House edge and RTP describe the same fact from opposite ends. When the stake, the payout basis and the rules are held fixed, RTP plus house edge equals 100 percent in theory. If a game’s return to player is 96 percent, its house edge is 4 percent; if the edge is 2.7 percent, the RTP is 97.3 percent. The two numbers are not rivals, but they answer different questions, and mixing them up leads to some persistent misunderstandings about what a percentage can tell you.
The same coin, two faces
RTP asks: out of all the money wagered, how much comes back to players? House edge asks: how much stays with the operator? Both are long-run, theoretical averages, calculated over an enormous number of identical bets. Neither says anything certain about a single session, a single evening or a single spin.
The phrase “same fixed stake and payout basis” matters more than it looks. Both figures only add to 100 percent when they are measured against the same denominator. Imagine a hypothetical game where every bet is a flat 10 units and every win is paid in cash. Over a million plays, 10 million units are wagered. If 9.6 million come back as winnings, RTP is 96 percent and the edge is 4 percent. The moment the denominator shifts, say because some outcomes return the stake plus a prize while others return only a prize, you need to be careful about which pot you are dividing by.
Gross return versus net profit
Two terms need defining before going further. Gross return is everything paid back to you on a winning bet, including the stake itself. Net profit is gross return minus what you staked. A hypothetical bet of 10 units that pays “even money” returns 20 units gross: your 10 back plus 10 in winnings. Net profit is 10 units. Lose the bet and gross return is zero, net result is minus 10.
RTP is conventionally framed in terms of gross return per unit staked. A 96 percent RTP means that, on average, each 100 units wagered produces 96 units of gross return and a 4-unit average shortfall. That shortfall is the edge. If you confuse gross with net, you might read “96 percent back” as “I keep 96 percent of my money,” when the more accurate reading is “for every 100 I put through the game, 96 passes back to me in payouts, on average.”
Strategy-dependent games change the question
For a pure chance game with no decisions, such as a slot or a roulette bet, the RTP is fixed by the rules and the paytable. You cannot play it better or worse. For games that involve decisions, RTP is not one number at all. It depends on how you play.
Blackjack is the classic case. Quoted house edges for blackjack usually assume a specific rule set and a specific playing approach, typically what is called basic strategy, meaning the mathematically optimal action for each combination of your hand and the dealer’s upcard. A figure such as “about half a percent” only holds if several assumptions hold at once: the number of decks, whether the dealer stands or hits on soft 17, what blackjack pays, whether doubling after splitting is allowed, and whether you actually follow the optimal play every hand. Change any one and the edge moves. Pay blackjack at 6 to 5 instead of 3 to 2 and the edge rises noticeably, because the most valuable hand is paid less.
So for strategy-dependent games, a published edge is better read as a conditional statement: if these rules and if optimal play, then approximately this long-run cost. A player who guesses at decisions, or who takes insurance as a habit, faces a higher effective edge than the headline number. The reverse is not available: no strategy pushes the edge below what the optimal-play calculation gives for a fixed shuffled shoe and a fixed rule set.
An original calculation: single-zero roulette
Roulette gives the cleanest worked example, because every probability is countable. Take a hypothetical single-zero wheel with 37 pockets: the numbers 1 to 36 plus one zero. A straight-up bet on one number pays 35 to 1, meaning 35 units of profit on a 1-unit stake, plus the stake returned.
The probability of winning is 1 in 37. The probability of losing is 36 in 37. Compute the expected net result per unit staked:
Expected net = (1/37 × 35) + (36/37 × −1) = 35/37 − 36/37 = −1/37, which is about −0.0270, or −2.70 percent.
That 2.70 percent is the house edge. In gross terms: a win returns 36 units (35 profit plus the 1-unit stake), so expected gross return per unit staked is 1/37 × 36 = 36/37, about 97.3 percent. RTP is 97.3 percent and the edge is 2.7 percent, adding to 100. Notice also that nearly every standard roulette bet gives the same result, because payouts are set against the true odds of 36 to 1 on a 37-pocket wheel while paying one unit less. A bet on red covers 18 of 37 pockets and pays even money: (18/37 × 1) + (19/37 × −1) = −1/37 again. Betting differently changes volatility, not the long-run cost per unit.
For a feel of scale, suppose that over a hypothetical evening someone places 200 bets of 5 units each. Total turnover is 1,000 units. The theoretical expected cost is 2.7 percent of 1,000, or 27 units. The actual outcome will wander well away from that: some evenings end well ahead, others lose several times the expected figure. The edge describes the center of the distribution, not its width.
Turnover, not deposit
This leads to the most commonly misapplied idea in the whole topic. House edge applies to turnover, the total amount wagered, not to the amount you deposit. Money is recycled. A hypothetical player deposits 100 units and plays roulette on red in 5-unit bets. If they win a bet, the returned stake and winnings go straight back into play. By the end of the session they may have wagered 600 units in total, even though only 100 ever left their bank balance.
Applying the 2.7 percent edge to the 600 units of turnover gives an expected cost of about 16.2 units, not 2.7 units. The cost grows with how much you put through the game, not with the size of the original deposit. That is why the same edge feels mild across a few bets and heavy across a few thousand. A 2.7 percent cost on 100 units of turnover is 2.7 units; on 10,000 units of turnover it is 270, and a player can reach that turnover while only ever depositing a modest sum, simply by recycling winnings.
The practical reading is this: edge multiplied by turnover estimates the expected loss, and turnover depends on bet size, speed of play and time at the table. A lower-edge game played quickly can cost more per hour than a higher-edge game played slowly. Hands per hour and spins per hour belong in any honest comparison.
What the percentages can and cannot tell you
An RTP figure or house edge is a statement about averages over very large samples, derived from the rules or the game’s mathematical model. It cannot predict your next result, it does not drift toward 100 percent in the short run, and it is not a promise about any individual session. When you see a figure attached to a particular game, ask what it assumes: the rule variant, the type of bet, the style of play and the denominator. Those assumptions, more than the percentage itself, decide what the number means.