The Decision You Make Before the Dealer Acts
You have been dealt two cards of the same rank. The dealer has one card face up. Before you decide whether to hit, stand, or double, you first have to answer a question most new players get wrong: do you keep these two cards together, or do you pay a second wager and turn them into two separate hands?
That is a split. It is the only blackjack decision that changes the number of hands you are playing in the middle of a round. It costs exactly one additional bet, equal to your original wager. It is not always available, and it is not always wise. The correct answer depends on the dealer's upcard, the exact table rules, and whether the casino lets you do anything useful with the split cards afterward.
What follows is the mechanical picture: what a split does, which pairs the published basic-strategy charts consistently treat as automatic splits, which ones depend on the rule set, and the procedural mistakes that quietly drain a bankroll.
What a Split Physically Does
When you split, you place a second wager next to your first. The dealer separates your two cards into two starting hands. You then play the first hand to completion—hitting, standing, or doubling as the rules allow—before moving to the second. Each hand is settled independently against the dealer's hand.
Three structural rules shape every split decision. First, the two cards must be of the same rank. Some tables also allow splitting any two ten-value cards, such as a king and a jack, but many do not; the table rules will say. Second, each split hand begins with only one of the original cards. You are not dealt a fresh two-card hand. Third, the second wager is a full bet, not a fraction, and you cannot split for less than the original stake.
What happens after the split varies by rule set, and those variations change whether a marginal split is worth making at all. The table below lays out the four rule options that most affect the decision.
| Rule option | What it allows | How it changes the decision |
|---|---|---|
| Double after split | You may double down on one or both split hands | Makes splitting more attractive on pairs where you might want to double a strong starting total |
| No double after split | Split hands may only be hit or stood | Reduces the value of splitting pairs that frequently build toward a double-down total |
| Resplit aces | A second ace dealt to a split ace may be split again | Changes the ace-pair rule at tables that otherwise restrict aces to one split |
| Hit split aces | You may draw additional cards to a split ace | Rare; most tables give each split ace exactly one card |
None of these options changes the core arithmetic of the game. Blackjack is a comparing card game where the objective is a hand total closer to 21 than the dealer's without exceeding it. The dealer's upcard is the single most important piece of information you have, because it tells you how likely the dealer is to bust or to build a strong total. A split decision is always made against that upcard, never in isolation.
The Pairs the Charts Treat as Automatic
Published basic-strategy charts derive from the specific rule set in use: deck count, dealer stand rules, doubling and splitting rules. A chart printed for a six-deck shoe with the dealer standing on soft 17 may give a different answer than a chart for a single-deck game where the dealer hits soft 17. Anyone who memorises one chart and applies it everywhere is playing the wrong game.
With that caveat, two pairs are treated as automatic splits across nearly every published rule set. A pair of aces should be split against any dealer upcard. A pair of eights should also be split against any upcard. The reasoning is mechanical, not superstitious. A pair of aces as a single hand is a soft 12, a weak starting total that forces you to draw. Split, each ace starts a hand with a strong building block, even though most tables then deal each ace only one card. A pair of eights totals 16, the worst starting total in the game against a strong dealer upcard. Split, each eight starts a hand with a far better chance of competing.
The reverse case is just as consistent. A pair of tens, or any two ten-value cards, is a strong 20. Splitting it breaks a near-certain winner into two uncertain hands. The charts treat this as a stand, not a split, against every upcard. A pair of fives totals 10, and the charts treat it as a double-down opportunity rather than a split, because 10 is a strong building total in its own right.
Where the Dealer's Upcard Decides
Most other pairs sit in a middle band where the split decision changes with the dealer's upcard and the table rules. The logic is consistent: split when doing so creates two hands that are each better positioned than the single hand you currently hold, and do not split when it merely doubles your exposure on a weak total.
The comparison below shows how the main pair groups divide. It is a structural summary, not a substitute for a chart matched to the exact table you are playing.
| Pair | General treatment | What drives the decision |
|---|---|---|
| Aces | Split against any upcard | Converts a weak soft 12 into two strong starting hands |
| Eights | Split against any upcard | Escapes the worst starting total in the game |
| Tens | Stand | 20 is too strong to break apart |
| Fives | Double instead of split | Total of 10 is a strong building hand |
| Twos, threes, sixes, sevens | Split against weaker upcards, hit or stand against stronger ones | Dealer bust probability against a low upcard |
| Fours | Usually hit or split only in specific rule sets | Total of 8 is weak; split value is marginal |
| Nines | Split against most upcards, stand against a dealer 7, 10, or ace | 18 is strong but can be beaten by a made dealer hand |
The nines row is the clearest example of how the dealer's upcard changes the logic. Against a dealer 7, 10, or ace, an 18 is likely to lose anyway if the dealer makes a hand, so the charts generally say stand and take the single outcome. Against a dealer 6, splitting nines lets you attack a weak upcard with two strong starting hands. The decision is not about whether 18 is a good hand in the abstract. It is about whether two hands starting with a 9 each beat the expected value of one hand starting with 18 against that specific upcard.
The Procedural Mistakes That Cost New Players
The first mistake is splitting tens. It looks aggressive and feels like a way to win twice, but it breaks a 20 into two hands that each need to draw well to reach a competitive total. The house edge on the original hand is already in your favour; the split hands are not.
The second mistake is treating a strategy chart as a universal law. A chart is a set of decisions derived from a particular rule set. Change the number of decks, the dealer's stand rule, or the doubling and splitting rules, and the correct decision on marginal pairs can flip. The fix is not to memorise harder. It is to read the table rules before you sit down, and to use a chart that matches them.
The third mistake is procedural: not knowing the hand signal for a split. In a face-up shoe game, you place the second wager next to the first and point with two fingers or otherwise indicate a split according to the table's convention. In a pitch game where cards are dealt face down, you place the second wager and place the two cards face up on the layout. Getting this wrong can lead to a misplayed hand, and a misplayed hand is settled according to what the dealer understood you to do, not what you intended.
The fourth mistake is resplitting without checking the rule. Some tables allow a third split after a second pair is dealt; many do not. If the table does not allow resplitting aces, and you receive another ace on a split ace, you are stuck with the hand as dealt. The time to learn this is before the cards arrive, not after.
What the Charts Cannot Do
A split decision changes the expected rate of loss on a hand. It does not change the built-in edge of the game, and it does not guarantee a win. Two strong starting hands can both lose to a dealer's made total. One weak hand can win when the dealer busts. The charts are about playing the hand that the probabilities favour, over a very large number of repetitions, not about controlling any single round.
That is the honest boundary. Basic strategy, including split decisions, reduces the expected loss on the hands you play relative to guessing. It does not reverse the house edge, and it does not turn a negative-expectation game into a source of profit. It is a method for losing more slowly, which is the only thing a strategy chart has ever been.
FAQ
Do you always split aces and eights?
Across nearly every published rule set, yes. A pair of aces as a single hand is a weak soft 12, and a pair of eights is a 16, the worst starting total against a strong dealer upcard. Splitting each pair creates two starting hands with a better expected outcome than the single hand they came from.
Can you split any two ten-value cards?
Some tables allow it, and many do not. A king and a queen are both ten-value cards, but whether the table treats them as a splittable pair is a rule printed on the layout or stated by the dealer. Even where it is allowed, the charts treat a total of 20 as a stand, not a split.
Does the number of decks change split decisions?
It can, on marginal pairs. Deck count is one of the rule-set variables that published basic-strategy charts are derived from, along with the dealer's stand rule and the doubling and splitting options. A chart built for one rule set may give a different answer on a pair like fours or nines than a chart built for another.
Play responsibly
Gambling should be treated as paid entertainment, never as a way to earn income or recover losses. 1-800-MY-RESET (1-800-697-3738)
18+ or 21+ depending on where you are; follow the minimum age that applies to you.
This article is general information about how these mechanics work. It is not legal advice and not a recommendation to gamble or to use any particular operator. Availability and legality differ by jurisdiction — check the rules that apply where you are.