What Does a Bonus Buy Return? Bought Bonuses by Provider, Measured Ticket by Ticket

What does a bonus buy return on average?

Across 10,200 logged purchases in 113 games, a bought bonus returned a mean of 0.93 times its price, and 73% of all tickets paid less than they cost. The median ticket came back at 0.49 times the price. Only 40 of the games that sell their bonus, 35% of them, averaged at or above the price over the logged series; in every other game the series ended with less money than it started with.

The method is the same for every game. On the provider’s official demo server, in one session at one stake, the rig buys the bonus at its listed price, usually fifty times per run, lets the feature play out and logs the total it returned; games re-run later add more tickets to the same pool. Returns are expressed as multiples of the ticket price, so a 100x ticket that paid 86 stakes shows as 0.86x. Natural triggers from the base game are logged separately and never enter these figures; how often those arrive by provider is the companion report.

Which provider’s bonus buys pay back the most?

By mean return per ticket, Quickspin leads with 1.22 times the price averaged over its tested games, and Thunderkick trails with 0.82. By how often a single ticket lost, the best provider is Thunderkick with 71% of tickets below price and the worst Relax Gaming with 78%. The two rankings disagree, and the gap between them is the whole story of bonus buying.

Read the mean and the median columns together. The mean is what a very long series of purchases tends towards; the median ticket is the one in the middle of the pile, the result half the tickets beat and half did not. In every provider the median sits far below the mean, because a bought bonus returns a small amount most of the time and a very large amount occasionally, and the occasional tickets carry the average. A provider whose mean is above 1.0x is not a provider where buying wins; it is a provider where the rare big ticket was big enough, in the logged series, to cover the many small ones.

Why does a bonus buy with a positive average still lose most of the time?

Because the distribution is lopsided. On 19 October 2026 the best single provider mean in the table belonged to Quickspin at 1.22 times the price over three games, and three quarters of its tickets had still paid less than they cost. Sakura Fortune 2 averaged 1.54 times its 55x price with 70% of tickets below it; Gates of Olympus Super Scatter averaged 1.51 times a 100x ticket with 58% below. A player buying ten tickets in such a game should expect seven of them to lose and should expect the series to come out ahead only if one of the remaining three lands in the tail.

The same arithmetic runs the other way at the bottom of the table. Big Bass Halloween 2 returned 0.42 times its price on average over the logged series with nine tickets in ten below the price; Book of the Fallen 0.50 and Fire Portals 0.57, both with four fifths of tickets below. These are not games whose bonus is broken, they are games whose tail did not show up in the logged series, and the re-test report shows how far a second series of the same game can land from the first. The current extremes are in the two tables further down, recalculated whenever a test is published.

Provider by provider: how the bought bonuses behaved

The order follows the number of tested games that sell a bonus; the numbers inside each paragraph recalculate from the current catalogue, the dated ones are the logs of 19 October 2026. Each provider name links to a page with every one of its tests.

Pragmatic Play

The biggest pool by far: 64 of 77 tested games sell their bonus, almost all at 100 times the stake, and together they account for 6,150 logged tickets. The mean return is 0.93 times the price, the median ticket 0.48, and 71% of tickets paid less than they cost; 24 games averaged at or above the price, led by Gates of Olympus Super Scatter. The provider’s spread is the lab’s spread in miniature. On 19 October 2026 Buffalo King Megaways averaged 1.16 times its price with 64% of tickets below, 5 Lions Megaways sat almost exactly at break-even, 1.01 times with 66% below, and Big Bass Splash returned 0.64 with 78% below, a game whose natural bonus arrives often but whose bought one paid back little in the logged series. Sword of Ares led the Pragmatic buys that day at 1.25.

Relax Gaming

17 games with a buy, a mean return of 0.87 times the price and 78% of tickets below it. The provider’s purchases are dominated by its Money Train line, where the bought feature is a respin grid with persistent multipliers rather than free spins, and the results split the way the games do: Iron Bank 2 averaged 1.28 times its price on 19 October 2026 while Money Train 2 averaged 0.63 with 86% of tickets below. The best Relax ticket series now belongs to Iron Bank 2, the worst to Money Train 2.

Nolimit City

The provider that prices its bonuses differently: several tiers per game instead of one, and the lab buys the one the game presents as its main feature, which is why the median price in the table sits below 100x. Over 15 games the mean return is 1.04 times the price with 76% of tickets below. On 19 October 2026 Tombstone: No Mercy averaged 1.51 times a 65x ticket and Das xBoot 1.34 times a 75x one, two of the three best series in the lab; the provider’s high ceilings mean its tails, when they land, land hard.

Quickspin, Big Time Gaming, Thunderkick, Blueprint Gaming and Hacksaw Gaming

Smaller pools of two to four games each, so a single series moves the provider mean, and the table should be read as a sketch for these five. Quickspin’s three buys, priced between 55x and 70x, led the lab on 19 October 2026 at 1.22 times the price on average, carried by Sakura Fortune 2. Big Time Gaming’s four 100x buys averaged 0.93, the same as Pragmatic’s much larger pool; Hacksaw’s three averaged 0.88, Blueprint’s two 0.83 and Thunderkick’s four 0.82. The live means are 1.22, 0.93, 0.82, 0.83 and 0.88 times the price respectively.

Play’n GO

No bonus buys at all: the provider does not sell its features, so its 5 tests in the lab contain natural triggers only, and the provider is absent from the table above by design rather than by lack of data. Its bonus frequency and pay are in the trigger-rate report.

Is a cheaper bonus buy a better one?

Not by itself. The cheapest median ticket among the providers in the table currently belongs to Quickspin at 60 times the stake, returning 1.22 times that price on average; the dearest, Pragmatic Play at 100x, returns 0.93. Price is set by the provider to match the feature’s expected value, and a 60x ticket buys a feature designed to be worth about 60 stakes in the long run, just as a 100x ticket buys one designed to be worth about 100. What a cheaper ticket does change is the arithmetic of a session: with the same bankroll a player affords more tickets, and more tickets mean more chances for the tail to show up before the money runs out, at the cost of a smaller tail when it does.

The declared figure to compare against is the feature’s own RTP, which many providers print in the game rules next to the base-game RTP and which is usually a fraction of a point higher. The lab records the declared RTP of the build it tests; the measured return of the bought series is what the tables here show, and the gap between the two on a single game is almost always sample size rather than design, for the reasons the declared-versus-measured report works through.

Which bonus buys paid the most, and which the least?

The two tables are the lab-wide extremes by mean return per ticket, with the median ticket, the share below price and the best single ticket of each series alongside; both recalculate from every published test. A game at the top of the first table earned its place with one or two tickets far above the rest, which is why the median column rarely agrees with the order.

Is buying the bonus better than waiting for it?

A bought bonus replaces the rounds a player would otherwise spend waiting for a natural one, and the two routes can be priced against each other with the lab’s figures. In the median game a natural bonus arrives once per 233 rounds, so at a 100x ticket the purchase costs about the same as the base-game rounds it skips, and the base game pays something back along the way while the ticket does not. The bought feature, on the other hand, is sometimes a richer version of the natural one, with guaranteed symbols or an extra tier, which the game rules state and the reviews note under Bonus lab. There is no general answer; the per-game pages set the mean return of the bought series next to the mean pay of the natural rounds, and the best bonus ranking sorts every test by the bought figure.

How these numbers update

Every number and table in this report recomputes from the published tests whenever a new one is added or a game is re-run; the figures quoted for 19 October 2026 are the static ones and are marked as such. Providers with two to four bought games will move the most, a provider joins the comparison table when it has at least two games with a bought series and 3 tests overall, and the register below lists every game that sells its bonus with its price, mean return and the share of tickets that paid less than they cost.