a cashier holds no British licence the round can land wherever its operator chooses to put it, and the figures above show why the hour it lands at matters more than its length. A reader who cannot control when the question is asked can at least control whether there is anything left to ask, by answering it before playing.

8

Counting 47 hours three different ways

Forty-seven hours is the Friday penalty and it deserves to be valued rather than merely stated, because the three available valuations disagree with each other by several orders of magnitude and the disagreement is the finding.

Two and a half pence, near enough, is the first valuation. A £120 balance held for 47 hours, valued at the four per cent annually this page works in, comes to 2.6 pence. On that reading the weekend is worth about a quarter of a press at ten pence and nobody should care about it at all.

One hundred and twenty-eight is the second valuation, being 47 hours divided by the 22 minutes this site conventionally allows for one phone sitting. A hundred and twenty-eight sittings is the amount of life the delay occupies in the only unit this site measures in, and it is a figure nobody would shrug at.

Fourteen thousand one hundred is the third, being 47 hours at the 300 presses an hour that twelve seconds a press implies. That is not a prediction that anybody presses anything. It is the size of the window during which a balance that has not left is a balance that can still be staked, and it is the valuation a payout page should be most interested in.

Two point six pence, 128 sittings and 14,100 presses are the same 47 hours. A page that picks the first of those is telling the truth and saying nothing. The second is what a reader feels. The third is what the money is exposed to, and it is worked out properly on the note about the pause before release rather than here.

9

Whose money it is while it stands in a queue

Every hour counted on this page is an hour in which a sum has been asked for and not received, and the question nobody asks is where it is sitting meanwhile. On the British side there is a published answer, and the answer is a rating rather than a guarantee.

A licensee has to tell its customers what would happen to money held in their accounts if the business failed, and the disclosure is graded. Condition 4.2.1 fastens on the grade that means a balance is unprotected, and what it obliges is a reminder of that fact to the customer twice a year. Paragraph 3, which is where that interval sits, began binding licensees on 31 October 2025. The ratings are given no starting date by the regulator at all, and a page supplying one has invented it.

Six months against 65.5 hours is the comparison worth drawing on a payout page. Being reminded twice a year that a balance is unprotected is not much, and it is infinitely more than nothing, which is what a cashier holds no British licence is obliged to tell anybody. The rating is a weak protection that exists. Its absence is not a stronger protection stated differently.

One hundred and twenty pounds in a queue for 65.5 hours is £120 on somebody else’s balance sheet for 65.5 hours, and the reader’s position during that stretch is that of an unsecured creditor who has been promised a payment. Nothing on this page suggests there is anything odd in that. It is how a payout is ordinarily put together, and the rating exists precisely because it is ordinary.

Forty-seven hours of that 65.5 are hours in which nobody is doing anything at all, which is where the two halves of this page meet. A delay with a cause is a delay somebody could be asked about. A delay produced by a shut window has nobody to ask, and it happens to a balance whose protection, out beyond the register, nobody has rated or disclosed.

10

The single input a reader actually sets

First, the rail is half an hour and it is not the subject. Authorisation to posting is 30 minutes on this model, any day of the week, weekend included, and that is genuinely quick. Everything slow on this page happens before the rail is reached.

Second, the hour of the request is the only input in a reader’s hands, and on this model it is worth up to 47 hours. Five afternoons out of seven answer in 18.5 hours. One answers in 41.5 and one in 65.5. Nothing distinguishes them except where the eighteenth hour of review lands against a working week.

Third, moving a Friday afternoon request back to Thursday afternoon costs 26 hours and returns 46, which is the one trade on this page that pays better than one for one. Every other improvement available here is bought at par or worse.

Fourth, the older route back to a card runs between 90 and 138 hours against 18.5 to 65.5, and its best and worst days are not the same days. Establishing which of the two is in use settles 71.5 hours, more than any other single question a reader can ask a cashier.

Fifth, 47 hours is 2.6 pence, 128 sittings and 14,100 presses depending on how it is counted, and the three valuations are all correct. A reader choosing which one to believe is really choosing what they think a weekend costs, and this site’s answer is the second and the third rather than the first.