Every blackjack player has heard it. The dealer flips an Ace, pauses with a practiced smile, and asks, "Insurance, anyone?" If you're holding a blackjack, the question becomes friendlier: "Even money?" Both offers sound like caution. Both are among the most expensive decisions you can make at the table.
I've watched disciplined basic strategy players give back hours of careful play to this one moment. They split correctly, double correctly, and stand on soft 18 when the chart says so. Then they buy insurance and hand the casino a 5.88% to 7.47% house edge on a bet that has nothing to do with the hand in front of them. This guide covers why insurance is a sucker bet, why even money is the same bet under a different name, and the one narrow situation where a skilled counter should take it.
The Psychological Illusion: A Side Bet Wearing a Seatbelt
The word "insurance" is the casino's strongest marketing tool. In everyday life, insurance means paying a small premium to avoid a catastrophic loss. Homeowners buy it. Drivers are required to carry it. Responsible people insure what they value. The casino counts on you carrying that association to the felt.
Blackjack insurance does not protect your hand. It is a separate side bet on a single question: does the dealer's hole card have a value of ten? You can bet up to half your original wager. If the dealer has blackjack, the side bet pays 2:1. If not, you lose it, and your original hand plays out exactly as it would have without the side bet.
Look at that framing honestly. Your hard 20, your soft 19, your stiff 16: none of them changes the insurance bet's odds in any meaningful way. A player holding 20 who buys insurance "to protect a strong hand" is making the same wager as a player holding 12. The only thing being decided is the dealer's hole card. The bet is merely placed next to a hand you care about.
Key Takeaway
Insurance is not risk management. It is a standalone bet on whether the dealer's hole card is a ten-value card. Your own hand's strength does not change the math. When you see it that way, you're no longer being sold peace of mind. You're being offered a 2:1 payout on an event that happens less than one-third of the time.
The Combinatorial Proof: Why 2:1 Isn't Fair
The math here is short, so you can check it yourself at the table. For a bet paying 2:1 to be fair, the event must occur exactly one time in three. That means the remaining cards need a non-ten to ten ratio of 2.0 to 1. Any ratio above 2.0 means the casino is underpaying you.
A standard deck has 16 ten-value cards (10, J, Q, K in four suits) and 36 non-tens. That's a 2.25:1 ratio before a single card is dealt, already worse than the 2:1 you need. The dealer's upcard is a known non-ten (the Ace), so removing it improves the ratio slightly, but never enough.
Take a 6-deck shoe: 312 cards, 96 tens, 216 non-tens. Remove the dealer's Ace and you have 215 non-tens to 96 tens, a ratio of 2.24:1. That's the basis for the standard 7.39% figure. If your own two cards are also non-tens, say a 9-7, then 213 non-tens remain against 96 tens, a 2.219:1 ratio. That's slightly better for the insurance bet, but still well short of the 2.0:1 needed to break even. Holding tens makes it worse, because you've pulled tens out of the pool that could be under the dealer's Ace.
Insurance's expected value is simple: tens remaining times 2, minus non-tens remaining, divided by total cards. With the dealer's Ace removed in a 6-deck game, that's (96 × 2 − 215) ÷ 311 = −7.39%. For the full proof, including how removing the player's cards shifts the figure, see the insurance math proof on BlackjackMath.org. Here, I'll focus on how it plays at the table.
House Edge on Insurance by Deck Count
Fewer decks means each removed card has a larger effect, which is why single-deck insurance is slightly less bad. Note the wording: less bad. Not good.
| Decks | Tens Remaining | Non-Tens Remaining (Dealer Ace Removed) | Non-Ten : Ten Ratio | Insurance House Edge |
|---|---|---|---|---|
| 1 | 16 | 35 | 2.19 : 1 | 5.88% |
| 2 | 32 | 71 | 2.22 : 1 | 6.80% |
| 6 | 96 | 215 | 2.24 : 1 | 7.39% |
| 8 | 128 | 287 | 2.24 : 1 | 7.47% |
| Fair Bet Requirement | — | — | 2.00 : 1 | 0.00% |
Put that in perspective. A good 6-deck game with solid rules gives the house around 0.5% against perfect basic strategy. Insurance in the same shoe carries an edge roughly fifteen times larger. On a $100 hand, a full $50 insurance bet costs about $3.70 in expected value. That's about the same as playing seven or eight full hands of basic strategy at $100 each. One reflexive "yes" can cost what an entire round of careful play earns back. You can run your own game's numbers on the house edge calculator on BlackjackMath.org and see how small the base edge is compared to this one bet.
Warning:
Watch for the theater around the insurance offer. Dealers are often trained to pause, sweep a hand across the layout, and look at each player in turn. This creates social pressure and a sense that "everyone else is protecting themselves." The longer the pause, the more players buy in. Decide before the Ace appears. Your answer is "no insurance," and you can give it with a flat wave before the dealer finishes asking.
The Even Money Trap: Same Bet, Better Disguise
Even money catches far more players than plain insurance, because it targets your best moment. You've been dealt a natural, which normally pays 3:2. The dealer shows an Ace and offers you a guaranteed 1:1 payout right now, before checking the hole card. It feels like locking in a win. Many otherwise solid players take it every time, reasoning that a sure profit beats a gamble.
Here's what the casino doesn't say: even money is insurance. Run both scenarios with a 1-unit bet on your blackjack:
- You buy insurance (0.5 units): If the dealer has blackjack, your natural pushes (0) and insurance pays 2:1 (+1.0). Net: +1.0 unit. If the dealer doesn't have blackjack, your natural pays 3:2 (+1.5) and you lose the insurance (−0.5). Net: +1.0 unit.
- You take even money: You're paid 1:1 immediately. Net: +1.0 unit.
The outcomes are identical in every case. Even money is simply insurance on a blackjack with the bookkeeping done for you. The casino gives it a friendlier name, and that framing works.
The Expected Value Calculation
Now compare taking even money with declining it. Taking it yields exactly +1.00 unit, every time, no variance.
Declining it means you either push (if the dealer has blackjack) or collect 3:2 (if the dealer doesn't). With your Ace and ten and the dealer's Ace removed from play, the dealer completes a blackjack roughly 30.8% of the time. You get paid 3:2 the other 69.2% of the time:
0.692 × 1.50 = +1.038 units
That 0.038-unit gap is the cost of the guarantee. Taking even money gives up 3.8% of your bet in equity on every natural you have against a dealer Ace. That loss comes out of the hand with the highest payout in the game, the one that's supposed to fund the rest of your play.
| Decision (1-Unit Blackjack vs. Dealer Ace) | Dealer Has Blackjack (~30.8%) | Dealer No Blackjack (~69.2%) | Expected Value |
|---|---|---|---|
| Take Even Money | +1.00 | +1.00 | +1.000 |
| Insure the Blackjack | +1.00 | +1.00 | +1.000 |
| Decline Even Money | 0.00 (push) | +1.50 | +1.038 |
The usual objection is: "I'd rather have a guaranteed win than risk a push." That reasoning makes sense once in your life. Blackjack isn't played once. You'll get a natural against a dealer Ace dozens or hundreds of times over a playing career. Over that run, the 3.8% gap shows up in your results. The players who decline even money finish ahead of the players who take it, and the difference comes from no other source than this decision. For how this fits into everything else you do against an Ace upcard, see my Playing Against Dealer Ace guide.
Sharp Tip (Marcus Vance):
If declining even money makes you uneasy, think in terms of 100 identical naturals rather than one. Take even money every time and you collect 100 units. Decline every time and you collect about 103.8 units. Nobody is asking you to gamble for fun. You're choosing the larger total. The push stings in the moment. The extra 3.8 units are permanent.
The Counter's Exception: When Insurance Turns Positive
There is exactly one situation where insurance is a good bet, and it belongs to card counters. Insurance is fair when the remaining cards are exactly one-third tens. In a neutral shoe, tens make up about 30.8% of the cards. As low cards (2 through 6) leave the shoe, the share of tens rises. Once it passes one-third, the 2:1 payout becomes favorable, and the side bet turns into an advantage for the player.
For Hi-Lo counters in multi-deck shoe games, the standard insurance index is a True Count of +3 or higher. At that point, take insurance on any hand, including your blackjack, which means taking even money is also correct at +3 and above. In single-deck games the index drops lower, because each card has more effect. Among all count-based playing deviations, insurance is consistently ranked the most valuable, ahead of standing on 16 against a 10. That's because a counter's large bets go out in exactly the ten-rich shoes where insurance turns positive.
This exception is narrow, and it requires an accurate count. You need a running count tracked correctly through the shoe, converted to a true count by dividing by the remaining decks, at the moment the dealer flips an Ace. "The last few hands felt like a lot of small cards" is not a true count. If you aren't counting, you have no information that shifts the odds, and the shoe should be treated as neutral.
Key Takeaway
For basic strategy players, insurance is always a sucker bet, and so is even money. No exceptions based on hand strength, bet size, mood, or what happened last shoe. The only legitimate trigger is a verified True Count of +3 or higher in a shoe game, and that applies only if you're actually counting.
How Casinos Build on This Mistake
Insurance is the original blackjack side bet, and it trained players to accept the modern ones. Perfect Pairs, 21+3, Lucky Ladies, and Royal Match all use the same approach: a separate wager placed next to your main hand, a payout that looks generous, and a house edge many times higher than the base game. Once a player is comfortable buying insurance, the next side bet is easy to sell. If you want to see how any side bet on your table compares, run it through our Side Bet Auditor. Most of them look as bad as insurance once the numbers are laid out.
The broader lesson is that table selection and bet selection both matter. A game with good rules (3:2 blackjack, dealer stands on soft 17, double after split allowed) loses its value if you keep making −7% side wagers on it. Use the Table Grader tool to evaluate the rules before you sit down. Then protect the edge you've chosen by turning down every bet that hands it back.
Warning:
Expect pushback from other players, and occasionally from dealers, when you decline even money. "Take the sure thing!" is common table folklore. It's also wrong. Table advice is free and is worth what you pay for it. Your strategy comes from the math, not from the person on third base.
Your Table Protocol
Here's the routine to commit to memory before your next session:
- Dealer shows an Ace, you hold any non-blackjack hand: Decline insurance. Wave it off immediately.
- Dealer shows an Ace, you hold a natural: Decline even money. Accept the push risk in exchange for the 3:2 payout.
- You're counting and the True Count is +3 or higher (shoe game): Take insurance for the full half-bet, and take even money on naturals.
- Any other situation: Return to step one.
That's the whole decision. It's one of the simplest corrections in blackjack, and one of the most profitable, because it removes a recurring loss with no added effort. You don't need to memorize index tables or practice deviations. You only need to recognize a sucker bet and say no.
Bottom Line
Insurance is a side bet on the dealer's hole card. It pays 2:1 on an event that needs a 2.0:1 non-ten-to-ten ratio to break even, while real shoes run around 2.2:1 or worse. That produces a house edge of 5.88% in single deck, 6.80% in double deck, 7.39% in six decks, and 7.47% in eight decks. Even money is the same bet with friendlier wording: it locks in +1.00 unit when declining is worth about +1.038 units, giving up 3.8% of your bet on your best hands. Unless you're counting accurately and the True Count is +3 or higher, the answer to "Insurance?" and "Even money?" is always no.
