Bonus buy slots: what a stated price can tell you

A feature purchase is the rarest thing in a casino lobby: a published number that means something. When a game offers to sell entry to its feature for a stated multiple of the stake, that multiple is a statement about the base game’s arithmetic, and it can be reasoned about without any inside information. This page works through what a price implies, what it cannot imply, and where bonus buy slots collide with the clauses that decide whether a balance leaves the site. Every worked figure below is a substitution, labelled as one, because our files hold no return figure, no trigger frequency and no game title for anything.

Bonus buy slots: what is actually being sold

Bonus buy slots do not sell an outcome. It buys entry to a sub-game with its own distribution of results, most of which are worse than the price paid. This is the first and most frequently muddled point: a feature is not a prize, it is a second lottery that the base game would otherwise have reached occasionally on its own.

So the honest question is never “is the feature good” but “is the price right”, and that question is arithmetic. Call the price P, expressed as a multiple of the base stake. Call the feature’s expected return R, in the same units. Then buying returns R ÷ P of what you paid. If the base game returns some fraction of stake, then a fairly priced purchase is one where R ÷ P equals that same fraction. A purchase priced above that is worse than spinning; one priced below is better.

Everything that matters about bonus buy slots is contained in that single ratio, and the reason nobody can compute it from outside is that R is never published. P is published, prominently. R is a property of the game’s implementation, and our files record no such property for any game, from any studio, on any site.

The ratio, worked with substituted numbers

Substitute a base return of 0.96 of stake — a number chosen to make the arithmetic legible, not read from any game — and a price of 100 times the stake.

• For the purchase to match the base game, the feature’s expected return must be 100 × 0.96 = 96 times the stake.
• If the feature actually returns 92 times the stake, the purchase returns 92 ÷ 100 = 0.92 of what was paid. Four percentage points worse than spinning.
• If it returns 99 times the stake, the purchase returns 0.99. Three points better.

Note the sensitivity. A four-unit error in R — four parts in a hundred — moves the purchase from slightly better than the base game to noticeably worse, and four units is well inside the range a designer can move without anyone outside noticing. The ratio is therefore extremely sensitive to a quantity that is never disclosed, which is the whole reason no comparison table can rank bonus buy slots by value.

The same arithmetic run at a different price shows the structure does not change. At a price of 50, matching a 0.96 base needs a feature returning 48 times stake. At a price of 400, it needs 384. P scales R exactly, so the ratio is the only quantity that carries information, and it cannot be computed from published data. What a return figure needs before it means anything is set out in the return figure a feature price is measured against.

What a price implies about the base game

Here the arithmetic runs in a more useful direction, because the price does reveal something about how the base game is built.

Suppose the feature triggers naturally once every N base spins on average, and suppose the feature accounts for a fraction f of the base game’s total return. Then per base spin the feature contributes f × (base return), and the expected value of one natural trigger is N × f × (base return). A price set to match the base game must equal that expected value divided by the base return — that is, N × f.

Substitute: base return 0.96, feature contributing f = 0.5 of it, natural trigger once every N = 200 spins. A matching price is 200 × 0.5 = 100 times the stake. Change the trigger to once every 300 spins with the same contribution and the matching price becomes 150. Change the contribution to 0.7 with a 200-spin trigger and it becomes 140.

Read backwards, then, a published price of 100 is consistent with a family of designs: trigger every 200 spins with half the return in the feature, or every 250 with 40% in the feature, or every 150 with two-thirds. It narrows the possibilities without identifying one. That is genuinely more than nothing — a price of 500 tells you the feature is rare and carries most of the game’s return, a price of 40 tells you the opposite — and it is the most that can be inferred honestly about bonus buy slots from published information. The other single-number pitch is taken apart in the other mechanic sold on a single headline number.

Turnover arithmetic: the same stake, a fraction of the trials

The effect nobody advertises is what a purchase does to the number of independent trials behind a given amount of money staked.

Substitute a unit stake and a price of 100. Spinning 1,000 units means 1,000 separate spins. Buying 1,000 units means 10 purchases. Same money through the game; one hundredth of the sample.

The consequence is not a change in average — the ratio section already established that the average is set by R ÷ P — but an enormous change in dispersion. With 1,000 trials, results cluster near the expectation; the arithmetic of averaging sees to that. With 10 trials, the outcome is dominated by whether one or two of them happened to land well. A player who buys is choosing a small-sample game, and a small-sample game has a much higher chance of both outcomes: ending far ahead, and ending with nothing, over the same total staked.

There is also a rate effect. A purchase resolves in the time a handful of spins would take, so money moves through the game far faster per minute of play. That is an arithmetic observation about turnover and it matters for the next section, because every clause we read about documents and caps is written in terms of amounts, not sessions.

What our files hold about the studios, and what we will not add

Thirty-three of our hundred operator records list the studios named on their own lobby. Counted after merging spelling variants, out of 33: Hacksaw 32, Pragmatic Play 28, Endorphina 26, Betsoft 26, Nolimit City 25, PG Soft 25, Booming 25, BGaming 24, AvatarUX 23, NetEnt 23, Relax 21, Play’n GO 21, Thunderkick 20, Quickspin 20, Big Time Gaming 18, Push Gaming 18, Peter & Sons 15, Kalamba 15, Bullshark Games 14, Blueprint 13, Wazdan 12, ELK 9, Print Studios 7, ReelPlay 5, 4ThePlayer 4.

No field in our library connects any of those studios to a mechanic, a feature price, or a game. We are not going to name which studios sell feature entry, because we would be supplying that from memory and setting it beside numbers that are readings. The census says which suppliers are switched on across these lobbies as of records checked on 26 August 2026, and that is the entire claim.

The counting traps are the usual ones and they are severe. Nolimit City appears across the lists under four different strings, Play’n GO under three — two of which differ only in whether the apostrophe is straight or curly — Peter & Sons under two, ELK under two, Print Studios under two, AvatarUX under two. There are 548 distinct provider strings in the library and a large minority of them are second spellings. Any page that counts strings rather than studios will report suppliers of bonus buy slots that do not exist. A comparison that can be checked instead is grid geometry as a comparison that can actually be checked.

Where a purchase collides with the terms

This is the part that belongs in an article about feature prices and almost never appears in one. A purchase moves a large amount per action, and the clauses we read are written in amounts.

Five records in the whole library state a figure at which additional identity checks apply: Bitcasino.io €2,500 in clause 6.6, Sportsbet.io €2,500 in clause 6.6, Livecasino.io €2,500 in clause 6.6, Rocketpot $2,500 in clause 11.4, Empire.io 2,000 USDT in clause 5.4. Substitute a purchase price of €100. Twenty-five purchases reach €2,500. At €50 a purchase, fifty reach it. At €400, seven do. Those are the only arithmetic statements that can honestly be made about how quickly a purchasing session meets a stated threshold, and they use the operator’s own published figure. All five are counted in the five stated thresholds a purchasing session runs into.

Then the caps, of which fourteen records publish one. Substitute a 40,000-unit result. Oshi Casino caps payouts at €4,000 a day in clause 12: ten payout days. Instant Casino, €4,000 a day in clause 8.1: ten. BetPlays, €5,000 a week in clause 6.11: eight weeks. PlayAmo, €7,500 a week in clause 12: six weeks. Slotum, €15,000 a week in clause 13: three weeks. Bitdreams, €20,000 a month in Section 11: two months. Bets.io caps a transaction at $10,000 in clause 10: four transactions. Wild Fortune, €15,000 a month in clause 11.8: three months.

And the periods. Fairspin gives five days to supply documents in clause 7.6. Vave gives fourteen in clause 8.7, with a three-day processing maximum in clause 8.1 and instalments above 50,000 USDT over up to thirty days in clause 8.8. Bets.io gives thirty calendar days in clause 5. Oshi Casino’s clause 12 allows confiscation if information is not supplied within forty days. Those are readings, dated, and they are the half of bonus buy slots arithmetic that the game screen never shows. They are collected in every cap and period we have read, with dates.

Three claims the arithmetic does not support

That a purchase is better value than spinning. It is better value only if R ÷ P exceeds the base return, and R is never published. Any page asserting the comparison has either obtained R from the studio or invented it. Our files hold no return figure for any game, so we assert nothing in either direction.

That a purchase is worse value than spinning. The symmetric error, and just as common in cautionary writing. The same missing quantity blocks the claim both ways. What can be said is that the price is highly sensitive to R and that R is undisclosed, so a buyer is accepting an unverifiable ratio.

That buying reduces risk by removing the wait. It removes the wait and multiplies the dispersion, because it cuts the number of trials behind a given amount staked — one hundredfold on the substitution above. Fewer trials is more variance, always. Whatever bonus buy slots are, they are not a way of smoothing results.

What is left is short and worth stating plainly. A stated price on bonus buy slots bounds the family of designs the game could belong to. It does not reveal value. It raises turnover per action, which meets published thresholds faster on the arithmetic above. And once a large result exists, the caps and document periods in the operator’s own numbered clauses decide what it is worth and when — which is why every figure in this article that is a reading rather than a substitution comes from a terms page with a clause number and a date attached. Those are the fields behind top no KYC casinos read clause by clause, not screen by screen.