When You Split, How Often Do Both Hands Win?

Splitting a pair turns one hand into two and puts a second bet on the felt to do it. The natural question is simple: how often do both of them win?

The answer depends on the pair and on the dealer's upcard. Split aces, and both hands win far more often than both lose. Split 8s against a 10, and both hands lose about twice as often as both win, yet the play is still correct. To answer with numbers instead of impressions, we ran the same simulation engine behind our article on how often you win a blackjack hand and isolated every round that began with a split.

What counts as a split round

The engine played about 62 million rounds under six deck rules: dealer stands on soft 17 and peeks, blackjack pays 3:2, double on any two cards, double after split, split to four hands, one card to each split ace and no resplitting aces, no surrender, cut card at 75 percent. Hands were played by basic strategy from the app's 4 to 8 deck chart, which splits a pair only against the upcards it lists.

A split round is every hand that grows from one split pair: split 8s, resplit a third 8, double one of the hands, and it still counts once, with the doubled result folded in. A round is every hand won when all its hands won, every hand lost when all lost, and mixed otherwise, a bucket covering anything where at least one hand pushed or the hands disagreed.

Splits are rare. All eight split pairs together account for 2.48 percent of rounds. For pairs other than aces, about 86 percent of split rounds end with exactly two hands, averaging 2.16 hands per round; aces are never resplit under these rules, so every split ace round is exactly two hands.

All eight pairs at a glance

The average result is the net for the whole round, in percent of one starting bet; the share of rounds is how often the pair was both dealt and split.

As a check, the Wizard of Odds publishes exact expected values for splitting every pair (blackjack appendix 9, six decks, dealer stands on soft 17, resplit to four hands except aces, one card to split aces). Taking its double after split column, weighted over the same upcards, the exact values give A,A +38.45, 9,9 +25.81, 8,8 -7.58, 7,7 +4.90, 6,6 -0.65, 4,4 +14.00, 3,3 +3.24 and 2,2 +7.04 percent. Ours land within one point for every pair except 4,4, which sits 1.85 points away and is the pair split least often: close agreement, not exact.

Aces are the best split of the eight at +38.11, nines next at +26.63; the always-split 8s average -7.31, and 6,6 dips just below zero at -0.55. Splitting is not reserved for hands that finish ahead, but for hands where two bets lose less, or win more, than the one hand held together.

Aces: the split where both hands usually agree

Aces are never resplit in this simulation, so every split ace round has exactly two hands, with these joint outcomes:

At every dealer upcard, without exception, both aces win more often than both lose.

Because each ace gets exactly one card and aces are not resplit, a split ace round always nets between -2 and +2 units. The net result: +2 units 41.18 percent of the time, +1 3.98, zero 26.64, -1 8.18, -2 20.02. Fold those back into outcomes and the round is won 45.2 percent of the time (both win at 41.2 plus win and push at 4.0), pushed 26.6 and lost 28.2. A split ace round pushes more than a quarter of the time, mostly one hand winning while the other loses.

The margin is wide. Per the Wizard of Odds exact tables, the best play that keeps the aces together is worth +19.1 percent against a 6 and splitting +68.2; against a 10, -6.6 against +18.2; against an ace, -2.2 against +12.7. The fence around the play, one card and no 3:2 on a made 21, is already priced in, as our article on always splitting aces and 8s explains.

Eights: the split that loses both hands most often

Split 8s look very different. Here is how the two hands finished in the 86.3 percent of 8,8 split rounds that ended with exactly two hands:

The single most common outcome is both hands losing. Against weak upcards the split is strong, +40.14 percent against a 6, but against a 9, 10 or ace the profile turns ugly: every hand loses 1.7 to 1.9 times as often as every hand wins. Against a 10, both hands win about one round in five (20.5 percent) and both lose about two in five (39.5).

The round level net for 8,8: +2 units 20.98 percent of the time, -2 26.54, zero 13.21, +1 7.92, -1 10.25, +3 8.97, -3 7.90, +4 2.00, -4 1.52, larger results rarer. Taken together, the round is won 40.3 percent of the time, pushed 13.2 and lost 46.5. Results beyond two units need a double or a resplit. With double after split, an 8 that catches a 3 becomes an 11, which the chart doubles against every upcard except an ace.

Why split 8s against a strong upcard anyway

None of that makes the split wrong. Against a 10 there is no winning play available; basic strategy picks the least losing one. The Wizard of Odds exact values after the peek:

Splitting loses least in all three columns; our simulation's -45.91 against a 10 sits about 1.6 points from their exact -47.5. The reasoning is the framework our article on expected value lays out: a decision is a comparison between options, not a judgment about whether the hand is good. The full 10 column is walked in our guide to playing against a dealer 10.

The marginal pairs

The remaining rows are quieter. A pair of 6s split against 2 through 6 averages -0.55 percent, and splitting is still the chart's play; our pair splitting guide lists where each of these pairs splits and how taking away double after split moves the borderline. Fours are the rarest split at 0.088 percent of rounds, taken only against a 5 or 6. Nines, split against 2 through 6 plus 8 and 9, are the second best split at +26.63 and have their own article on when to split 9s.

What the tables show is that how a split feels is a poor guide: both hands losing is the most common outcome for 8s, and the split is right anyway. The chart already knows this; the work is drilling the rows until your hands do too.

Frequently asked questions

How often do both hands win when you split aces?

In our simulation, both hands won 41.2 percent of split ace rounds and both lost 20.0. Counting the whole round, it was won 45.2 percent of the time, pushed 26.6 and lost 28.2. Both hands won more often than both lost against every dealer upcard, and the average result was +38.11 percent of the starting bet.

Should you still split 8s against a dealer 10?

Yes. Against a 10, both split hands win about 20.5 percent of the time and both lose about 39.5, and in our simulation the split lost 45.91 percent of the starting bet on average. The alternatives are worse: per the Wizard of Odds exact tables, standing on 16 is worth -53.7 percent, hitting -53.5 and splitting -47.5. Splitting loses least.

What are the chances both hands lose when you split?

It depends on the pair. Two hand split 8 rounds ended with both hands losing 34.4 percent of the time and both winning 31.1. For aces it was 20.0 percent both losing against 41.2 both winning. Against a 9, 10 or ace, split 8s lose both hands 1.7 to 1.9 times as often as they win both.

Can splitting give you more than two hands?

Yes. Many tables allow resplitting if you draw a third card of the same rank. In our simulation, about 86 percent of split rounds for pairs other than aces ended with exactly two hands, averaging 2.16 hands per round. Split aces are different: each ace receives one card, and resplitting aces was not allowed, so those rounds stayed at two hands.

Can you practice split decisions without risking money?

Yes. 21 Trainer deals practice hands with no real money involved and gives instant right or wrong feedback with an explanation on every decision, splits included. The charts redraw for the rules you select, such as double after split and split limits, so the split decisions you practice follow the rule set you plan to play.

Keep reading

• Why You Always Split Aces and 8s
• Splitting Pairs: Always, Never, and It Depends
• What Share of Blackjack Rounds Do You Win, Push and Lose?