Before the numbers
Casino games are games of chance with a built-in house edge. No strategy, system, pattern or timing changes the odds in the player’s favor over time, and most players lose money over time. Nobody should gamble money they cannot afford to lose.
This article uses only invented numbers to show how payouts work. It reports no real data and makes no claim about any particular game, casino or machine.
Payout ratios
A payout ratio such as 8 to 1 states how much a winning bet earns compared with the amount staked. At 8 to 1, a winning $1 bet earns $8 in profit and the $1 stake is returned, for $9 in total. A losing bet loses the $1 stake.
Some places write the same payout as “9 for 1”, which counts the stake in the figure. The difference between “to” and “for” is a frequent source of confusion, and it is worth checking which convention applies before comparing payouts.
True odds and fair payouts
The true odds of an event come from its probability. If an event has a 1 in 10 chance, the true odds are 9 to 1 against: nine unfavorable outcomes for every favorable one. A fair payout would then be 9 to 1, because in the long run the wins and losses would balance exactly.
Casino payouts are set below the fair level. That gap is the source of the house edge. A hypothetical game shows how it works.
A made-up ten-number game
Imagine a game where a number from 1 to 10 is drawn at random, and a player picks one number. The chance of winning is 1/10.
Suppose the game pays 8 to 1. For a $1 bet:
- With probability 1/10, the player gains $8.
- With probability 9/10, the player loses $1.
Expected value per $1 bet = (1/10 × 8) − (9/10 × 1) = 0.8 − 0.9 = −$0.10.
So on average the player loses 10 cents per dollar wagered, a house edge of 10%.
Now look at it from the return side. A winning bet returns $9 in total (profit plus stake). The average return per $1 wagered is 1/10 × 9 = $0.90. This is the return to player, or RTP, which here is 90%.
The two numbers are linked simply:
house edge = 100% − return to player
A 90% return means a 10% edge, and a 96% return means a 4% edge.
If the game paid the fair 9 to 1, the average return would be 1/10 × 10 = $1.00, a 100% return and no edge. Every cent of shortfall in the payout moves the return below 100%.
A real structural example
A textbook case is single-zero roulette. There are 37 pockets, so a bet on one number has a 1/37 chance. The fair payout would be 36 to 1, since there are 36 unfavorable outcomes. The usual payout is 35 to 1, and the return works out as 36/37 of each stake, which is about 97.30%. The edge is therefore 1/37, or about 2.70%. The payout ratios compared with true odds follow the same pattern across the whole layout of bets.
What a percentage means in practice
Return percentages describe the long-run average across a very large number of bets. They are not a promise about a single session.
Suppose a hypothetical game returns 96% on average, and someone makes 200 bets of $1. The total wagered is $200, so the expected return is 0.96 × 200 = $192, and the expected loss is $8. That is the average. Actual results scatter widely around it: some sessions finish ahead and many finish behind by more than the average. With 100 bets of $1, the average loss would be $4, but the spread of results is usually far larger than that, which is why the amount wagered and the variance matter as much as the percentage. The same arithmetic means that money won early and wagered again becomes part of the total wagered. The percentage applies to every bet, including the ones made with winnings.
Why higher is not the same as good
A higher return is better than a lower one in the narrow sense that expected loss per dollar is smaller. It still does not turn the expectation positive. A game at 98% costs 2 cents per dollar wagered on average. Fast play can multiply the total wagered so quickly that the dollar cost per hour is larger than a lower-return, slower game would cost.
Two other cautions apply.
- Published percentages for slot-style games are normally long-run design figures. They describe the program’s average over an enormous number of plays, so they say little about one person’s evening. This is where hit frequency and volatility come in, because the same return can be delivered through frequent small wins or rare large ones.
- A percentage that is advertised or implied can be framed to look attractive. Framing effects can make a payout look more generous than the expected loss justifies.
Reading payout information carefully
When looking at a game’s published payouts:
- Check whether the payout ratio includes the returned stake.
- Check which bets in a game have which payouts, since a game can have several bets with different edges.
- Remember that the edge is built into the payout; it does not appear as a fee.
- Treat the return figure as an average over a huge number of bets, not as a forecast.
The mathematics here is simple, but it explains a lot: because payouts sit below true odds, the expected result of every wager is a small loss, and no choice of bet changes the sign of that expectation. If gambling is causing distress, a local support service can help.