Search for “best payout casino” and you get lists ranked by RTP. Ninety-seven percent at the top, ninety-two percent at the bottom.
The ranking is accurate. The conclusion almost everyone draws from it is wrong.
A 97% game and a 92% game are not “good” and “bad.” They are “slow” and “fast.” Both end in the same place. This article shows you the arithmetic, gives you a table that prices every common gambling bet on one scale, and ends with the one number in that table that surprises most people.
Table of contents
Use the links bellow to jump to each section
Table of contents
What RTP actually measures
RTP means return to player. It is the percentage a game is designed to return over a very long run.
The two words doing the damage are “of what.”
RTP is measured against everything you wager, not against what you deposited. That is the entire misunderstanding, and the industry has no reason to correct it.
If you deposit $100 and bet it once, RTP and your deposit describe the same money. But nobody bets once. You bet, you win some back, you bet that too. The money goes through the machine again and again, and the machine takes its percentage on every pass.
The number nobody prints: turnover
Here is the formula that matters.
Expected loss = house edge x total amount wagered.
Not deposit. Wagered.
Take a realistic session. You deposit $100 and play $1 spins on a 97% RTP slot. The house edge is 3%. Online slots run fast, and 500 spins an hour is a conservative figure, because many players are faster than that.
- Spins per hour: 500
- Amount wagered per hour: $500
- House edge: 3%
- Expected loss per hour: $15
Your $100 is not being risked once. It is being risked five times over in the first hour.
Now run the same session on other games:
| Game | RTP | House edge | Expected loss per hour at $1 x 500 spins |
|---|---|---|---|
| High payout slot | 97% | 3% | $15 |
| Typical online slot | 96% | 4% | $20 |
| Average slot | 94% | 6% | $30 |
| Low payout slot | 92% | 8% | $40 |
How long your $100 survives
Flip the formula around and it gets clearer. Divide your bankroll by the house edge and you get the total turnover your money can support before it is gone.
- 97% RTP: $100 / 0.03 = $3,333 of turnover. At 500 spins an hour, that is about 6.7 hours.
- 96% RTP: $100 / 0.04 = $2,500 of turnover. About 5 hours.
- 92% RTP: $100 / 0.08 = $1,250 of turnover. About 2.5 hours.
Read those three lines again, because they are the honest version of every “best payout casino” list ever written.
The difference between the best rated casino on the page and the worst rated one is not winning instead of losing. It is six hours of play instead of two and a half.
RTP does not change the destination. It changes the travel time.
That is not an argument for playing the 92% game. A slower loss is genuinely better than a faster one, and if you are playing for entertainment then buying more hours for the same money is a real benefit. It is just not what the word “payout” makes people think they are buying.
Three things RTP does not tell you
1. Hit frequency. How often the game pays anything at all. Two games can both return 96% while one pays small amounts constantly and the other pays almost nothing until a rare large hit.
2. Volatility. How those returns are spread out. A high volatility 97% game can empty a $100 bankroll faster than a low volatility 94% game, because the swings are wider. In a short session, volatility matters more than RTP does.
3. Your actual result. This is the important one. RTP is a limit that the game approaches over millions of spins. Your session is a few hundred spins. You are drawing a tiny sample from a very large distribution, and a tiny sample tells you almost nothing about the average behind it.
That is why a player can have a winning night on a 92% machine and a losing month on a 97% machine, and why neither result proves anything about either machine.
If you want to check the published RTP for a specific title rather than trusting a site-wide average, most operators list it in the game information panel. Free demo versions are useful here too. A classic like the free Cleopatra slot runs the same paytable as the real money game, so you can read the numbers before you risk anything. Read the figure for the exact title, not the figure for the casino.
Every common bet, priced on one scale
Here is the table I wish existed on every gambling site. It converts each bet into the same unit: expected cost per $100 staked. Now you can compare a slot spin, a roulette bet and a point spread directly.
| Bet | House edge | Cost per $100 staked |
|---|---|---|
| Craps, free odds behind the line | 0% | $0.00 |
| Blackjack, basic strategy, good rules | ~0.5% | $0.50 |
| Baccarat, banker | 1.06% | $1.06 |
| Baccarat, player | 1.24% | $1.24 |
| French roulette, even money with la partage | 1.35% | $1.35 |
| Craps, don’t pass | 1.36% | $1.36 |
| Craps, pass line | 1.41% | $1.41 |
| European roulette | 2.70% | $2.70 |
| Online slots, typical range | 3% to 8% | $3.00 to $8.00 |
| Standard sports bet at -110 | 4.55% | $4.55 |
| American roulette | 5.26% | $5.26 |
| Baccarat, tie at 8 to 1 | 14.36% | $14.36 |
| Four leg parlay, all legs at -110 | 16.98% | $16.98 |
Two rows in that table deserve their own explanation.
The sports bet at 4.55%
At -110 you risk $110 to win $100. Your break even win rate is 110 divided by 210, which is 52.38%. A bettor with no edge wins about half the time, so the expected result of every bet is a loss of $5 for every $110 staked. That is 4.55%.
A sports bettor with no edge is playing a more expensive game than European roulette. Most gambling content will never print that sentence.
The parlay at 16.98%
This is the row that surprises people, so here is the full working.
Four legs, each at -110, which is decimal 1.9091. If all four win, the parlay pays 1.9091 to the fourth power, which is 13.28 including your stake. That is the standard +1228 payout US books offer, and it is exactly the true multiplication of the four prices, so it looks fair.
It is not fair, because the house edge multiplies too.
A coin flip bettor wins each leg 50% of the time, so all four legs win 0.5 x 0.5 x 0.5 x 0.5 = 6.25% of the time. Expected return = 0.0625 x 13.28 = 0.830. That is an 83% return, which is a house edge of 16.98%.
Check it the short way: 0.9545 to the fourth power = 0.830. Same answer. The 4.55% compounds.
So the most expensive bet in the entire table is not a slot machine. It is not even the baccarat tie, the bet everyone knows to avoid. It is a four leg sports parlay, the single most heavily marketed product in betting. And plenty of books pay less than +1228 on four teamers, which makes it worse still.
Payout speed is not payout value
The other half of every “best payout” article is withdrawal speed. Getting paid in one hour instead of three days is a genuine improvement in service, and slow payouts are a legitimate reason to avoid an operator.
But be clear about what you are comparing. Speed changes when you are paid. It does not change what you are paid. One is a customer service metric. The other is a price. Only the price affects your result, and reviews spend most of their words on the one that does not.
The same goes for payment method comparisons, bonus size and interface design. All real, none of them price. Payout rules also vary a lot by market, and a page covering casinos in Iceland will not describe the same terms as one covering Greece or Canada. Compare those details after you have compared the cost of the games, not instead.
Where the arithmetic finally flips
Everything above describes a fixed price. A slot’s 4% is a design specification. A roulette wheel’s 2.70% is exactly 1 divided by 37. Nobody estimated those numbers, so nobody can have gotten them wrong.
A betting line is different. The 4.55% you pay at -110 is a toll, and behind the toll sits a number that a trader or a model or a market estimated about an uncertain future event. Estimates can be wrong. Wheels cannot.
That is the whole reason one game can be beaten and the other cannot, and it is worth understanding properly. I wrote it up in detail here: Can a betting model beat a casino?
Three things to take away
- Measure turnover, not deposit. Your expected loss is the house edge times everything you wager, and everything you wager is far larger than what you put in.
- Rank every bet by cost per $100 staked. One scale, one number. It cuts through bonus offers, payout badges and review scores in about ten seconds.
- Watch the bets where the edge compounds. Parlays, accumulators, teasers, tie bets and side bets all stack a small edge on top of itself until it becomes a large one. The marketing is loudest exactly where the price is worst.
Want to know whether a specific price is wrong?
That is the only question worth asking, and it is the one the Underdog Chance AI Betting Assistant is built to answer. Give it a game, and it produces a fair price, compares it against the number on the board, and tells you whether there is enough of a gap to bet, or whether the correct answer is to pass.
The Starter Pass gives you 15 full analyses for $7, so you can run it on real games before deciding whether it fits how you bet.
Profit Strategy Revealed
Discover the strategy that took me from broke to pro—see why most lose and winners prevail.