A data-driven analysis by Tracksino tracked 5,000 bonus-buy purchases across high-volatility slots. The findings were sobering: players returned an average of 91.4% of their bonus-buy spending. High-volatility titles specifically returned only 87.3%, and just 23% of bonus purchases were profitable (returned more than the 100x cost).
A separate analysis of 312 online players by Kyruus Health found that 68% overestimated the frequency of high payouts after purchasing a bonus round, "simply because the visual signal is louder."
These numbers illustrate a structural reality: the bonus buy is not a shortcut to profit. It is a high-variance mechanism that concentrates both wins and losses into short, intense sessions. The 77% loss rate across tracked purchases means that for every player who hits a large multiplier during bought free spins, approximately three others lose their entire buy-in.
The comparison with regular play is straightforward. In base game mode, you trigger free spins naturally (roughly once every 447 spins) and your bankroll depletes gradually at the house-edge rate. With bonus buy, you compress that entire wait into a single 100x payment. The expected return per bonus round does not improve. You are simply paying for speed.
A single bonus buy at 100x costs 50 to 100 spins' worth of base-game budget. If your session bankroll supports only 200 base-game spins, one bonus buy consumes half your budget on a single event with a ~77% chance of loss. That is the trade-off. Some players prefer it. But going in without understanding the maths is a mistake.
There are three ways to enter the free spins round:
- Natural trigger: Land 4+ scatters during base game play.
- Ante Bet: Enable the Ante Bet option to double the scatter landing frequency (at a 25% increase to your stake).
- Bonus Buy: Purchase direct entry for 100x your current stake (unavailable in some jurisdictions, including the UK).
It should be noted that the Tracksino analysis of bonus-buy outcomes comes from an industry data tracker and does not represent a peer-reviewed academic study; its sampling methodology is not fully disclosed.