Walk into any casino discussion and you'll hear the house edge cited like a protection spell. Blackjack is 0.5%. Baccarat banker is 1.06%. Roulette double-zero is 5.26%. Players repeat these figures with confidence, and the figures are accurate—as far as they go. The problem is that the house edge describes something theoretical, and what casinos actually collect from players in the real world is a different number entirely: the hold percentage.
Understanding the difference isn't a semantic exercise. It's the reason players consistently walk away with less than the math they memorized would predict.
What the House Edge Actually Measures
The house edge is the casino's expected mathematical advantage on a single bet, expressed as a percentage of that bet. If a game has a 2% house edge and you bet $100, the casino expects to keep $2 over a large enough sample. That number is precise and reproducible under controlled conditions.
What it does not account for is time, rebetting, or what happens to money once it's won. The house edge is a per-bet figure. Real sessions are not a single bet.
What Hold Percentage Measures
The hold percentage is what the casino actually retains from the total buy-in across a session. Casinos track this internally across tables and games, and it is consistently higher—often dramatically higher—than the house edge would suggest.
A blackjack table with a 0.5% house edge might show a hold percentage of 15–20% over a shift. Slot floors with edges in the 6–10% range sometimes hold 25–30% of total money brought to the machines. These numbers aren't the result of cheating or manipulation. They're the result of a mechanism the house edge formula doesn't capture: recycling.
The Recycling Effect
Here's the mechanism. A player buys in for $200 at a $10 blackjack table. They win early and build up to $280. They continue playing. They lose back to $140. They keep playing. Eventually they leave with $60.
The casino's hold on that session: $140, or 70% of the buy-in.
Now, did the house edge suddenly jump to 70%? No. The house edge stayed at 0.5% per hand. But the player didn't bet $200 once—they recycled their initial buy-in and their winnings repeatedly across dozens of hands. Every dollar they won was fed back into the machine and exposed to the edge again. And again. The 0.5% edge had many more opportunities to do its work than the initial $200 would imply.
This is why the hold percentage is so much higher than the house edge. The edge is applied not just to buy-ins, but to every dollar that cycles through the betting circle, including money the player temporarily won.
Why This Matters Practically
The hold percentage exposes a planning error that is nearly universal among recreational players: anchoring session expectations to the house edge rather than to total handle.
Total handle is the sum of every bet placed during a session, not the buy-in. A player who buys in for $200 and plays 100 hands of $10 blackjack has generated $1,000 in handle even if they never add to their stack. At 0.5% edge, the expected loss on $1,000 in handle is $5—not $1, not $10. If they run hot early and ride a bigger stack for part of that session, handle increases further, and so does expected loss in dollar terms.
Players who budget based on buy-in are underestimating their exposure. Players who budget based on handle are working with the number that actually matters.
Calculating Your Own Handle
Estimating handle is straightforward. Multiply your average bet size by the number of hands or decisions per hour, then multiply by the hours you intend to play.
$15 average bet × 80 hands per hour × 3 hours = $3,600 in handle
At a 1% house edge, expected loss over that session is $36. At 5%, it's $180. The edge percentage that looks small in isolation produces real dollar figures when it's applied to the full handle rather than to a buy-in.
For slots, the same logic applies but the numbers move faster. A machine cycling 600 decisions per hour at $1 per spin generates $600 in handle per hour before a single dollar of winnings is reinvested. At an 8% edge, expected loss is $48 per hour—from what felt like a low-stakes machine.
The Rebetting of Winnings
The component that inflates hold most aggressively is the reinvestment of winnings. When a player wins a hand and leaves the chips in the betting circle, those winnings become new handle. The house edge applies to that amount too. Over a session where someone runs well early, a substantial portion of what they're ultimately losing is money they technically won at some point during the session.
This isn't a psychological illusion. It's arithmetic. The more a player wins mid-session and continues playing, the more total handle they generate, and the more total expected loss accumulates.
The practical implication: winning early in a session does not change the structure of the remaining session. The house edge on every future bet is unchanged. Players who feel that an early profit gives them a cushion to play with are correct in a narrow sense—but they're also generating more handle, which increases total expected loss in dollar terms.
What to Do With This
None of this argues against playing. It argues against using the house edge as a session-loss estimator when it was never designed for that purpose.
A more useful pre-session habit: estimate your handle, apply the edge, and use that figure as your expected-loss baseline. Set a walk-away point based on handle rather than on how your stack compares to your buy-in. Recognize that every hand you continue playing after winning increases the amount of money you've exposed to the edge, not just maintained.
The house edge is an honest number. The hold percentage is what happens when that honest number meets time, recycling, and the human tendency to keep playing as long as there are chips in front of you. Knowing both—and knowing why they diverge—gives you a cleaner picture of what you're actually agreeing to when you sit down.