The decision that changes how many hands you are playing
You sit down, buy in, and the first two cards arrive as a pair. The dealer asks whether you want to split. Most new players answer from feel: a pair of tens feels strong, a pair of fours feels weak. The better answer comes from knowing that a split is not a single decision about one hand. It is a decision to turn one hand into two separate wagers, each with its own draw, its own outcome, and its own exposure to the dealer's total.
When the question is which hands should always be split, the standard answer covers two starting pairs: aces and eights. The reasoning is not a slogan. It follows from how blackjack hands are built and what each of those pairs becomes when it stays together.
What splitting does to the hand
In blackjack, the objective is a hand total closer to 21 than the dealer's without exceeding it. When the first two cards match in rank, most rule sets let you separate them into two starting hands, each beginning with one of the paired cards. You place a second wager equal to the original, and the dealer then deals a second card to the first split card and plays it out before moving to the next.
The split changes three things at once. First, it doubles the amount of money at risk on that round. Second, it replaces one total with two independent totals, each of which can win or lose on its own. Third, it consumes a decision point: after the split, the normal options of hit, stand, double, or split again depend on the exact table rules in force.
Because the split doubles the wager, it should only happen when the two resulting starting points are better than the single total they replaced. That is precisely where aces and eights stand apart from most other pairs.
Why aces and eights are the standard always-split pairs
A pair of aces is a soft total of either 2 or 12. As a single hand, it is awkward: any ten-value card makes 12 again, and the hand has no natural strength. Split the aces, and each new hand starts with an ace, which is the most flexible first card in the game. A ten on either ace makes 21. The catch is that many tables restrict what happens after splitting aces, often allowing only one additional card per ace and no resplitting. That restriction does not make splitting wrong; it just means the value of the split is realised through the single draw each ace receives.
A pair of eights totals 16, widely understood as the least comfortable starting total in blackjack. Standing on 16 against a strong dealer card loses often, and hitting 16 risks an immediate bust on any card above 5. Splitting the eights turns one poor total into two hands that each begin with an 8, a far more workable starting point. The split does not promise a win. It replaces a hand with a high expected loss with two hands whose combined expected loss is smaller.
The rule to remember is structural, not mystical: split when the pair's total is worse than the two single cards would be. Aces and eights fail as totals and succeed as starting cards. Most other pairs do not fail in the same way.
Where the always-split rule bends
No blackjack rule survives contact with a different rule set. The instruction to always split aces and eights is a default that assumes the common version of the game. A few variations change the arithmetic enough that a published strategy chart will show a different line.
One example is what the dealer does on a soft 17. Another is whether doubling after a split is allowed. Another is whether the game uses one deck, two decks, or a shoe, because the proportion of ten-value cards left changes with deck composition. Another is whether resplitting is permitted and how many times. Each of these changes the expected loss on a split hand, and therefore changes the correct play in borderline situations.
The comparison is easiest to see when the options are laid out side by side.
| Pair | Total if kept | What the split creates | Why the default play is what it is |
|---|---|---|---|
| Aces | Soft 2 or 12 | Two hands each starting with an ace | An ace is a strong first card; the pair's total is weak |
| Eights | 16 | Two hands each starting with an 8 | 16 is a poor total; 8 is a neutral starting point |
| Tens | 20 | Two hands each starting with a 10 | 20 is already strong; splitting risks making two weaker hands |
| Fives | 10 | Two hands each starting with a 5 | 10 is a good doubling total; splitting breaks it apart |
The table shows the principle: split when the pair is a liability as a total, not when the pair simply contains cards you like. Tens and fives are the mirror image. A pair of tens is already 20, a total that wins most of the time without any further action. Splitting it risks turning one strong hand into two mediocre ones. A pair of fives totals 10, which is often a doubling opportunity, and splitting it produces two weak starting points.
How the round moves from buy-in to settlement
Understanding the split decision is easier when the whole round is visible. You buy in at the table and receive chips. You place a wager in the betting circle. The dealer gives each player two cards, usually face up, and takes two cards with one face down. If your two cards are a pair, the split option appears when the table rules permit it.
To split, you place a second wager equal to the first beside the original bet. The dealer separates the two cards and deals a second card to the first split card. You then play that hand to completion under the table's rules. After it settles, the dealer deals a second card to the other split card and you play that hand. Settlement happens hand by hand: each split hand is compared with the dealer's total independently, and each pays or loses on its own.
The sequence matters because a split is not a single result. One split hand can win while the other loses. The total outcome of the round is the sum of the two settlements, which is why the decision to split is really a decision about expected loss across two wagers rather than one.
The procedural mistakes that cost new players
The first mistake is splitting a pair because the cards feel valuable. A pair of tens feels valuable, and it is, but as a total of 20. Splitting it destroys the value the pair already has. The question is never whether the individual cards are good. It is whether the pair's total is worse than the two single starting points the split would create.
The second mistake is assuming the split changes the house edge on the game. It does not. The house edge is a property of the game under its specific rule set. Splitting correctly changes the expected loss on that particular hand. It does not convert the game into a positive-expectation situation, and no split decision can do that.
The third mistake is treating a remembered rule as universal. A player who has memorised "always split aces and eights" without checking the table rules will occasionally make the wrong play. The correct play derives from the exact rule set in use: the number of decks, whether the dealer stands or hits on soft 17, whether doubling after a split is allowed, and whether resplitting is permitted. Published basic-strategy charts are built for specific rule sets, and a chart for one game is not automatically correct for another.
Reading a strategy chart without treating it as scripture
A basic-strategy chart is a grid that lists the player's total down one side and the dealer's upcard across the top. Each cell gives the play with the lowest expected loss for that combination under the rule set the chart was built for. The split decisions occupy their own section, because they apply only to paired starting cards.
The chart is not a promise about the next hand. It is a statement about the long-run average across a very large number of identical situations. A player who splits eights against a dealer ten will still lose that hand often. The chart says only that, across many repetitions, splitting loses less than standing or hitting would.
When the table rules differ from the chart's assumptions, the correct play can change. The way to check is not to guess. The table's rules are printed or announced at the table, and the game's help screen or rules page states the rule set for an online table. The player's job is to match the chart to the game in front of them, not to force the game to match a remembered chart.
FAQ
Is it ever wrong to split aces?
Under the common rule set, no. The restriction that usually applies after splitting aces, such as receiving only one additional card per ace, reduces the value of the split but does not remove it. The pair's total as a single hand is weak enough that the split remains the lower expected-loss play.
Why do some players split nines but not tens?
A pair of nines totals 18, which is a good but not dominant total. Against certain dealer upcards, splitting the nines produces two starting hands that each have a better combined expectation than standing on 18. A pair of tens totals 20, which is already strong enough that splitting usually makes both hands worse. The difference is the strength of the total before the split.
Does splitting change how much I can win on the round?
It changes the structure of the round, not the game's mathematics. A split doubles the wager and creates two independent settlements, so the round's outcome is the sum of two hands rather than one. It can produce two wins, two losses, or one of each. The expected loss across the two hands is what the strategy chart minimises, not the chance of any single outcome.
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.