A blackjack table does not become more expensive merely because the dealer offers insurance. The extra expected cost appears when the player takes the insurance wager at a time when the dealer’s hidden card is not ten-valued often enough to justify the 2:1 payout.
That distinction matters. Insurance is a separate side bet on the dealer’s hole card. It is not protection for a weak 16, a reward for holding 20, or a modification of the main-hand house edge when the offer is declined.
The 2:1 payout creates an exact break-even point
Let:
- T = unseen ten-value cards (10, jack, queen, king);
- U = total unseen cards;
- p = T/U = probability that the dealer’s hole card is ten-valued.
For a $1 insurance wager, a win produces $2 profit and a loss costs $1. Expected value is therefore:
EV = 2p - (1 - p) = 3p - 1
The break-even condition is:
3p - 1 = 0
so:
p = 1/3 = 33.33%
That is the core insurance test. If more than one-third of the unseen cards are ten-valued, a 2:1 insurance bet has positive expectation. If fewer than one-third are ten-valued, the casino has the edge on that side bet.
Expressed per dollar insured:
Insurance house edge = 1 - 3(T/U)
The formula is composition-dependent. There is no single immutable insurance edge for every point in every shoe.
A fresh six-deck example shows why basic strategy declines it
Six decks contain 312 cards, including 96 ten-value cards. Suppose the dealer shows an ace and the player’s two cards are 9 and 7. Three visible cards have been removed, but none is ten-valued.
That leaves:
- 309 unseen cards;
- 96 unseen ten-value cards.
The dealer-blackjack probability is:
96 / 309 ≈ 31.07%
Insurance EV per $1 is:
3 × (96/309) - 1 = -21/309 ≈ -0.0680
So the side bet loses about 6.8 cents per $1 insured under those exact fresh-shoe conditions. On a $25 insurance wager, the expected loss is about $1.70.
The player’s visible cards matter because they are no longer available to be the dealer’s hole card. If one player card is ten-valued, only 95 tens remain unseen; if both are ten-valued, only 94 remain. A visually strong player hand can therefore make ordinary insurance slightly worse, not safer.
| Player’s two cards remove | Unseen tens | Unseen cards | Insurance EV per $1 |
|---|---|---|---|
| No ten-value card | 96 | 309 | about -$0.068 |
| One ten-value card | 95 | 309 | about -$0.078 |
| Two ten-value cards | 94 | 309 | about -$0.087 |
These are fresh-shoe illustrations. Later in the shoe, exposed cards can move the exact ratio in either direction.
”Insurance is offered” and “insurance is taken” are different house-edge questions
When published blackjack house edge is quoted for a rule set and correct basic strategy, the optional insurance offer is not normally treated as though every player automatically takes it.
If the player declines insurance, the main hand continues under the ordinary rules: dealer standing or hitting soft 17, doubling rules, split rules, blackjack payout, surrender, number of decks, and so on. The mere appearance of the insurance line has not added a second wager.
When the player accepts insurance, the round now contains two economic components:
Combined round EV = main-hand EV + insurance EV
For a $50 original wager and a $25 insurance wager in the 96-of-309 example, insurance contributes approximately:
$25 × (-21/309) ≈ -$1.70
Whatever the main hand is worth under the exact rules and decision, the side bet adds that separate negative expectation.
This also explains why there is no useful single number called “the house edge when insurance is offered” without specifying behavior. A player who never takes it has no insurance action. A player who always takes it creates additional action only on dealer-ace rounds. A composition-aware counter may sometimes take it when the ratio crosses the break-even point.
The value of insurance does not depend on whether your hand is 12, 16, or 20
Insurance is sometimes sold psychologically as protection for a “good” hand. Mathematically, the player’s total is irrelevant to the insurance wager except to the extent that the visible card ranks alter deck composition.
If the dealer has blackjack, insurance wins whether the player’s main hand is 12 or 20. If the dealer does not have blackjack, insurance loses regardless of how the main hand later resolves.
That is why insurance should be analyzed as its own bet. Combining the emotional value of “saving” a strong main hand with the side-bet math is one of the most common reasoning errors in blackjack.
Read Insurance Bet for the rule overview and When to Take Insurance for the decision framework.
Even money is insurance expressed as a guaranteed result
When the player has a natural blackjack and the dealer shows an ace, some tables offer even money. The player can accept a guaranteed 1:1 win rather than wait to see whether the dealer also has blackjack.
Economically, that is the same insurance problem packaged differently.
In a fresh six-deck example, the player’s blackjack removes one ace and one ten-value card. With the dealer’s ace also exposed, 309 cards remain unseen and 95 are ten-valued. If a 3:2 blackjack payout applies, declining even money has expected profit per $1 original wager of:
1.5 × (1 - 95/309) ≈ 1.039
That is about $1.039 expected profit versus the guaranteed $1 from even money. The difference is small on one hand but systematic over repeated offers.
The result changes if the blackjack payout itself is different, which is why the main table rules still matter. See why never take even money for the dedicated comparison.
Card counting is the legitimate exception because it estimates composition
Insurance is one of the clearest places where card composition can change the correct decision. If enough low cards have been removed while ten-value cards remain, T/U can rise above one-third.
A card counter is not predicting the dealer’s hidden card. The counter is estimating whether the side bet has crossed from negative to positive expectation.
That is also why “take insurance whenever the count is positive” is too crude. Counting systems use specific insurance indices; shoe games require a running-count-to-true-count conversion; penetration and deck estimation matter; and mistakes in counting can erase a thin theoretical advantage.
The card-counting basics page explains running count, true count, and composition estimation without confusing them with hand-by-hand prediction.
Procedure can change without changing the one-third break-even rule
The common insurance structure allows a wager up to half the original blackjack bet and pays 2:1 when the dealer has blackjack. Current Massachusetts rule materials, for example, specify a 2:1 insurance payout and describe timing around the dealer’s ace and hole-card check.
Actual casino procedures can still differ in ways that matter to the main hand:
- a U.S.-style hole-card game may check for blackjack before player action;
- a no-hole-card game may expose dealer blackjack only after doubles or splits have been made;
- some rule sets protect certain additional wagers against dealer blackjack while others do not;
- chip denominations can affect the exact physical amount accepted as “half” the original wager;
- electronic tables may present insurance through a timed interface rather than a felt insurance line.
Those procedural differences can affect exposure on the main wager, especially after doubles and splits. They do not change the arithmetic of a separate 2:1 insurance wager: it needs a dealer-blackjack probability above one-third to be favorable.
Insurance can raise session cost even though it appears on only some rounds
A player who automatically takes every insurance offer does not add a losing side bet to every blackjack hand. Insurance appears only when the dealer’s upcard triggers the offer. That makes session-level cost a frequency problem.
A simple accounting model is:
Expected insurance loss over a session = number of insurance wagers × average insurance stake × average insurance house edge
This is more informative than adding the insurance edge directly to the main-game house edge. The denominators are different: one is priced per dollar of insurance action; the other is normally priced per dollar of initial blackjack action.
The practical conclusion is precise. The offer itself does not hurt the player. Blind acceptance usually does. To compare the rules that price the main game, use Blackjack House Edge by Rules.