Splitting is the single biggest source of leak on the basic strategy chart for recreational players. The chart only has ten pair rows — but they include the two highest-EV plays in the game (split aces, split eights) and the most counter-intuitive ones (split 9s against 9, don't split 10s ever).
This is the splits-only cheat sheet for 4-8 deck blackjack with DAS (double after split) allowed. Memorize the two universal rules first, then the eight conditional rules. Drill until each pair-vs-upcard is a reflex.
The two universal rules
These are non-negotiable. They apply at every blackjack table, every rule set, every count, every time:
- Always split Aces (A,A). Two starting Aces is the highest-EV starting hand in the game. Splitting turns a hand worth +18% expected return (treated as 12) into two hands worth roughly +50% each.
- Always split Eights (8,8). A pair of 8s is a 16 — the worst total in blackjack. Splitting it into two hands starting with 8 is a loss-mitigation move that turns a -54% hand into two -10% hands. You'll often lose both halves; you'd lose more by standing.
The two never-split rules
- Never split Fives (5,5). A pair of 5s is a hard 10 — one of the best starting totals in the game. Split it and you have two hands starting with 5, each one a stiff against most dealer cards. Treat 5,5 as hard 10: double against 2-9, hit against T-A.
- Never split Tens (T,T). 20 is the second-best hand in blackjack. Splitting it gives up a near-certain win in exchange for two hands that each have to beat the dealer independently. A common tourist instinct; mathematically a disaster.
The conditional pairs
These six pairs split against some dealer upcards and not others. Memorize the upcard ranges.
2,2 and 3,3 — small pairs
Split against dealer 2-7. Hit against dealer 8-Ace. Without DAS, narrow to splitting only against dealer 4-7. The logic: a 4 or a 6 (your starting card from a 2,2 or 3,3 split) is a hit-and-pray hand against a strong dealer card, and you don't get the upside of doubling on a 10 or 11 in the no-DAS case.
4,4 — the small-edge split
Split only against dealer 5-6 (and only if DAS is available). Hit against everything else. A pair of 4s is a hard 8, which doesn't bust on any one-card draw, so the split is only worth it against the two cards where the dealer is most likely to bust.
6,6 — split low
Split against dealer 2-6. Hit against 7-Ace. Without DAS, don't split 6,6 against dealer 2. The pair-of-6s split is a defensive play — you're turning a 12 (which busts a lot) into two starting-6 hands that have more flexibility.
7,7 — split through the middle
Split against dealer 2-7. Hit against 8-Ace. A pair of 7s is a 14 — almost as bad as a 16. The split is similar in spirit to the 8,8 split: turn one terrible hand into two mediocre ones.
9,9 — the trick pair
Split against dealer 2-6, 8, and 9. STAND against dealer 7. STAND against dealer 10, J, Q, K, A. The 9,9 row is the most-misplayed cell on the chart.
The logic: 18 is a winning total against dealer 7 (because the dealer's most common total against a 7 upcard is 17). It's a probable losing total against dealer 8 or 9, so you split to try to improve. Against dealer 10 or Ace, you keep the 18 because both halves of the split would face stronger dealer hands than 18 already loses to.
Pair splitting with no DAS
DAS (double after split) lets you double on a split hand if the first card you draw makes a doublable total. Without DAS, the splitting strategy tightens because you lose the upside of doubling a soft 16 or hard 10 on a split half:
- 2,2 and 3,3: split vs 4-7 instead of 2-7.
- 4,4: never split.
- 6,6: split vs 3-6 instead of 2-6.
- Everything else stays the same.
Why splitting is hard to learn
Splits are the row of the chart with the most counter-intuitive logic. The other cells follow patterns (stand on hard totals against bust cards, hit on stiff totals against strong cards, double when the dealer has a weak upcard). The split row mixes loss-mitigation (8,8 vs anything), exploitation (split 9,9 vs 6), and abstention (stand on 9,9 vs 7) in ways that don't follow one rule.
Drilling pairs in isolation, repeatedly, is the fastest path to muscle memory. Most basic-strategy drill apps shuffle in pairs at the same rate they appear in real play — which means a typical session has maybe 3-5 pair decisions out of 50 hands. That's not enough reps.