Bonus buys,
honest math.
Bonus buys look mathematically fair: pay 100× to skip the base game and enter the bonus round directly. In reality the "feature-buy RTP" on most slots is 0.5–2% worse than base play. Plus the variance per buy is compressed (you skip the small base-game wins) but not actually lower. Below: when buying makes sense, when it doesn't, and the cost-x ladder for every bonus-buy slot we index.
Why bonus buys look fair on paper
A typical bonus-buy slot has its bonus round priced at a cost designed to match the game's headline RTP. On Sweet Bonanza, you'd need to spin roughly 230 base-game spins to organically trigger one bonus round (the documented 1-in-230 trigger rate). Pragmatic prices the bonus buy at 100× the base bet — which works out fair-ish on the surface: 230 spins × 0.435% per-spin expected loss ≈ 1 base bet of expected drawdown to get to one bonus, which costs 100× to buy directly.
The math looks like the operator is just front-loading the variance: skip the base-game grind, pay the equivalent of the long-run loss to get to the payoff event. Same RTP either way, in theory.
In practice, providers usually publish a separate "feature-buy RTP" that's 0.5–2% worse than the base RTP. Which means the bonus buy isn't free — it's a small, mostly-hidden premium charged for the convenience.
Feature-buy RTP, and why it’s lower
Pragmatic Play, Hacksaw Gaming, and NoLimit City all publish two RTP figures for their bonus-buy slots: base RTP and feature-buy RTP. The feature-buy RTP is consistently lower. Sample numbers from public game documentation:
Base RTP Feature-buy RTP Delta Gates of Olympus (Pragmatic) 96.50% 96.50% 0.00% Sweet Bonanza (Pragmatic) 96.51% 96.48% 0.03% Wanted Dead or a Wild (Hack) 96.38% 96.27% 0.11% Mental (NoLimit) 96.06% 95.55% 0.51% San Quentin xWays (NoLimit) 96.03% 95.30% 0.73% Tombstone N' Cash (NoLimit) 96.18% 94.71% 1.47%
The pattern is clearest at the top end of the volatility spectrum. The "harder" the bonus round (more multipliers, more upside, harder to clear), the bigger the feature-buy premium. By buying, you're paying for variance compression — getting to the bonus deterministically rather than paying the variance to find it organically — and the provider charges for that compression.
This isn't a scam. It's a documented design choice with the math published in the in-game info panel. But it does mean "bonus buy is RTP-neutral" is folk wisdom that doesn't survive contact with the actual numbers.
Bonus-buy cost tiers
When buying makes sense, when it doesn’t
When the buy makes sense
- Time efficiency. A 100-spin grind to one bonus takes ~5 minutes. A buy delivers the same bonus event in ~30 seconds. For high-volume play sessions, the time you save is real and the small RTP premium is worth it.
- Variance compression. Base play has a long tail of breakeven sessions where you grind out 200 spins without hitting the bonus. Buys give you the bonus-round outcome distribution directly. If you want clean exposure to the bonus round's variance, that's what buys give you.
- Streamer/content reasons. The hit-rate of bonus rounds on stream is the show. A 100-spin base-game grind isn't TV. For content creators, the buy is the production decision more than the math one.
When the buy is a tax
- Bankroll discipline. A 100× buy on a $5 base bet is a $500 commitment on a single event. Most recreational bankrolls can't absorb 3-4 dud buys in a row. Base play smooths the variance with small wins between bonus rounds.
- When the feature-buy RTP delta is >1%. Tombstone N' Cash (1.47%) and similar high-volatility NoLimit games charge a real premium. Over volume, that's 5–10x worse long-run than the base RTP suggests.
- When you actually want to play the slot. Base game is part of the experience for most slots — feature integration, anticipation, near-misses, pre-bonus narrative. Buying skips all of that. If the buy is just a way to get to a payoff event fast, you're not playing the slot, you're playing a $100-bet single event.
What a buy session actually looks like
The distribution of bonus-round outcomes across most upper-tier slots is roughly:
Sweet Bonanza bonus-round outcome distribution (estimated from the streaming community's catalogued bonus pool patterns): Pays 0–10× ~20% "dud" — covers <10% of buy cost Pays 10–50× ~50% breakeven-to-modest — covers 10–50% Pays 50–200× ~22% profitable — covers 50–200% Pays 200–1k× ~7% big — multi-buy recovery Pays 1k×+ ~1% monster — pays for hundreds of buys The expected return on a 100× buy across that distribution should land near the published feature-buy RTP (~96.5%), but on any 5–10 buy session, modal outcome is "two duds, three breakevens, one profitable" — net loss of roughly 30-40% of total stake committed.
That distribution is what makes bonus buys feel random in practice. The skewed-positive tail has wins big enough to recover a session, but they land less than 10% of the time. The middle of the distribution is the grinder — modest payouts that slowly bleed against the buy cost.
Streamers who buy 30 bonuses in a session and end even or up are riding the right tail. The same 30 buys for a recreational player ending down 60% is also normal variance. Same math, same slot, same feature-buy RTP. The variance just feels different when the sample size is small.