One word, five questions, and a build that answers any of them badly answers another well

1

Five questions hiding inside a single adjective

Ask what makes a reel build better than another and five answers come back, each of them reasonable, each measurable in pounds, and no two of them pointing at the same build. The word carries all five at once and names none of them, which is why a page promising the best ten of anything is answering a question nobody asked in a form nobody can check.

Here are the five, stated as questions rather than as adjectives. Which build makes an hour cost least. Which build has the most happening in that hour. Which build gives the best chance of a payout large enough to change the evening. Which build keeps a fixed balance alive the longest. And which build hands back the most cash on an ordinary press that pays. Every one of those is a real thing somebody might want and every one of them has a number.

The trouble is not that the five are vague. It is that they are tied together by arithmetic, so improving one forces another down, and the forcing is exact rather than approximate. A build with four times as many paying presses pays a quarter as much on each of them. A build with a serious chance of a very large return takes the funding for it out of the ordinary line, and a fixed balance consequently lasts a third as long. Those are not editorial judgements. They are consequences of one published percentage and a fixed total.

No title appears on this leaf, here or anywhere below, for the reason this whole site exists: the fact sheet behind it carries no claim about any individual build, so a sequence of ten resting on figures nobody published would be an invented quantity sitting beside checkable ones. What follows instead is the criterion, taken apart and priced, which is the part of the question that keeps working after any particular build has been replaced.

Worth saying plainly before the arithmetic starts: none of the five criteria is the right one. Somebody who wants an evening to last has a different correct answer from somebody who wants a story to tell, and no arithmetic adjudicates between them. What it can do is say what each choice costs in the currency of the remaining four criteria, which is the business of this leaf.

One pair in the five does not fight, which is worth flagging early so that the page is not read as a counsel of despair. The cheapest hour and the longest-lasting balance pull in the same direction, both being helped by a smaller press. Everything else on the list trades, and part eight sets out every pairing in one table.

2

Settings everything below is priced against

Four quantities are fixed for the length of this leaf and each is labelled wherever it is spent, an assumption spent quietly in a later line having a way of hardening into a fact that nobody ever checked. Ninety-six per cent is the return and 10p the press, both of them standing assumptions across this site rather than readings taken off anything. The balance is £20.00. And the rate of pressing is one press every four seconds, which gives 900 presses in an hour and is this leaf’s own setting rather than a published floor.

From those four settings everything else follows by multiplication. Nine hundred presses at 10p put £90.00 over the reels in an hour, of which the margin keeps four hundredths, so an hour is expected to cost £3.60. The average return on one press is 9.6p, which is the published percentage applied to the stake and is the single figure every criterion below has to be reconciled with.

Nine point six pence a press is the constraint doing all the work on this leaf, so it is worth stating in the strongest form. Whatever a build does with its payouts, however often it pays, however large the largest payout is, the average across presses comes to 9.6p at this press value and this return. A build cannot put more into one part of its distribution without taking it out of another, and the whole of parts four to seven is that sentence worked through with figures attached.

The balance matters because two of the five criteria are about a fixed sum rather than about a fixed length of time. Twenty pounds at a margin of four hundredths is expected to put £500.00 over the reels before it has gone, or 5,000 presses, which at 900 an hour is five hours and thirty-three minutes. Those three figures describe the same quantity in three units and part six uses all of them.

One caution on the pressing rate. Four seconds a press is slower than a licensed build must allow and faster than most people keep up over a long sitting, so it sits somewhere sensible without being measured. Change it and every hourly figure moves in direct proportion while nothing priced against the balance moves at all, so the two families of criteria do not contaminate each other.

A fifth quantity is used from part five onward and is introduced here so that it does not arrive unlabelled. Where a build holds part of its return in a rare large payout, this leaf describes that part as a share of the 96 and converts it into a probability by dividing. Nothing anywhere publishes such a share for any build, so every figure drawn from it is an illustration of a mechanism rather than a reading of a market.

3

Cheapest hour, and the quantity it is tied to

Take the first meaning. Best is whichever build lets somebody spend an hour at the reels for the smallest expected outlay. That is the easiest of the five to compute and the only one with a published quantity on both of its inputs.

Expected cost of an hour is presses times stake times margin. On this leaf’s settings that is 900, 10p and four hundredths, giving £3.60. Halve the press to 5p and the hour costs £1.80. Double it to 20p and the hour costs £7.20. Leave the press at 10p and move the return instead: two points in either direction puts the hour at £1.80 or at £5.40, by the same arithmetic approached from the other input.

Notice what is missing from that table and from this criterion generally. Nothing about how the build pays, how often it pays, or how large its largest payout is enters the calculation anywhere. Cost per hour depends on three numbers and all three are available before a penny moves. A reader whose only question is what an hour costs is done, and needs none of the rest of this leaf.

The trade appears the moment a second criterion is admitted. Cost per hour falls when pressing slows or the stake shrinks, and both also reduce what happens in the hour. Press every eight seconds rather than every four and the hour costs £1.80 and holds half the events. The cheapest hour available is the one in which nothing happens at all, which is a reductio rather than a recommendation and shows that cost per hour read alone has a degenerate answer. Part four puts the events back and fixes the relation between the two.

4

Most happening per hour, and the identity behind it

The second meaning is the one most people have in mind without saying so. Best is whichever build keeps something happening, where something means a press that gives money back rather than taking it. Call the proportion of presses that pay anything the event rate, and this is where the arithmetic becomes unforgiving.

The constraint is one line. Event rate times the average amount handed back on a paying press comes to the return rate. Where one press in four pays, the average paid press returns 3.84 stakes, which on a 10p press is 38.4p. One in three gives 28.8p. One in ten gives 96p. One in twenty gives £1.92. Four times the event rate buys exactly a quarter of the event.

So a reader asking for the busiest hour is asking for the smallest returns, and the two requests are the same request wearing different clothes. There is no build anywhere that pays often and pays well at a fixed return, because the product of the two is pinned by the only figure the rules screen publishes.

Convergence is the whole of that drawing. Four builds that feel entirely unlike each other arrive at one figure on the right, the right being the published percentage and the left being a way of distributing it. A reader picking between those four event rates is picking a texture rather than a price.

Put the event rate against the first criterion and a second fixed relation appears. Events in an hour are the press rate times the event rate, and the hour’s cost is the press rate times stake times margin, so both scale with the press rate and the ratio between them does not. At one press in four and the settings held here, an hour brings 225 paying presses for £3.60, which is 62.5 events for every pound of expected cost. At one in ten it is 90 events for the same £3.60, which is 25 to the pound.

62.5 paying presses for each pound of expected cost at one press in four, against 25 at one in ten

Events for each pound of cost is therefore the honest form of this criterion, and it depends on the event rate alone. Nothing a reader does to the stake or the pace changes it. What it buys is smaller events, in exact proportion, which is the sentence the first paragraph of this part was heading towards.

5

The chance of something large, and what pays for it

The third meaning is the one a lobby advertises hardest. Best is whichever build offers a real chance of a payout big enough to be worth describing afterwards, and the figure usually quoted is a multiple of the press. A build announcing a thousand times the stake is announcing £100.00 on a 10p press.

What nobody announces is the probability, and the probability follows from the share of the return the large payout is given. Let a tenth of the 96 per cent be delivered by that payout, which is 9.6 points of return. Dividing 9.6 points by the thousand-times multiple gives a probability of 0.0096 per cent a press, or roughly one press in 10,417. At this leaf’s pressing rate that is one expected occurrence every eleven hours and thirty-five minutes.

The ordinary line has to be funded out of what remains. With 9.6 points committed to the large payout, everything else shares 86.4 per cent rather than 96, so the margin on an ordinary press is thirteen point six hundredths instead of four. That single change is the engine behind parts six and seven, and it has not required a build to be worse in any sense a rules screen would record.

Read the last two columns together and the trade is visible without any further working. Every point of return moved into the rare payout is a point taken off the ordinary line and added to the margin a press carries, and the margin a press carries is what decides how long anything lasts. The reach is not free and it is not paid for by the operator. It is paid for out of the same 96 pence.

One figure is worth computing because it is the one a reader actually wants. On the tenth-share row, a £20.00 balance buys about 1,471 presses before the ordinary line exhausts it, and the chance of meeting the large payout inside that many presses is roughly one in eight. That is this leaf’s own working from the rows above, approximate because it ignores what happens after the payout lands, and it is a far smaller number than the advertisement implies and a far larger one than nothing.

6

How long twenty pounds lasts, in pounds

The fourth meaning is survival. Best is whichever build keeps a fixed balance in play longest, which is the criterion somebody choosing an evening’s entertainment is really applying even when they describe it some other way.

Survival has a clean formula and it is the same compounding that governs any recycled balance. Any sum is expected to send itself divided by the margin over the reels before it has gone, each returned pound being staked a second time. At the published 96 per cent, £20.00 produces £500.00 of turnover, which is 5,000 presses, which is five hours and thirty-three minutes at this leaf’s pressing rate.

Now apply part five. Where a tenth of the return sits in a payout arriving once in 10,417 presses, the ordinary margin is 13.6 hundredths rather than four, so the same £20.00 produces £147.06 of turnover and 1,471 presses, which is an hour and thirty-eight minutes. The balance has not changed. The published return has not changed. Only the shape has, and the shape is settled nowhere in writing.

Turnover £20.00 buys with nothing in the tail £500.00 The same sum with a tenth of the return there £147.06 Play given up for one chance in eight £352.94

Three hundred and fifty-two pounds ninety-four comes out twice from workings with nothing in common. Subtract the two turnovers and £500.00 less £147.06 leaves £352.94. Or forget turnover and count presses: 5,000 against 1,471 is a shortfall of 3,529 presses, and 3,529 presses at 10p is £352.94 again, to the penny once the rounding is carried properly. The agreement matters because it fixes the difference on the shape of the distribution rather than on anything else that moved.

Survival also responds to the stake, and here it agrees with the first criterion rather than fighting it. At 5p a press the same £20.00 buys 10,000 presses instead of 5,000, and at 20p it buys 2,500. Turnover is unchanged at £500.00 in all three cases, because turnover is decided by the balance and the margin alone, which is a result worth holding on to: the stake moves the count of presses and leaves the money untouched.

Which gives the honest summary of this criterion. A reader who wants a balance to last should want a small press and a build with little in its tail, and the second of those two wishes cannot be checked anywhere. The first can be checked on the stake panel in seconds, which makes it the only part of the survival question a reader has any purchase on.

7

The size of an ordinary paying press

The fifth meaning is the quietest and it is the one that decides whether an hour feels worth sitting through. Best is whichever build hands back the most cash on a press that pays at all, leaving aside how often that happens and leaving aside the rare payout entirely.

Part four fixed the relation at a flat return: event rate times average paying press equals the return rate, so at one press in four the ordinary paying press returns 38.4p on a 10p stake. Part five then takes some of that away. With a tenth of the return committed to the large payout, 86.4 per cent is left for the ordinary line, and at the same one press in four the ordinary paying press returns 34.6p rather than 38.4p.

Three point eight pence a paying press sounds negligible and compounds into something that is not. At 900 presses an hour and one press in four paying, an hour holds 225 paying presses, so the difference is £8.64 of cash returned in an hour, against a build whose expected cost for that hour is the identical £3.60 either way. The reader whose build holds a tenth of its return in the tail receives less back per event, every event, for the whole of the sitting.

Every figure in that table is the same 96 per cent arranged differently, which is why no row is better than any other in any sense the arithmetic recognises. Reading left to right within a row shows the trade in part four. Reading top to bottom within a column shows the trade in part five. The two trades are independent and a build has to take a position on both.

One point of honesty about this criterion. A reader has more direct evidence on the size of an ordinary paying press than on anything else here, twenty minutes producing a rough sense of it, and a short sample says least about precisely this. A sitting holding two of the better ordinary returns reads as a generous build and is nothing of the sort.

8

All five set against one another in one table

Everything above produces one table and the table is the point of the leaf. Each row takes a criterion and asks what winning on it does to the other four, in the figures this leaf has been using. Where a cell is a cost it says so in pounds or in presses. Where a pair does not trade at all, the cell says that too, because a page claiming everything fights everything would be as unhelpful as one claiming nothing does.

Read down the diagonal and each cell is the criterion stated in its own units, which makes the table self-describing. Read across a row and the costs of winning on that criterion appear in the units of the other four. Read down a column and the same quantity is seen from every other direction, which is where the cells saying no effect earn their place: three of the twenty cells record a genuine independence and the rest record a cost or a help.

Two findings fall out of the table that the parts above did not quite state. The first is that the busiest hour is almost orthogonal to everything except the size of a paying press, so a reader who wants plenty happening pays for it in one currency only and that currency is small returns. The second is that the reach is the most expensive criterion in the list: it costs £352.94 of play on a £20.00 balance and thins every ordinary return by a tenth, and it buys a probability of roughly one in eight within that balance.

The honest use of the table is not to pick a winning row. It is to pick a row knowing what the other four cells in it say, and then to check whichever of the five quantities can actually be checked, which part nine settles and which is a shorter list than the table implies.

9

Which of the five has a published figure behind it

Five criteria, and the count of them with a published quantity behind them is one. Setting that out plainly is the point at which this leaf stops being arithmetic and starts being about what a reader can establish.

The cheapest hour is checkable. It needs the return and the stake panel, both of them one movement from the lobby, plus a pressing rate the reader supplies themselves. The busiest hour needs an event rate, which a build may publish, since the disclosure duty can be satisfied with the odds on winning events in place of a return figure, and which in practice is published rarely. The reach needs a probability nobody prints. Survival needs the share of the return held in the tail, printed by nobody. The ordinary return needs the event rate and the tail share together.

So four of the five criteria rest on quantities settled nowhere in writing, and that is the plain answer to the question this leaf’s search phrase poses. Any ten builds put in an order by one of those four criteria would be ordered on figures nobody could have read, and a reader holds no instrument for telling that case apart from a sound one.

Three British rules bear directly on this leaf’s criteria and they are worth stating because of what they do to the measuring rather than to the outcome. Sound and animation of the celebratory sort are barred on any press whose return fails to exceed the stake laid on it, which means the count of events on the screen is obliged to match the count in part four’s arithmetic. A build may not set out to push anybody into chasing a loss, into a bigger stake, or into carrying on once they have indicated they would rather stop, and a fresh choice is needed each time funds go back in, which is what keeps part six’s balance a fixed quantity. And limit-setting has to be put to a customer as the default choice, at account level, over spans among which a day, a week and a calendar month must all appear, with any reduction landing at once and any rise waiting out a cooling-off period of a day and then a further positive act.

One concession, stated as plainly as the rest. A rulebook published by one offshore regulator and read for this site does require cooling-off periods reachable in a single click and a deposit limit reducible with immediate effect, so on the third of those three the position is closer than a reader might expect. On celebrating a small return and on encouraging a continued sitting it carries nothing, and those two are the ones part four and part six lean on.

10

Settling on a criterion before settling on anything

The method is short because the hard work was done in parts three to eight. Name the criterion first, in a sentence, before looking at any build at all. Then check whichever of its inputs is published, price the ones that are not across a band, and accept that the answer is a range rather than a name.

Step one, write down which of the five questions is being asked. Cheapest hour, busiest hour, best reach, longest survival, largest ordinary return. A reader who cannot pick one has not got a question yet, and a page offering ten of something is offering an answer to whichever of the five the reader supplies silently.

Step two, take the inputs that exist. The return figure and the stake panel are both a movement from the lobby and between them settle the cheapest hour exactly. Where a build publishes the odds on winning events instead of a return, convert: the event rate and the average paying press multiply to the return, so either one plus the return gives the other.

Step three, band the rest. For the tail share, run the figures at nothing, at a twentieth and at a tenth, which gives survival at £500.00, £227.27 and £147.06 of turnover on a £20.00 balance. For the event rate, run one press in three, one in four and one in ten. Printing three answers is not weakness. It is the only honest output where the input is unpublished.

Step four, do the subtraction the site is built on. Print the figure the criterion gives at its best reading, print it at its worst, and rule off the gap. On the reach and survival pair that gives £500.00, £147.06 and £352.94, and the second of those figures is what a reader actually meets where the build they picked holds a tenth of its return in a rare payout.

Two habits keep this from collapsing back into a ranking. Never compare two builds on two different criteria, which is the commonest error on this subject and makes any build look better than any other depending on which cell of part eight’s table is quoted. And label the pressing rate and the tail share inside whichever sentence puts them to work, each time it does, neither having a source and each moving a whole family of figures.

What a reader is left holding is a criterion, one checkable figure, and a band on the rest. That is less than a list of ten and it is the part that survives contact with the arithmetic, because the thing a list of ten was promising was a single answer to a question that has five, and the five pull against each other by amounts this leaf has now printed to the penny.