Blackjack Insurance: The Real Math Behind the Casino's Worst Bet

Blackjack insurance is the oldest side bet in the game, old enough that it predates the phrase “side bet” entirely. It has been silk-screened onto felts since the 1950s, dressed up as a courtesy: the dealer shows an ace, the table tenses, and the layout offers to protect your hand for half your wager. What it actually offers is a separate proposition on the dealer’s hole card, priced so the house keeps about 7.4 cents of every dollar that crosses the line.

That makes it worse than almost anything you can do with the main bet.

What blackjack insurance actually is

The mechanics take one sentence: when the dealer’s upcard is an ace, you may bet up to half your original wager that the hole card is a ten, jack, queen or king, and the bet pays 2 to 1 if you’re right. Dealer has blackjack, insurance cashes while your main bet dies, and the round nets out near zero. Dealer doesn’t, the premium disappears and the hand plays on.

Notice what is missing from that description: your cards. Blackjack insurance never looks at them. It is not insurance in any actuarial sense, because nothing about your hand is covered; a 20 and a stiff 15 buy exactly the same policy at exactly the same price. The industry name is a marketing decision that has survived seventy years because it works. Players who would never bet on a coin flip at bad odds will happily “protect” a good hand, which is the same transaction with a comforting label on it.

The wager showed up on Las Vegas layouts in the 1950s and then spread to every jurisdiction that deals the game, outliving every side-bet fashion since. It survives because it earns, and because it is cheap to run. The modern propositions need a progressive meter, a licensing deal with a distributor and a trained pitch; insurance needs one line of paint and a moment of fear. It collects at the exact instant the table feels most vulnerable, which is a better salesman than any jackpot.

Structurally it belongs on the full side-bet board with everything else we grade on this site: a fixed-odds proposition, resolved before strategy starts, with a built-in gap between the payout and the probability.

The math: 30.8% against a 33.3% break-even

A 2 to 1 payout breaks even when the bet wins exactly one time in three. So the whole question is: how often is the dealer’s hole card worth ten?

Four of the thirteen ranks are. That is 4/13, or 30.77%, and no amount of table feel changes it. On a six-deck game, once the dealer’s ace is accounted for, 96 of the 311 unseen cards are tens, which nudges the figure to 30.9%. Either way you are buying a one-in-three payout on a bet that wins less than one time in three, and the shortfall compounds into the house edge. Wizard of Odds prices it at about 7.4% for six decks, roughly 5.9% for a single deck and 7.5% for eight, and this page’s headline number follows that schedule.

Put dollars on it. Suppose you bet $20 a hand and insure for the full $10 every time an ace shows, and suppose the ace shows a hundred times. About 31 of those hands cash the insurance for $20 profit each, roughly $620 coming in. The other 69 hand the house your $10 premium, about $690 going out. You are down some $70 on $1,000 of insurance action without a single unlucky thing happening. That is the 7.4% expressed in chips instead of decimals, and it never improves with volume.

Your own hand usually makes it worse, not better. The hands players most want to insure are strong ones, and strong hands tend to contain ten-value cards. Every ten you hold is a ten that cannot be in the hole. Insuring your 20 means paying the posted price for odds that are slightly longer than the posted price already fails to cover.

The gap sounds small when written as 30.8 versus 33.3. Per dollar wagered it is enormous: a good six-deck main game runs a house edge of around half a percent with sound play, so the insurance line bleeds money about fifteen times faster than the game it sits on. The working blackjack odds look forgiving by comparison almost everywhere else on the layout.

Even money is the same bet in a different jacket

The one moment insurance gets genuinely persuasive is when you hold a natural and the dealer shows an ace. The dealer offers “even money”: a guaranteed 1 to 1 payout right now, instead of risking the push. Most of the table will tell you to take it. Most of the table is wrong.

Run the two branches of insuring a blackjack for half your bet. Dealer has the ten: your natural pushes, insurance pays 2 to 1 on half a bet, and you collect exactly one bet. Dealer doesn’t: your natural pays 3:2, the half-bet premium is gone, and you collect exactly one bet. One bet, guaranteed, either way. Even money and insuring your own blackjack are the same wager; casinos simply folded the arithmetic into a single phrase that sounds like a favor.

Declining is better because the push only happens 30.8% of the time. The other 69.2% you collect the full 3:2, which averages out to about 1.04 bets per blackjack against even money’s flat 1.00. Four cents on the dollar, surrendered for certainty. Casinos train dealers to announce the offer cheerfully and settle it fast, because certainty is the easiest product on the floor to sell.

If you want one blunt rule: never take even money. It is the same donation as insurance with better lighting.

When counters actually take blackjack insurance

Everything above assumes you know nothing about the shoe’s composition. Card counters know something, and blackjack insurance is where that knowledge pays best.

A running count tracks the ratio of tens and aces to small cards remaining. When the shoe grows ten-rich, the probability that the hole card is a ten climbs past 4/13, and at some point it crosses the 33.3% break-even and the bet flips positive. In Hi-Lo the standard trigger is a true count of +3. Counting literature going back to Edward Thorp’s Beat the Dealer in 1962 treats the insurance index as the single most valuable deviation in the game, worth more than any hit-or-stand departure in the Illustrious 18.

Which produces a delicious irony. The bet the casino advertises in the middle of the felt is the one wager a skilled player can most reliably turn against it, and pit surveillance knows this. A player who insures a stiff 16 for exactly half a bet at a deep, ten-rich shoe is waving a flag; insuring garbage hands makes no sense except through a counter’s eyes, and some professionals deliberately skip marginal insurance calls as camouflage. You, playing without a count, get none of this. The index belongs to people doing arithmetic on every card that hits the discard tray.

Why basic strategy says never

Basic strategy is the set of plays that maximizes your return with no information beyond the visible cards, and on insurance its verdict has never wavered: decline, every time, with every hand. There is no hand strength that changes the hole card’s odds, no bet size that improves the price, and no streak that makes 30.8% into 33.4%.

Two popular workarounds deserve a quick burial. “Insuring for less” changes the amount leaked, not the rate; a $5 premium loses 7.4% of $5 with the same certainty as a $10 premium loses 7.4% of $10. And “I only insure my strong hands” is the ten-removal problem from earlier wearing a strategy costume: the strong hands are the ones holding the cards the bet needs. The single defensible case is a table-maximum bettor who genuinely cannot stomach the swing of losing a monster hand to a dealer natural; for that player insurance is a variance-reduction fee, purchased knowingly at 7.4%, the way a business buys real insurance it expects to lose money on. If that sentence does not describe your bankroll, it does not describe your decision either.

What actually drives most insurance bets is mental accounting. A player with a good hand feels they have something to lose, so half a bet feels like a fair premium against disaster. But the disaster and the premium are independent events. You are not protecting the 20; you are placing a new, separate, overpriced wager while feeling prudent about it. The casino collects either way, one line at a time, roughly 7.4 cents per insured dollar.

Skip it. Take the money you would have fed the insurance line over a year of sessions and it will cover more than a few of the pushes that even money was supposed to spare you.

One last detail worth noticing next time you sit down: “Insurance Pays 2 to 1” is the only bet whose full terms are printed in an arc across every standard blackjack layout on earth, positioned where no seat can miss it. Casinos do not buy that real estate for bets that lose them money.