Basic strategy is mostly a set of conditional rules: this play against these upcards, that play against those. Two rows are different. Aces and 8s split against every dealer upcard, from a 2 to an ace, with no branching and no exceptions under standard rules. They are the only rows on the pair chart with that property, which is why they are the first thing anyone teaching the game says out loud.
Repeating the rule is easy. Understanding it is more useful, because the two rows are correct for completely different reasons: one is the best hand in the game trying to get out of a bad container, and the other is the worst hand in the game trying to become anything else.
Aces: turning the worst 12 into two of the best 11s
Held together, a pair of aces is a soft 12. That sounds harmless and is close to useless. You cannot bust it, but you also cannot do much with it: hitting a soft 12 leaves you in the same place as often as not, and standing on 12 is surrender by another name.
Split them and each ace becomes the start of a soft 11, which is the strongest opening card in blackjack. The reason is arithmetic about the deck. Ten-value cards (tens, jacks, queens, kings) are sixteen of every fifty-two cards, far more than any other rank, so the single most likely card to land on a split ace is a ten. That makes 21 immediately.
And when it is not a ten, an ace is still forgiving in a way no other card is. A 9 gives you 20, an 8 gives you 19, a 7 gives you 18, and even a low card leaves you a soft hand that can be improved without risk of busting. You have converted one hand with almost no upside into two hands that each start from the best position the game offers. That advantage is large enough to overwhelm every dealer upcard, which is why the row never branches.
The catch: one card only, and no 3 to 2
Splitting aces comes with a restriction almost every casino imposes: each split ace receives exactly one card, and that is the end of the hand. No hitting, no doubling, no drawing your way out of a bad landing. If you split aces and catch a 4 and a 5, you are standing on soft 15 and soft 16 and there is nothing further to be done.
The second detail catches people out at the table. A ten on a split ace is 21, but it is not a blackjack. A natural is specifically an ace and a ten dealt as your first two cards, and a hand assembled after a split does not qualify. It pays even money, not the 3 to 2 that a natural pays. Nobody is cheating you when the dealer pays one to one on it; that is the standard rule everywhere.
Both restrictions are real costs, and both are already baked into the strategy. The published charts that tell you to split aces against every upcard were computed with the one-card rule in place. Splitting is still correct, comfortably, which tells you how large the underlying advantage is.
21 TrainerDrill the whole pair block with instant feedback and the reason behind every split. Free on the App Store.
Get the app
Eights: escaping the worst hand in blackjack
The 8s row is correct for the opposite reason. There is nothing promising about a hand starting with an 8. What there is, is something genuinely terrible about the alternative.
Two 8s played as one hand is a hard 16, and hard 16 is the worst total in blackjack. It cannot be stood on comfortably because it loses to every dealer total from 17 up, and it cannot be hit comfortably because any card above a 5 busts it. Every option is bad. Our article on 16 against a dealer 10 works through just how bad in detail.
Splitting throws the 16 away. Each new hand starts at 8, which is not strong but is at least a live hand: it can draw to 18, 19, 20 or 21 without busting on a single card, and it can improve on nearly anything it receives. Trading one hopeless hand for two ordinary ones is worth a second bet, and the calculation holds against every upcard.
Against a 10 or an ace, splitting 8s still loses
This is where the rule stops feeling intuitive, so it is worth stating plainly: splitting 8s against a dealer 10 is a losing play. You will lose money on that hand over the long run, and you will lose it twice as often as you would lose a single bet, because there are now two bets on the felt.
It is still correct. Basic strategy does not choose the winning play, because against a 10 there is no winning play available. It chooses the least losing one, and the published expectation for two hands starting at 8 is better than the expectation for one hand of hard 16, even after doubling your exposure. Standing on 16 against a 10 is worse. Hitting it is worse. Splitting loses least.
If that reasoning feels wrong, it is the same reasoning that governs half the chart, and it is worth internalizing once rather than fighting hand by hand. The framework is laid out in our piece on expected value: every decision is a comparison between options, not a judgment about whether the hand is good.
One genuine exception: at a table offering late surrender, some published charts prefer surrendering a pair of 8s against a dealer ace under H17 rules rather than splitting. Surrender is not offered everywhere, so this only applies if it is on your table, and it is a narrow call rather than a general revision of the row.
Rule details that change the value of these splits
The plays themselves are stable. What varies is how much they are worth, and a few rules move the number.
| Rule | Effect on these two rows |
|---|---|
| Resplit aces allowed | Lets a third ace start another hand instead of stranding you on soft 12; worth about 0.08 percent per the commonly published rule effect tables |
| Split to four hands | Matters more for 8s than aces, since a third 8 is common enough to see |
| Double after split (DAS) | Helps 8s meaningfully: an 8 catching a 3 becomes a doublable 11 |
| European no hole card (ENHC) | The dealer's blackjack can take your additional split bets, so some published charts change the 8s play against a ten or an ace |
DAS is the one most players underrate. Half the value of splitting 8s comes from the hands that develop into doubling positions, and a table that forbids doubling after a split quietly takes that away. The ENHC caveat is the one to check when playing outside the United States: because the dealer does not check for blackjack before you act, the extra chips you commit on a split are at risk in a way they are not in the American game, and charts for that variant are computed separately. Use a chart built for your actual rules rather than assuming the American version transfers.
How to lock the two rows in
These are the two easiest rows on the chart to memorize and two of the easiest to fumble anyway, because the hands are rare and the 8s row asks you to make a play that feels wrong against a face card. The fix is repetition against the uncomfortable upcards specifically: 8s against a 9, a 10 and an ace, until splitting is what your hands do before your brain files an objection.
The rest of the pair chart is genuinely conditional, and it is worth learning alongside these rows rather than after them, which our pair splitting guide covers row by row. Aces and 8s are just the two you never have to think about, which is exactly why they are the place to start.
Set your rules, see the chartCharts redraw for DAS, split limits and ENHC so you drill the game you actually play.
Download
Frequently asked questions
Why do you always split aces in blackjack?
Because a pair of aces played as one hand is a soft 12, which is close to worthless, while two hands starting with an ace each are soft 11s, the single best starting position in the game. Any ten-value card, the most common denomination in the deck, turns a split ace into 21. Standard charts split aces against every dealer upcard without exception.
Why do you always split 8s, even against a 10?
Because hard 16 is the worst total in blackjack and two hands starting from 8 are better than one hand of 16, even against a strong upcard. Splitting 8s against a 10 is not a winning play in absolute terms, it is a losing play that loses less than standing or hitting a 16. Where late surrender is offered, some charts prefer surrendering 8s against an ace, so check your rules.
What happens if you get a ten on a split ace?
You have 21, but it is not a blackjack. Almost every casino deals only one card to each split ace and no further draws, and a hand made after splitting pays even money rather than the 3 to 2 a natural pays. That is a real cost, and it is already priced into the strategy: splitting aces is still correct against every upcard despite the restriction.
Can you resplit aces if you get another one?
It depends on the table. Many games do not allow it, so a third ace simply becomes a soft 12 on that hand with no further cards. Where resplitting aces is permitted, it is a genuine player benefit worth roughly 0.08 percent per the commonly published rule effect tables. Read the rules before you sit, because the two versions play differently on a hand that already comes up rarely.
Are aces and 8s always split under every rule set?
Under standard American rules with a dealer hole card, yes, both rows split against every upcard. The exception worth knowing is the European no hole card game, where the dealer's blackjack takes the additional bets you put out, and some published charts for that variant change the 8s play against a ten or an ace. Always use a chart computed for the exact rules you are playing.