Blackjack Side Bets and Insurance: Probability, Payout Math and Variance

Blackjack insurance, Perfect Pairs and 21+3 are related to the same deal, but they are not the same kind of wager. Insurance is offered only after the dealer exposes an ace and concerns the dealer's hidden card. Perfect Pairs concerns the player's first two cards, while 21+3 uses those two cards together with the dealer's upcard. In every case, the side wager is settled separately from the main blackjack hand.

The examples below use one clearly labelled, illustrative ruleset: a six-deck shoe, an insurance wager of up to half the original stake, and these example side-bet tables. Actual names, qualifying combinations and payouts can differ by casino, software and jurisdiction, so the layout in front of you takes priority.

How payout math works

Three terms prevent most confusion. The stake is the amount placed on the side bet. A gross payout is the amount returned by the winning wager, including the original stake when the table describes a total return. A net result is what remains after the stake is removed:

Net result = payout received − stake

For a table quoting a profit multiple, a 5:1 result on a $2 stake means $10 profit and usually a $12 total return. If a table instead says “pays 5:1” but defines the figure as the total return, the result is different. Always establish the convention before comparing outcomes.

Insurance: a wager on the dealer's hole card

Insurance becomes available when the dealer's upcard is an ace. It is a separate wager that the dealer's concealed card has a ten value: a 10, jack, queen or king. Under the illustrative ruleset, the insurance stake can be half the original blackjack stake and a successful insurance wager pays 2:1 profit. Thus, a $10 insurance stake returns $30 in total: $20 profit plus the $10 stake.

The insurance wager is settled independently. If the hidden card has a ten value, the insurance bet wins; otherwise it loses. The main hand is still settled under the ordinary blackjack rules. This is why insurance should not be described as simply another way for the original hand to win.

Calculating insurance probability

The calculation is a sample-space question: count the unseen cards that qualify, then divide by the number of unseen cards. In an initially complete six-deck shoe there are 312 cards and 96 ten-valued cards, because each deck contains four such ranks. Once the dealer's ace and the player's two cards are known, 309 cards remain unseen. If neither player card is ten-valued, the illustrative probability is:

96 ÷ 309 ≈ 31.07%

If one of the player's cards is ten-valued, the numerator becomes 95; if both are ten-valued, it becomes 94. Other exposed cards, a partially used shoe or a different number of decks changes both the numerator and denominator. This is why there is no single universal insurance probability. For a broader reference point, compare this method with blackjack odds and probabilities while keeping the deck assumptions visible.

Perfect Pairs: counting the player's first two cards

Perfect Pairs is won when the player's first two cards have the same rank. The illustrative table below divides that event into three mutually exclusive categories. “Same rank” means both cards have the same number or letter; “same colour” means both are red or both black; and “perfect” means the same rank and suit, with the matching cards coming from different decks.

This is an illustrative table, not a universal casino rule. Under its profit-multiple convention, the $2 mixed-colour example produces $10 profit and returns $12 in total. The $2 perfect-pair example produces $60 profit and returns $62.

The combination count shows how the categories are separated. With six decks, each rank has 24 cards. The total number of unordered two-card deals is C(312, 2) = 48,516. For any one rank, there are C(24, 2) = 276 matching-rank pairs, so all 13 ranks produce 3,588 matching-rank combinations.

For the perfect category, each suit has six copies of a rank. Four suits multiplied by C(6, 2) gives 60 combinations per rank, or 780 across all ranks. Coloured pairs account for 72 combinations per rank, or 936 overall. The remaining 1,872 matching-rank combinations are mixed-colour pairs. Dividing each count by 48,516 gives the frequency of that tier under these assumptions. The categories must be separated before adding them; otherwise the same deal could be counted more than once.

21+3: three cards and poker-style outcomes

The 21+3 wager uses the player's first two cards plus the dealer's upcard. The cards are evaluated together for a three-card poker-style combination, not for whether the main blackjack hand totals 21. Under this illustrative table, the highest applicable category takes priority:

Again, this is an example ruleset. A straight flush is also a flush and a straight, while a suited three of a kind is also three of a kind. The tier-priority rule prevents overlapping wins from being paid separately. On a $2 stake, a flush pays $10 profit and returns $12 in total. A suited three of a kind pays $200 profit and returns $202.

To calculate a frequency, define the sample space first. If the six-deck shoe is complete and card identities are treated as equally likely, the number of unordered three-card combinations is C(312, 3). Count the combinations matching each category, assign a qualifying combination to its highest applicable tier, and divide each tier count by that denominator. If known cards have already been removed, reduce the denominator and recalculate the available ranks and suits. A documented calculator can help, but it must state whether it uses one deck or multiple decks and how it resolves overlapping categories.

Win frequency is not the same as volatility

The probability of any win is the sum of the probabilities of the mutually exclusive winning tiers. The probability of a particular tier is its own favourable-outcome count divided by the sample space. A frequent low-paying result and a rare high-paying result can therefore coexist without making the high-paying result frequent.

For example, imagine a simplified sequence of ten side-bet rounds in which a small winning tier appears twice and a larger tier does not appear at all. The player may describe that as “two wins,” but eight losing rounds still create an uneven session. In another ten-round sequence, one rare high tier may appear once, producing a very different balance despite only one winning round. That unevenness is variance: results move between losing rounds, occasional small returns and infrequent large returns.

Looking at a payout table without its outcome frequencies can therefore give a misleading impression of how a wager behaves. Conversely, a relatively common winning tier does not remove the possibility of long stretches without a qualifying combination. The mechanics can be measured with counts, but short sequences can still look very different from the underlying distribution.

What to check on the table

Before calculating any blackjack side bets and insurance, read the posted rules rather than relying on a familiar name. Check the qualifying cards, the payout convention, the priority used when categories overlap, and whether the table is fixed or progressive. Also check which rules source defines the game, because the same label can represent different combinations or payout tiers in different settings.

Keep the side-bet calculation separate from decisions about the main hand. Resources on basic strategy explained address the ordinary blackjack hand, while a session budget and bankroll and session planning concern how a player manages money over time; neither changes what a posted side-bet table means. The most reliable calculation is the one that states its deck model, known cards, category definitions and payout convention explicitly.

Gambling is for adults aged 18+ only (or older where local law requires). Play responsibly: set a budget and time limit in advance, treat gambling as entertainment rather than a way to make money, and contact a local support organisation if gambling stops being fun.