Seconds at the front of a journey that is measured in days

1

Three seconds, and the fraction of a trip they are

Put 3 seconds on authorising a payment with a handset, which is an assumption of this leaf and not a measurement anybody made. Put 90 seconds on the alternative, meaning finding a card, typing a long number, and waiting for whatever code arrives. Also assumed, and not published on a page this site may cite. The front end has therefore been shortened by 87 seconds, and 87 seconds is the entire quantity this leaf has to work with.

Now put the money back where it came from. Assume a balance waits 24 hours before it is released and then takes 48 hours to arrive, both assumed rather than read off anything. That is 72 hours on the way out, or 259,200 seconds, against three seconds on the way in.

Divide and the proportion is stark enough to need no adjective. Three seconds into 259,200 is one part in 86,400, and 86,400 seconds is exactly one day, so the relationship can be stated without any division at all: the fast end stands to the slow end as one second stands to a whole day. Expressed as a share of the complete round trip of 259,203 seconds, the authorisation is 0.00116 per cent of it.

The trip read off its first leg 3 seconds The trip with the far end counted in 72 hours and 3 seconds Untouched by anything the handset does 72 hours

None of that is an argument against a fast front end. Three seconds is pleasanter than ninety and nobody should pretend otherwise. The argument is about where a reader’s attention is being pointed, because the one quantity a payment screen shows off is the one quantity that was never the constraint, and the three that are the constraint are settled somewhere a reader cannot see until afterwards.

The question here, then, is not whether the front end is quick. It plainly is. The question is what share of a reader’s own waiting the quickness removes, and on the assumptions above the answer is roughly one thousandth of one per cent.

2

The journey those seconds belong to, leg by leg

A payment into a gambling account and a payment back out of it are usually discussed as one thing, which is how the front end comes to stand in for the whole. Break the trip into legs and the proportions settle themselves without any argument from this leaf.

The reader’s own play is deliberately missing from that table. It takes whatever it takes, it belongs to the reader rather than to any party in the chain, and including it would let a long session disguise how the rest of the trip divides. What is left is the part the reader waits through without being able to do anything about it.

Two of the five rows carry the whole of the waiting and neither of them is anywhere near the handset. One belongs entirely to the gambling account, which decides when a request stops being a request. The other belongs to parties further along, whose names the reader never saw and whose timings appear on no screen the reader passed.

The last row is the one with no author, and it is the only row a reader would actually want to know in advance. No party in the chain publishes it, because no party in the chain is responsible for more than their own leg, and a figure covering all of them would be a figure somebody had to stand behind.

259,200 assumed seconds on the way out against 3 on the way in, from the same account

3

Who fixes the length of each leg, and where

Sorting those legs by where a reader could have settled each one is the exercise this site performs on every subject, and on a question of duration it produces an unusually empty answer.

Whether a handset appears as a choice is settled at the cashier, a row of pictures being something that exists only once an account does. How long the authorisation takes is settled by the handset and is the one length a reader has felt for themselves. How long a balance waits before release is settled nowhere in writing. How long the journey afterwards takes is first encountered on the way out, learnt by its effect rather than by reading anything.

Five rows, and the two that carry 99.99 per cent of the waiting are the two with no address a reader can walk to. This part of the site keeps turning up that shape, and it is worth saying plainly rather than implying: the quantity best advertised is the quantity best known, and the two are the same quantity for a reason.

One row deserves a note in its own defence. The authorisation length is marked as felt rather than published because a reader genuinely does know it, having experienced it, which makes it the only length on the table with any evidence behind it at all. Three seconds is the one honest number in the whole trip, and it governs 0.00116 per cent of the trip.

4

What a faster front end buys, in seconds and in pounds

Take the 87 seconds the handset saves and ask what share of the trip it removes. Against a round trip that would otherwise run to 259,290 seconds, 87 seconds is 0.0336 per cent. A reader who shortened the front end by that much has shortened their own waiting by about one part in three thousand.

Put the same quantity against the money and the answer is cleaner still, because it is a round nought. Take a payment of £30.00 matched in full, giving £30.00 of credit. Forty multiples of that credit want £1,200.00 of turnover, which at the 10p a press assumed sitewide is twelve thousand individual presses.

Each staked pound surrenders four pence, so turnover of £1,200.00 is expected to shed £48.00, which leaves the credit at an expected minus £18.00. Arithmetic from the other side agrees. Forty applications of a four per cent edge come to 1.6; taking that away from one gives minus 0.6; and 0.6 of thirty pounds is eighteen pounds facing backwards. The weighting stays at one in every line here.

Credit read off the match £30.00 Expected close at forty times minus £18.00 Distance between the readings £48.00

Of that £48.00, the amount moved by authorising on a handset rather than by typing is £0.00, and not approximately nought. The difference is made of a multiple and a return, neither of which has any term in it for how the money arrived. A payment route can decide whether a credit exists, which is a different leaf’s subject entirely, but once a credit exists the route has no further part in the arithmetic.

That gives the honest ordering for anybody deciding what to look at first. One unread clause is worth £48.00 on this payment. The entire speed advantage of the front end is worth £0.00 on the same payment and 0.0336 per cent of the waiting. A payment screen presents the second of those and is silent about the first.

5

The wait before release, which is the first real lever

Two things in the trip are worth money or time, and this part is the first of them. Before a payment out leaves, there is an interval in which a request exists and the money has not moved. It is put at 24 hours here, assumed rather than published, and it is the single largest thing any one party in the chain controls alone.

Suppose that interval were two hours instead of twenty-four. The trip falls from 72 hours to 50, a saving of 22 hours or 79,200 seconds. Set that against the 87 seconds the handset saves and the ratio is about 910 to one. One decision taken by one party, about one interval, is worth nine hundred times the whole of the front end.

The ratio is worth stating the other way round as well, because it is more striking read backwards. To match a single hour removed from the wait, a front end would have to be shortened by 3,600 seconds, which is forty-one times longer than the entire authorisation it replaced. There is no version of a payment screen capable of that, because the thing it is shortening was never more than a minute and a half.

What makes the interval interesting rather than merely large is that nobody publishes it. A length nobody states is a length nobody can compare, lean on or argue with, and a reader meets it by watching a screen report that a request is being handled. It is the clearest instance on this leaf of a figure that governs everything and is settled nowhere in writing.

One fair qualification. An interval before release is not automatically an obstruction. Reviews exist for reasons, some of them legal, and a chain with none at all would have properties of its own that nobody should want. The complaint here is about the figure being unstated rather than about the figure being positive, and those are different complaints with different remedies.

6

When the paperwork is asked for, which is the second

The other lever is not a length at all. It is a position: whether a demand for documents falls at registration or at the exit. The work is identical in both cases and the cost to the reader is not, because at registration nothing is waiting behind it and at the exit a balance is.

Assume a demand arriving at the exit adds 48 hours to the trip, which is assumed here and published nowhere. That is 172,800 seconds, or about 1,986 times the 87 seconds the handset saved. One decision about where in the order a check sits is therefore worth nearly two thousand times the entire speed advantage of the front end.

The right-hand column is a constant and it is printed on purpose. Not one of the three levers changes the £48.00 by a penny, because the £48.00 is produced by a multiple working on a credit and none of the three touches either. Time and money are separate accounts on this subject and improvements in one do nothing whatever to the other.

The top bar in that drawing is the one a payment screen advertises and it is drawn at two hundredths of an hour. The two underneath it are drawn at twenty-two hours and forty-eight, and neither is mentioned anywhere a reader passes on the way in. That is the whole of this leaf in a picture, and the numbers are this leaf’s own assumptions rather than anybody’s published tariff.

7

Why the way out is not the way in reversed

Anybody meeting the figures above tends to ask why the exit is not just the entrance played backwards, and the answer is short. Two directions, two sets of parties, two authorisations and two clocks, with the handset taking part in exactly one of them.

On the way in, the thing being proved is possession: that whoever is holding the device is entitled to use what is stored on it. That proof takes three seconds and is complete at the moment the payment is authorised. On the way out, nothing needs proving in that sense, because the money is already inside an account and the question has become where it should be sent and whether it should be sent yet.

Those are different questions with different answerers, which is why no amount of improvement at one end has any effect on the other. A quicker proof of possession does not shorten a review that has not started. It does not shorten a journey between parties who were not involved in the authorisation. It settles nothing about the destination either.

The rule governing which instrument a payment out may be sent to is a separate subject and is worked through on another leaf of this site, where it is derived from first principles and priced. Nothing here repeats that derivation, and nothing here depends on it: the arithmetic above holds whether the destination is fixed, free or decided by a document nobody has read.

What this leaf adds to that subject is a single structural observation. Authentication is asymmetric by construction. It is demanded where money is leaving a reader’s control and is not demanded where money is arriving back in it, so the quick, polished, heavily advertised part of the chain is the part that only ever runs in the direction the house benefits from.

Stating it that way is not an accusation, because the asymmetry has an ordinary explanation. Somebody paying in has to prove they may, and somebody being paid does not have to prove they want it. The consequence is still worth printing: of the two ends of the trip, the one with the engineering on it is the end that takes three seconds, and the one with 72 hours on it has no engineering a reader can see at all.

8

Identity at the front, and the order a licence fixes

Part six handled the position of a check as a lever. Inside the British licence that position is not a lever but a rule, and naming the rule is the only way to put a price on its absence.

Second nought of the trip, ahead of a single press, is where somebody’s identity has to be established and checked, and the condition names three items that must at least be covered: who the person is, the address behind them, and the day they were born. Second one, before any money may be taken at all, is where a warning is owed about which papers could be wanted later, at what point, and in what form. Both duties govern position in a sequence, which is exactly what part six was treating as a lever.

Play nobody paid for sits inside the same rule. The announcement of the current regime puts age checking ahead of gambling funded by a credit or a free bet exactly as it does for gambling funded out of somebody’s own pocket, and ahead of the practice versions kept in a lobby. Before it, the window for finishing an age check ran to seventy-two hours from registration, and none of that window was kept. Dated 7 February 2019, the announcement names the seventh of May as the day the rules start and prints no year against it; the only year its own date leaves available is 2019.

A page written for players puts the same thing in a sentence and gives three grounds for it: how old somebody is, whether they have shut themselves out, and whether they are the person they say. That page is older than the condition, published 23 September 2020 with its last revision on 13 April 2022, and the age of it is worth stating whenever anybody leans on the wording.

Fold that into the trip and the ordering duty prices itself in seconds. Where the rule holds, the check costs the reader minutes at registration and nothing at the exit. Where it does not, the identical check may cost an assumed 48 hours at the exit, which is 1,986 times the whole of the front end’s advantage. The work done is the same work. Only its position has moved.

9

The round-trip sum, for anything authorised on a handset

The procedure takes a sheet of paper and about two minutes, and it works on any payment route that advertises its own speed. List the legs of the round trip. Put a length against each one, from your own experience where you have any and marked as a guess where you do not. Add them. Then divide the leg being advertised by the total.

On this leaf’s assumed lengths that gives three seconds over 259,203, which is one part in 86,400 and 0.00116 per cent of the trip. Any figure of that order means the advertised leg is not the constraint, and the constraint is whichever leg carries the largest share. Here the two largest legs carry 33.33 per cent and 66.67 per cent between them, and neither is published by anybody.

Step two is the money, a separate sum that must never be mixed with the first. Work out the credit as the match presents it. Do it again with the clause applied. Print both and rule the gap off beneath. On £30.00 at forty times that produces £30.00, then minus £18.00, then £48.00 of difference, and no leg of the trip changes any of those three figures.

Two habits keep the result from flattering anybody. Never let the advertised leg stand for the trip, which is the error the whole leaf is about. And never convert a saving in seconds into a saving in pounds, because the two accounts have no exchange rate between them: 87 seconds is worth 87 seconds, and £48.00 is worth £48.00, and no amount of the first buys any of the second.

The action this leaf points at is unglamorous and free. Before paying anything in, ask how long a payment out takes and when documents are asked for. Both answers are about the far end of the trip, both decide 99.99 per cent of the waiting, and neither is on the screen showing how quickly a handset can authorise something.

0.00116 per cent of an assumed round trip happening at the end everybody advertises