Slots explained

Slot machine payback at Canadian casinos: what each province actually sets

Casino newsUpdated October 28, 2026

Only a few Canadian jurisdictions publish a number. Ontario's technical standard sets a minimum theoretical payout of 85% for every slot game, Nova Scotia's casino regulations set 86%, and BCLC puts the lifetime average on B.C. machines above 92%. For the other provinces we found no published floor. None of these figures predicts what you will get back in a single visit.

What payback actually measures

Payback, also called return to player, is the share of all money wagered on a game that the game is designed to hand back as prizes over its whole working life. The rest is the hold, the portion the casino keeps. OLG's PlaySmart page on slot odds puts it plainly: machines arrive with a chip that pre-determines the payback over the machine's lifetime, the house sets the machine to pay back a percentage of what is played, including the jackpot, and the amount kept is called the hold.

The key word is lifetime. On PlaySmart, that lifecycle is one that can run to many millions of spins. A payback figure is a property of the game's maths, not a promise about the next hour. We cover the odds side of this in more depth in our guide to how slot odds work; this piece sticks to what Canadian regulators and lottery corporations have actually put in writing.

Province by province: the published numbers

Payback rules in Canada are set province by province, and they are not published in the same form. Some appear as a minimum in a technical standard or a regulation, some only as an average quoted by the operator, and several provinces publish nothing we could find. The table keeps those categories apart, because a floor and an average are very different promises.

Two details in the Ontario standard are worth knowing. The 85% floor applies to every bet configuration a game offers, not just the most generous one, although the standard's guidance tolerates a decrease of up to 0.5% in return for a higher bet. And where a game's paytables differ by more than 4% in return between wager categories, separate meters have to track them. In other words, a machine can legitimately return slightly different amounts depending on how you bet, but none of those options may fall below the floor. The Nova Scotia rule applies to the province's two casino properties, including Casino Nova Scotia.

The B.C. and Alberta figures are a different kind of number. BCLC's 92% is an average across machines, so individual games can sit above or below it. AGLC's 92% describes the VLT network as a whole, with programmed rates such as 92.00% or 91.89% that vary with the number of lines, the denomination and the maximum bet. Neither is a guarantee that any single machine pays 92%.

Payback is not what you take home from one visit

Every source that publishes a figure also warns against reading it as a session forecast. On PlaySmart, OLG is explicit that an 85% payback does not mean a player who puts in $100 will walk away with $85, and that payback does not affect short-term results. BCLC's FAQ makes the same point about its 92%: it does not mean you will leave with 92% of the money you started with. AGLC's page for VLT players frames its rate as an average over millions of plays, not over one session.

The reason is variance. A game's return is built from a long list of outcomes, from frequent small wins to very rare top awards, and a large share of the theoretical return can sit in prizes that almost never hit. GameSense, BCLC's player-information program, gives the example of a penny machine with a 1 in 12 chance of any prize but a 1 in 649,400 chance of the top prize. Ontario's standard caps top-award odds at 17 million to 1 unless the odds are disclosed, and PlaySmart puts the odds of a top award as low as 1 in 17,000,000 spins. Over a few hundred spins, whether one of those rare prizes lands shapes the result far more than the payback percentage does.

GameSense also offers a rough comparison: a typical average house advantage of 8% for slot machines, against 0.5% for blackjack, 1.06% for baccarat and 5.3% for roulette in its own table. It adds that machines can be programmed to pay out at higher or lower odds, so 8% is an average, not a setting.

Why a machine cannot be made tight on demand

A persistent casino myth holds that staff tighten machines on busy nights and loosen them to lure players back. The published rules point the other way, at least in the provinces that set them out.

• No reacting to results. Under the AGCO standard, devices must not alter any function based on the actual hold percentage, and features that automatically adapt a game's payout are not permitted.
• Changes are restricted and logged. Payout percentage, lines and denominations are critical game options that only authorized personnel may set or change, through a restricted technical procedure, and configuration changes are recorded as significant events. Slot monitoring systems must not allow critical configurations to be altered.
• No switches. Nova Scotia's casino regulations prohibit hardware switches that alter pay tables or payout percentages, and program storage cannot be altered except by procedures the Executive Director has approved. The Executive Director also carries out random testing of machines.
• Not mid-play, not for weekends. On PlaySmart, OLG points out that it does not have complete control over machine settings because the AGCO regulates them, and that casinos are not allowed to change a machine's payout while someone is playing or winning on it, or to adjust machines for weekends.

Each spin is also independent. In GameSense's explanation, a random number generator produces thousands of numbers per second and the outcome is fixed at the instant the game is activated. AGLC makes the same point for VLTs: play patterns, the time of day and a machine's position have no bearing on results. A machine that has not paid for an hour is not due, and one that just paid is not cold.

Using payback information on a casino floor

Because the published figures are floors or averages, their practical value for a player is limited but real.

1. Read the paytable. PlaySmart cautions that a higher denomination does not always mean a higher payout in Ontario, so the paytable on the game is a better guide than the coin size.
2. Treat the floor as a floor. An Ontario game cannot be configured below 85% for any bet configuration, and a Nova Scotia machine must meet 86% unless the Executive Director approves otherwise. That rules out the very worst settings but tells you nothing about which machine on the floor sits higher.
3. Size your budget to the swings, not the return. Even on a game that keeps a modest share over its lifetime, a short session can lose much more than that share of what you bring, so decide the amount in advance and stop when it is gone.
4. Separate jackpot games. Part of a game's theoretical return may be tied up in a top award you are unlikely to see. We explain how that works for linked and standalone progressive jackpots in a separate guide.

The gaps in the public record

For Manitoba, Saskatchewan, New Brunswick, Prince Edward Island and Newfoundland and Labrador, we could not locate a published payback minimum or average for slot machines. In Quebec, the most specific number we found is a review by the Institut national de santé publique du Québec citing an average machine return of 92%, which is research rather than a current rule. For Alberta there is a published 92% average for the VLT network, but we found no equivalent figure for casino slot floors.

That does not mean machines in those provinces run without limits. It means the limits, if there are any, are not published in a form we could find and cite. If you want a figure for a particular venue, the casino's responsible-gambling information centre is the place to ask, and our guide to land casinos in Canada lists the main properties by province.

Frequently asked questions