Every slot lists an RTP figure and almost nobody reads it correctly. Some players treat 96% as a promise to hand back 96 of every 100 deposited. Others assume the number is quietly tuned against them. Neither is true. Here is what RTP measures, how the house edge follows from it, and why neither figure predicts an evening.
RTP and house edge are one number seen from two sides
Return to Player is the share of total wagered volume a game pays back in winnings over an infinitely long run. An RTP of 96% means that for every 100 units staked by everybody, across all time, the game returns 96. It is a property calculated from the game's payout model, not a statistic collected from last week's results.
The detail that trips people up: the base is turnover, not deposits. Money you win and stake again re-enters turnover and is charged the same percentage a second time.
House edge = 100% − RTP. A 96% game carries a 4% edge, which is an expected loss of four units per hundred staked — staked, not deposited.
Worked example: deposit 100 USD, play 500 rounds at 1 USD. Turnover reaches 500 USD, so the expected loss is 20 USD against a deposit of 100. Identical edge, wildly different outcomes depending on how many times the bankroll is recycled, which is why session length decides the cost of a game as much as the percentage does.
Why the percentage says nothing about your session
RTP is a long-run average and it converges slowly. For a medium-volatility slot the spread of observed returns stays wide even after tens of thousands of spins. Across a hundred rounds, an actual return of 20% and one of 300% both sit comfortably inside a 96% model.
The practical consequence: RTP is a tool for comparing games, not for planning a night out. The single thing it governs directly is the average speed at which turnover erodes a bankroll.
Volatility does more work than RTP
Two games at 96% can feel nothing alike. A low-volatility title pays small and often and the balance drifts down smoothly. A high-volatility one takes money in long stretches and hands it back in a single hit.
On a limited bankroll that gap matters more than a point of RTP. Under the same expected value, the chance of going broke before the large payout arrives is several times higher in the volatile game. Reading the return percentage while ignoring the payout distribution is the most common analytical mistake players make.
Where percentages turn into money: wagering
Bonus terms are where the edge stops being abstract. Take a deposit bonus with x40 wagering: a 100 USD bonus demands 4,000 USD of turnover. At a 4% house edge the expected loss across that turnover is 160 USD, more than the bonus itself.
That is not a quirk of one operator, it is arithmetic that holds everywhere, and it dictates how the terms should be read: the lower the edge of the game used for clearing, the cheaper the turnover. Add the 5 USD maximum bet during wagering — 4,000 USD means at least 800 bets — and the 7-day bonus lifetime, and the honest framing appears. A bonus buys playing time, not expected profit.
What actually shifts the maths
Three things. Picking games with a smaller edge: payout multipliers in Tonza Originals are set by the game rules, and every round can be recomputed through provably fair. Cashback through VIP ranks, up to 5%: it leaves the game's RTP untouched, but returning part of what was lost lowers the effective cost of turnover.
And one thing that changes nothing. Martingale, doubling and every other staking system leave the house edge exactly where it was. They reshape the distribution of results — many small wins and a rare heavy loss instead of steady attrition — while expected value stays put. The only variable they move is how abruptly the loss arrives.
