The decision you're making at the insurance prompt
You have a strong hand, the dealer shows an ace, and the table offers you insurance. The surface question is whether to protect your original bet. The underlying question is whether a side wager with its own built-in mathematical disadvantage is worth making right now. Those are two different questions, and the second one is the only one that changes your long-run result.
Insurance is a separate bet, not a modification of your main hand. You place up to half your original wager on a small marked area of the layout. The dealer then checks the hole card. If the dealer has blackjack, the insurance bet pays 2 to 1. If the dealer does not have blackjack, the insurance bet is collected and the hand continues. Your original bet is settled normally either way.
What makes insurance unusual is that it is one of the few bets in a casino where the outcome is already heavily constrained before you act. You can see one dealer card, and you can see your own cards. The only unknown is the hole card. That means the decision can be analysed with unusual precision, which is exactly what card counters do.
How the house edge is built into the insurance bet
Insurance pays 2 to 1, but the true odds of the dealer having a ten-value card underneath an ace depend on the composition of the remaining deck. In a freshly shuffled game with a full deck, the dealer's hole card is a ten-value card roughly 4 times out of every 13, because there are 16 ten-value cards in a 52-card deck and 13 ranks total. A fair bet on that event would pay slightly better than 2 to 1. The 2-to-1 payout is therefore short of the true odds, and that shortfall is the house edge on the insurance bet.
That edge is not small. The insurance side bet carries a materially higher house edge than a basic-strategy blackjack hand under common rule sets. You are not protecting anything by taking it; you are placing a second, worse bet alongside a first one. The main hand's expected result does not change whether you take insurance or not. Only the side bet's expected result is added on top.
What the count changes
Card counting works because removing cards from the deck changes the proportion of ten-value cards remaining. When the deck is rich in tens and face cards, the dealer is more likely to have a ten in the hole, and the insurance bet becomes less bad. At some point it crosses from negative expectation to positive expectation. That crossing point is what a counter is tracking when the insurance prompt appears.
The arithmetic is a conditional probability problem. Suppose, as an example, that a counting system assigns a value to each visible card and the running count is divided by an estimate of decks remaining to produce a true count. If the true count is high enough that the proportion of ten-value cards remaining exceeds one in three, the insurance bet's expected value turns positive. Below that threshold, it remains negative. The exact threshold depends on the counting system and the number of decks, which is why published counting references give a specific index number for the insurance decision.
For a player who is not counting, none of this is knowable with enough precision to change the decision. Without a reliable estimate of the remaining deck composition, the insurance bet is simply a side wager with a built-in disadvantage, and the basic-strategy answer is to decline it every time.
RTP and house edge on the same bet under one rule set
Return to Player (RTP) and house edge are complements: they sum to 100% for the same bet under the same rule set. If a side bet returns a small fraction of wagers to players over a very large number of trials, the house edge on that same bet is the remainder. The two numbers describe the same relationship from opposite sides. RTP is the share returned; house edge is the share retained.
This relationship holds only for the same bet under the same rules. You cannot take the RTP of a basic-strategy blackjack hand and subtract it from the house edge of the insurance side bet to get anything meaningful. They are different bets with different payout structures and different probabilities. The insurance prompt is a moment when the table offers you a second product with its own separate RTP, and that product's RTP is worse than the main game's under normal deck conditions.
Why short sessions look nothing like long-run averages
A single insurance decision is one trial. The house edge on insurance expresses itself over tens of thousands of identical decisions, not over the three or four insurance prompts you see in an evening. In a short session, you may take insurance twice, win both times, and walk away convinced the bet is profitable. Or you may decline it every time and watch the dealer turn over blackjack on four consecutive aces. Neither outcome tells you anything about the bet's true expectation.
Variance is the reason. A 2-to-1 payout on an event that occurs roughly one time in three produces large swings in either direction over small samples. The long-run average is real, but it is approached slowly, and a player's memory of short sessions is a poor estimator of it. This is why the insurance bet survives in casinos: it occasionally pays, and those occasional wins are memorable enough to keep the bet popular despite its mathematical disadvantage.
Decisions per hour and total exposure
Every insurance bet you place adds a separate wager to your total action. If you are playing 80 hands per hour and the dealer shows an ace on a fraction of those hands, each insurance decision is an additional bet of up to half your original stake. Over a long session, those side bets accumulate into a meaningful amount of total money wagered, and the house edge on that accumulated amount is applied regardless of what your main hand does.
This is the quiet cost of side bets generally. A player who would never knowingly sit at a game with a high house edge will happily place a side bet with that same edge because it arrives as a small, occasional prompt attached to a game they already trust. The insurance decision is not about the hand you are holding. It is about whether you want to add a second, worse bet to your total exposure. The answer for a non-counting player is almost always no, and the reason is arithmetic, not caution.
Why progression systems do not fix the insurance bet
Some players treat insurance as part of a progression: take it after losses, double it after wins, or use it to "lock in" a profit on a strong hand. None of these patterns change the underlying probability of the dealer's hole card. A progression system redistributes outcomes across a session, but it does not alter the expected value of any individual bet. If the insurance bet has a negative expectation, a progression of insurance bets has a negative expectation equal to the sum of its parts.
The same logic applies to the even-money offer that appears when a player has blackjack and the dealer shows an ace. Even money is insurance in disguise: it is a guaranteed 1-to-1 payout on the main bet instead of the usual 3-to-2, funded by the same side-bet mechanics. The decision is identical to the insurance decision, and the same counting threshold applies. A player who declines insurance but takes even money is making the same bet twice under different names.
What a counter is doing at the prompt
A card counter treats the insurance prompt as a signal, not as a protection offer. The decision is made before the dealer asks, based on the running count and the number of decks remaining. If the true count is at or above the system's insurance index, the counter takes insurance at the maximum allowed amount. If it is below the index, the counter declines without hesitation. The strength of the counter's own hand is irrelevant to the decision, because the insurance bet is settled independently of the main hand's outcome.
This is a useful discipline for any player to understand, even one who never counts. The insurance decision is one of the few moments in blackjack where the correct play is entirely separable from the cards in your hand. The only question is whether the remaining deck makes the side bet's expected value positive. If you cannot answer that question with a reliable count, the default answer is no.
Reading the table rules before you sit down
Insurance payouts and availability are stated in the table rules, which are printed on the layout or available from the dealer. The standard payout is 2 to 1, but some tables offer variations, and the exact wording of the insurance rule affects the bet's house edge. The number of decks in play also matters, because it changes the conditional probability of the dealer's hole card and therefore the counting index at which insurance becomes profitable.
Before you rely on any counting threshold, confirm the deck count and the insurance payout at the table where you are playing. A threshold calculated for a six-deck game does not transfer cleanly to a double-deck game, and a threshold calculated for a 2-to-1 payout does not transfer to a different payout. The arithmetic is specific to the rule set in front of you.
FAQ
Does taking insurance protect my original bet?
No. Insurance is a separate side bet that pays if the dealer has blackjack. Your original bet is settled on its own terms regardless of whether you take insurance. If the dealer has blackjack and you do not take insurance, you lose your original bet. If you do take insurance, you win the side bet but still lose the original bet unless you also have blackjack, in which case the main bet pushes.
Is insurance ever a good bet for a non-counter?
Under normal deck conditions, the insurance side bet carries a higher house edge than the main blackjack game played with basic strategy. Without a reliable count of the remaining deck, the player has no way to know whether the bet's expectation has turned positive. The basic-strategy answer is to decline insurance every time, and the even-money offer is the same decision in different clothing.
What count makes insurance worth taking?
The threshold depends on the counting system, the number of decks, and the insurance payout. In general terms, the bet becomes positive when the proportion of ten-value cards remaining in the deck is high enough that the 2-to-1 payout exceeds the true odds of the dealer's hole card. Published counting references give a specific true-count index for the insurance decision under standard rule sets, and that index should be confirmed against the table rules before play.
Gamble responsibly. If you or someone you know has a gambling problem, call the helpline printed in the footer below.
Play responsibly
Gambling should be treated as paid entertainment, never as a way to earn income or recover losses. 1-800-MY-RESET (1-800-697-3738)
18+ or 21+ depending on where you are; follow the minimum age that applies to you.
This article is general information about how these mechanics work. It is not legal advice and not a recommendation to gamble or to use any particular operator. Availability and legality differ by jurisdiction — check the rules that apply where you are.