What Insurance Actually Is
When the dealer shows an ace, the table offers you insurance — a side bet equal to half your original wager that pays 2:1 if the dealer has blackjack. It is marketed as "protecting" your hand, but it is a completely separate bet with its own math. Your original wager is unaffected by whether you take insurance.
To break even on a 2:1 payout, the dealer needs a ten-value hole card at least 1 in 3 times (33.33%). In a standard deck, ten-value cards (10, J, Q, K) make up 4 out of 13 ranks, or 30.77%. Since 30.77% is less than 33.33%, insurance is a losing bet under normal conditions.
The Mathematics: Why Insurance Is Negative EV
Consider a 6-deck shoe with 312 total cards. There are 96 ten-value cards (16 per deck × 6 decks) and 216 non-ten cards.
When the dealer shows an ace, you know one card (that ace). Of the remaining 311 cards, 96 are tens and 215 are not. The probability of a ten in the hole is 96/311 = 30.87%.
- If you bet $5 insurance on a $10 hand:
- 30.87% of the time the dealer has blackjack: you win $10 from insurance (2:1 on $5) but may lose $10 on your hand → net depends on your hand
- 69.13% of the time the dealer does not: you lose your $5 insurance bet outright
- Expected value of the insurance bet alone: (0.3087 × $10) − (0.6913 × $5) = $3.087 − $3.457 = −$0.37 per bet
That is a 7.4% house edge on the insurance bet — one of the worst bets on the table. For comparison, a basic strategy player faces only 0.4-0.6% house edge on their main bet.
Even Money vs Insurance: The Same Decision
When you have a blackjack and the dealer shows an ace, the dealer will offer "even money" — an immediate 1:1 payout instead of waiting to see if the dealer also has blackjack. Players love taking even money because it guarantees a profit. But mathematically, taking even money is identical to taking insurance.
Here is why: if you decline even money and the dealer does not have blackjack (69% of the time), your blackjack pays 3:2. If the dealer does have blackjack (31% of the time), you push and win nothing. The expected value of declining even money is higher than accepting it.
- Taking even money: guaranteed $10 on a $10 bet = $10 every time
- Declining even money: 69% × $15 + 31% × $0 = $10.35 on average
You give up $0.35 per occurrence by taking even money. Over hundreds of dealer aces, that adds up. The emotionally satisfying "guaranteed win" is mathematically worse.
When Card Counters Take Insurance
Insurance becomes a positive-EV bet when the remaining deck has more than 1-in-3 ten-value cards. Card counters track this ratio through the true count (TC). The threshold varies slightly by the number of decks and specific count system, but general guidelines for Hi-Lo are:
| Game | Insurance Becomes +EV at | Notes |
|---|
| 6-deck shoe | TC +3 or higher | Most common threshold cited in literature |
| 2-deck hand-dealt | TC +3 | Count accuracy matters more in short decks |
| Single deck | TC +2 or higher | Fewer cards = faster count swings; rare opportunity |
At TC +3 in a 6-deck game, the density of tens in the remaining cards crosses the 33.33% break-even point. Above that, insurance has positive expected value and a counter should take it every time.
Insurance is actually one of the Illustrious 18 deviations — it is among the highest-value count-based decisions because it occurs frequently and the EV swing is large when the count is high.
Common Mistakes
- "Insuring a good hand." Players often take insurance when they have 20 or blackjack, reasoning that they have a "lot to lose." The strength of your hand is irrelevant to the insurance decision. Insurance is a separate bet on whether the dealer has a ten underneath — your hand does not change the probability.
- Always taking even money. Guaranteed money feels safe, but as shown above, declining even money has higher expected value over time.
- Confusing insurance with protection. Insurance does not protect your hand. It is a side bet. If you would not place a random side bet at an 7.4% house edge, you should not take insurance.
- Taking insurance at low counts. Some new counters take insurance whenever the count is positive. A TC of +1 or +2 is not sufficient — the ten density is still below break-even. Wait for TC +3 or higher.
Frequently Asked Questions
Should basic strategy players ever take insurance?
No. Without information about the remaining deck composition, insurance is always negative EV. Decline it every time.
Is insurance ever worth it at a 6:5 table?
The insurance bet itself pays the same 2:1 regardless of the blackjack payout rules. However, at a 6:5 table you should already avoid the table entirely. If you are stuck at one, the insurance math is unchanged — still decline without a count.
Does the number of other players affect the insurance decision?
No. Other players' cards are random draws from the same shoe. Whether you can see them or not, your insurance decision depends only on the true count (for counters) or should be declined (for non-counters).
Why do dealers call it "insurance" if it is a bad bet?
The name is deliberate marketing. Casinos profit enormously from insurance bets — the 7.4% house edge is among the highest for any standard table game wager. The framing as "protection" encourages players to take it out of fear rather than math.
Related Guides