Last updated: 14 October 2026 · By Power.Win Editorial Team
Quick Answer: No, blackjack insurance loses money: it carries about a 7.4% house edge in a six-deck game. It pays 2 to 1, but the hole card is a ten only about 30.9% of the time, and 2 to 1 needs 33.3%. On Power.Win’s provably fair blackjack the card order is committed before the deal, so seen cards never make it cheaper.
The dealer turns an ace face up and asks the table a question with a three-second clock on it. Insurance sounds like a seatbelt. It’s priced like a lottery ticket, and the price is the whole story.
Here is the number everything else comes from. A six-deck shoe holds 312 cards, and 96 are ten-value: ten through king, sixteen per deck. Once the dealer’s ace is face up, 311 cards are still unseen, so the chance the hole card is a ten is 96 ÷ 311, or 30.87%. Insurance pays 2 to 1, and a bet paying 2 to 1 needs to land one time in three, 33.33%, just to break even. You’re offered a price for something that happens less often than the price requires.
Is blackjack insurance worth taking?
No. Over any meaningful number of hands insurance loses about 7.4 cents of every dollar you put on it, and the gap that creates the loss is the 30.87% above against the 33.33% the payout demands.
The arithmetic is one line. Put up a dollar: you collect two dollars 30.87% of the time and lose one dollar the other 69.13% of the time, so 0.3087 × $2 − 0.6913 × $1 = −$0.0739 per dollar, a 7.4% house edge.
That edge doesn’t move with how well you’re playing or how good your own hand looks. It’s a fixed price on a single question, and the question has nothing to do with your cards. Scale is what makes the figure matter: 7.4% is an order of magnitude above what the main hand costs under ordinary rules, which is why a handful of insurance bets can outweigh a whole session.
What is insurance in blackjack, and when does the dealer offer it?
Insurance is a separate side wager, of up to half your main bet, offered only when the dealer’s up card is an ace, and it pays 2 to 1 if the dealer’s hole card turns out to be a ten-value card.
The sequence at the table runs like this:
- Everyone is dealt two cards; the dealer gets one face up and one face down.
- The up card is an ace, so the insurance offer opens.
- You may put up to half your main stake in the insurance line.
- The hole card is resolved. A ten-value card means the dealer has a natural: insurance pays 2 to 1, and your main hand loses unless you also hold a natural.
- Anything else means insurance loses, and the hand plays on normally.
The name does the damage, because nothing is being insured. Your main bet is not protected, reduced or hedged in any way that survives contact with the arithmetic: it goes on to win, lose or push exactly as it would have done. What you’ve bought is a second, independent bet on one face-down card.
Be precise about what “ten-value” covers, because that is where most estimates go wrong: tens, jacks, queens and kings all count, sixteen cards in every fifty-two. That abundance is why the bet feels close to even. It reaches roughly three-in-ten, then stops short of the one-in-three it needs.
What does insurance pay, and what are the true odds?
Insurance pays 2 to 1. The true odds against the hole card being a ten in a six-deck shoe are about 9 to 4. You are being paid 8 to 4 on a 9 to 4 shot, and that missing quarter of a unit is the entire house edge.
Set it out in the shoe’s own counts. After the dealer’s ace, 311 cards are unseen: 96 are ten-value, 215 are not. The odds against a ten are 215 to 96, or 2.24 to 1 — close enough to 9 to 4 that the fraction is the honest way to say it.
| Figure | |
|---|---|
| Ten-value cards in a six-deck shoe | 96 of 312 |
| Unseen cards after the dealer’s ace | 311 |
| Chance the hole card is a ten | 30.87% |
| True odds against a ten | 215 to 96, about 9 to 4 |
| What a fair payout would be | about 2.24 to 1 |
| What insurance actually pays | 2 to 1 |
| Break-even payout chance needed | 33.33% |
This is the part the usual explanations leave out. A bet is fairly priced when the payout matches the true odds, so at 2.24 to 1 paid as 2 to 1 you are short 0.24 of a unit every time it lands. Multiply that shortfall by how often it lands (0.24 × 0.3087) and you get 0.074: the same 7.4%, reached from the other direction. Wizard of Odds puts the loss at 7.4 cents per dollar insured at six decks.
What does insurance cost per $100 staked?
Every $100 you stake on insurance costs you about $7.40 on average. Because the maximum insurance bet is half your main bet, the cost per hand scales directly with the hand you were already playing.
| Main bet | Max insurance (half) | Pays at 2 to 1 | True odds against a ten | Cost per $100 insured | Expected cost of this bet |
|---|---|---|---|---|---|
| $10 | $5 | $10 | about 9 to 4 | $7.40 | $0.37 |
| $20 | $10 | $20 | about 9 to 4 | $7.40 | $0.74 |
| $50 | $25 | $50 | about 9 to 4 | $7.40 | $1.85 |
| $100 | $50 | $100 | about 9 to 4 | $7.40 | $3.70 |
The right-hand column is the insurance stake multiplied by 7.4%, and it is deliberately small. That is the trap rather than an argument for the bet: thirty-seven cents on a $10 hand does not feel like a decision worth refusing, so it gets taken, and the cost only becomes visible when you count how often you take it.
Against the other bets on the felt, insurance sits at the expensive end. Our guide to blackjack side bets ranks it alongside 21+3 and Perfect Pairs with each pay table priced out, and insurance lands fifth of the eight pay lines priced there — above 21+3 and three of the four Perfect Pairs tables, below three dearer lines. That page is the catalogue; this one is the price tag on its own.
What does taking insurance every time cost over 100 hands?
Over 100 hands at $100 a hand, insuring every dealer ace costs about $30 on under 4% of the money you put at risk, producing about 41% of your total expected loss.
The dealer’s up card is an ace one time in thirteen, so 100 hands brings the offer up about eight times. Insure all eight at the maximum $50 and you’ve staked $400 on insurance alongside $10,000 on main hands.
| Main hands | Insurance bets | |
|---|---|---|
| Decisions in 100 hands | 100 | about 8 |
| Total staked | $10,000 | $400 |
| House edge | 0.43% | 7.4% |
| Expected cost | $43.00 | $29.60 |
The 0.43% is Wizard of Odds’s figure for Atlantic City rules with the dealer standing on soft 17, played to correct strategy: a reference point for that rule set, not a figure Power.Win publishes.
Read the columns side by side and the shape of the session changes. The $400 is under 4% of the $10,400 staked in total, yet it accounts for $29.60 of a $72.60 expected loss: about 41% of the damage from a twenty-fifth of the money. Eight decisions that each felt like $3.70 did nearly as much to the balance as a hundred hands.
The main hand’s edge does move, which is the honest counterweight: it moves with your decisions, covered in our blackjack basic strategy guide, and with the table’s rules, including whether the dealer hits or stands on soft 17. Neither touches the insurance price, which is set by the shoe’s composition and the 2-to-1 payout alone.
Why is even money the same wager as insurance?
Even money, the flat 1-to-1 settlement offered when you hold a natural against a dealer ace, is insurance wearing a different name. Taking it costs about 3.9% of your main bet, and it’s marginally worse than ordinary insurance, not better.
Hold a natural against a dealer ace and you have two choices. Take even money and you collect one unit, immediately and certainly. Decline it and you are paid 3 to 2 unless the dealer also has a natural, in which case the hand pushes and you collect nothing.
Price the second option. The ten-density here is slightly lower than usual, and the reason is in your own hand: a natural is an ace and a ten-value card, so one of the tens the bet needs is already in front of you. Three cards are known, leaving 309 unseen and only 95 ten-value — 30.74%.
- Dealer also has a natural (30.74%): the hand pushes, you collect 0
- Dealer does not (69.26%): your natural pays 3 to 2, you collect 1.5 units
- Expected return: 0.6926 × 1.5 = 1.0389 units
So declining returns about 103.88% of your stake on average, against a flat 100% for taking even money. That difference, roughly 3.9% of your main bet, is what the certainty costs you.
Run the same 30.74% through the insurance formula and it comes out worse than the standard case: 2 × 0.3074 − 0.6926 = −0.0778, a 7.8% house edge rather than 7.4%. Holding a natural does not make insurance a better bet; it makes it slightly dearer, because your own ten has been taken out of the pool the bet needs.
When is taking insurance the correct play?
Only when you know the shoe has grown richer in ten-value cards than it started, enough to push the chance past 33.33%. That is a real condition, not a theoretical one, but it needs information the hand in front of you cannot supply.
The threshold is calculable. Insurance breaks even when ten-value cards make up a third of what is unseen, and at the start of a six-deck shoe 96 of 311 is 30.87%, so the shoe is short. Hold the ten count at 96 and ask how many non-tens have to leave before the ratio reaches a third:
- Need: 96 ÷ (311 − x) ≥ 1 ÷ 3
- So: 311 − x ≤ 288
- So: x ≥ 23
In other words, roughly 23 more low and middle cards than ten-value cards have to have left that shoe before insurance stops being a losing bet. Past that point it tips the other way — which is why insurance is the decision that tracking the shoe pays off on soonest. Doing so means keeping a running tally of what has been dealt, and our guide to how card counting works sets out how that tally is kept and what it can and cannot tell you.
The condition attached is the hard part: the 23-card threshold only exists if cards leave a shoe and do not come back. Where a table reshuffles continuously, the composition resets before any shortfall can build, so the ratio never drifts far from 30.87%. The same holds of an online shoe where each hand’s order is fixed independently of the last. In both cases the 7.4% is a flat price on every offer at that table, and the correct play is the simple one: decline it.
What does a committed shoe mean for insurance on Power.Win?
On Power.Win’s provably fair blackjack, the card order is settled before the first card is dealt and each hand is committed on its own, so nothing you have already seen tells you anything about the hole card, and the insurance price never improves with what you’ve seen. Power.Win’s documentation describes a “standard 52-card deck” and publishes no shoe size, so the exact edge depends on a figure it doesn’t state — the 7.4% above is the six-deck case.
The blackjack table under PWR Originals is one of the in-house games, and those are provably fair: the games-blackjack page states that the card order is committed before the first card is dealt and can be proved afterwards. Three inputs fix it — a server seed, hashed with SHA-256 and shown to you as a hash before play; a client seed generated in your browser and replaceable by you; and a nonce that increases with each hand.
The same page is blunt about what follows: “Nothing about how you play alters the sequence. Taking a card cannot cause a different card to be placed there, because the position was already decided.” That closes the one door through which insurance could ever have become correct — cards already seen tell you nothing about the next hand, so no shoe drains of its tens and the 23-card threshold has nothing to count.
This is a different product from the live dealer blackjack in the casino section, which is supplied by certified third parties. The two guarantees are not the same claim: provably fair means the in-house game commits its sequence in advance and lets you check it afterwards, while certified means an independent lab has tested a third party’s game. Both are legitimate, but only the first lets you reproduce the hand.
The live game also exposes a hands control beside the bet entry, so you can play more than one hand in a round, each carrying its own stake, which means one, two or three hands in a round, and so more insurance offers per round. The blackjack documentation publishes the table’s own rules, and Power.Win’s framing of strategy applies here exactly: correct play “narrows the gap. It does not close it, and no sequence of decisions turns the game positive.”
Power.Win is operated by 3-102-968317 SRL, Costa Rica, licensed by the Tobique Gaming Commission, licence 0000209. Power.Win isn’t available to players in the United States: section 5 of its Terms and Conditions lists the United States of America among the countries whose players aren’t eligible to deposit or play real-money games. Read this page as an explanation of what the insurance bet costs, not as an invitation to sign up.
Frequently Asked Questions
Does insurance protect your original bet?
No. It’s a separate wager on one question: whether the hole card is a ten. Your main hand plays on exactly as it would have done, and at six decks the insurance bet loses about $7.40 of every $100 staked on it, win or lose.
What is the most you can stake on insurance?
Half your main bet, and no more. On a $100 hand that’s $50, which pays $100 if the dealer’s hole card is a ten-value card. At a 7.4% house edge that $50 costs you about $3.70 on average every time you place it.
Should you take insurance when you hold a strong hand?
No, because the strength of your hand is irrelevant to it. Insurance settles only on the dealer’s hole card, so a 20 and a 12 face the identical 30.87% chance and the identical 7.4% cost. Holding a natural makes it slightly dearer still, at 7.8%.
Does the dealer check the hole card before you decide?
No, and that sequencing is the point of the bet. The offer opens while the card is still face down, so you are pricing it blind. Wizard of Odds puts the resulting loss at 7.4 cents per dollar insured in a six-deck game.
Is insurance ever cheaper on a single-deck table?
Slightly, and never enough. One deck holds 16 ten-value cards in 52, so after the dealer’s ace 16 of 51 unseen cards are tens: about 31.4%, against 30.87% at six decks. Both fall short of the 33.33% that a 2-to-1 payout needs.
Sources (Citations)
- Power.Win, Provably Fair Blackjack Explained (power.win/info/games-blackjack.html): PWR Originals blackjack as an in-house provably fair game; the committed card order and its server seed, client seed and nonce; “Nothing about how you play alters the sequence”; cards already seen telling you nothing about the next hand; the hands control, each hand carrying its own stake; live dealer blackjack as a distinct product from certified third parties; operator 3-102-968317 SRL, Costa Rica, Tobique Gaming Commission licence 0000209. Read 14 October 2026.
- Power.Win Blackjack documentation (docs.power.win/documentation/games/blackjack): the table’s published rules. Read 14 October 2026.
- Power.Win Terms and Conditions (power.win/terms): section 5, country eligibility. Read 14 October 2026.
- Wizard of Odds, Blackjack Insurance (wizardofodds.com/video/blackjack-insurance/): insurance costing 7.4 cents per dollar insured in a six-deck game. Read 14 October 2026.
- Wizard of Odds, Blackjack Basics (wizardofodds.com/games/blackjack/basics/): the 0.43% main-hand house edge for Atlantic City rules with the dealer standing on soft 17 and doubling after a split allowed, used only as a reference for that rule set. Read 14 October 2026.
- Shoe arithmetic here is computed from the standard six-deck composition, 312 cards and 96 ten-value, with every step shown inline above so you can reproduce each figure yourself.
Take the Lead, Gamble Responsibly
Insurance is the one bet on the table that is sold as caution, and reading its price is the difference between a careful decision and a costly reflex.
Work out what a session is worth risking before the first hand is dealt, and use the deposit limits, reality checks and self-exclusion tools on Power.Win’s responsible gambling page. If the betting has stopped being fun, treat that as the signal to stop; the support resources are listed on that same page.
By the Power.Win Editorial Team