Blackjack Split Calculator

The form defaults to Pair Rank 8 and Dealer Upcard 6. Rank 8 means a pair of eights; rank 11 would mean aces. Submit those two numbers. The engine returns Split? Yes, and If Not Splitting is suppressed in favour of “Play each hand normally” because the split is taken. Eights versus a six is the textbook illustration: a hard sixteen against a bust card is a miserable total, while two hands of eight, each eligible for a hit or double, have far higher expectation under DAS.

Rank is 2 through 11, not a hand total. Entering 16 will be rejected or clamped; you want 8. Dealer upcard uses 1 for ace and 10 for any ten-value card. Keep dealer at 6 and walk ranks: 11 (aces) always Yes; 10 always No with fallback Stand; 9 versus 6 is Yes; 7 versus 6 is Yes; 6 versus 6 is Yes; 5 versus 6 is No with fallback Double/Hit; 4 versus 6 is Yes; 3 and 2 versus 6 are Yes. That single dealer six therefore splits every pair except tens and fives.

Move the dealer to 7 with pair 9. Nines split versus 2–9 except 7, so 9,9 versus 7 is No. The fallback helper in this tool returns Hit for any pair that is not tens or fives. Read that as “do not split; then play the two-card total on the basic-strategy hard chart,” where 18 versus 7 is a stand. The fallback string is a coarse hint, not a second strategy engine. Aces versus anything, including ace and ten, still return Yes. Splitting aces is almost universally allowed for one card each; further hits may be forbidden.

This calculator does not model re-splitting, drawing to split aces, or DAS restrictions on those aces. It only answers the binary split question for the 6D S17 DAS pair chart. Eights versus ace or ten still split: you would rather have two eights than one sixteen against a strong dealer. Work a negative case to see the No path. Pair Rank 10, Dealer 6. Split? No. Fallback: Stand. Splitting tens turns a 20 into two tens that often become 12–16.

Pair Rank 5, Dealer 10. Split? No. Fallback: Double/Hit, because 5,5 is a ten and tens hit against 10 and ace. Pair Rank 4, Dealer 4. Split? No, because fours only split versus 5–6. Use 8 versus 6 as the Yes anchor, then perturb one field at a time so the exceptions become visible.

Pair splitting exists because a two-card total is not always the best use of two identical ranks. Two eights make sixteen, the worst common stiff. Two aces make either 2 or 12 as a hard total, or 12 as a soft, none of which capture the value of starting two hands at 11. The calculator’s job is to apply the 6D S17 DAS split chart to a rank and an upcard, not to estimate how often the dealer will bust.

Gambling-house practice treated pairs as ordinary totals until the Baldwin group and then Thorp showed that splitting eights and aces is mandatory in expectation, splitting tens is a gift to the house, and the middle pairs (2, 3, 6, 7, 9) depend on the upcard. Braun’s multi-deck work and the later Stanford Wong Professional Blackjack tables are the commercial descendants. Casinos added DAS (double after split) because it is popular and only mildly costly to the house when players actually know the chart; without DAS, some 2s, 3s, and 4s stop being splits.

This tool assumes DAS is on, matching the printed 6D S17 DAS contract used across the blackjack suite. It does not ask whether resplits are allowed, whether aces receive one card, or whether you may hit after splitting aces. Those details move EV by tenths of a percent and change a few close cells in full simulators. They do not change always-split-aces-and-eights or never-split-tens-and-fives.

Limits: the pair rank is a single integer 2–11, so you cannot enter mixed pairs (which are not pairs). You cannot enter a count. You cannot enter a no-DAS table. European no-hole-card rules, where the dealer does not peek and a player’s double or split is lost to a dealer blackjack, make some splits against ace and ten worse; this engine does not apply ENHC penalties. Spanish 21 and match-the-dealer side bets are out of scope.

Historically, splitting was also a bankroll decision: two bets replace one. Recreational players skip 8,8 versus 10 to “avoid putting more money out.” That is a variance preference, not a basic-strategy improvement. The calculator will still say Yes. Treat the output as the zero-count EV ranking under DAS, then decide separately whether your unit size allows two bets. The math does not care about comfort; the session might. Default 8 versus 6 remains the teaching Yes.

Splitting a pair is a comparison of two expected values, not a superstition about identical ranks. One computes the mean settlement of playing the two-card total as a single hand, then the mean of turning the pair into two independent hands, each then played by the rest of the basic-strategy table, with doubling after the split allowed. The cell is Split when the two-hand mean is larger. Aces and eights win that comparison against every upcard. Tens lose it against every upcard: twenty is already a magnificent total. Fives lose it as well; two fives are a ten, and the doubling policy for ten dominates two weak starting fives.

Derivation of the middle pairs follows the same comparison under six-deck combinatorics with the dealer standing on soft seventeen. Twos, threes, and sevens split versus dealer two through seven, where the dealer is bust-prone or mediocre and two small starting totals can still be doubled. Sixes split versus two through six. Fours split only versus five and six, because doubling after split is what makes an eight that becomes nine, ten, or eleven worth the extra unit, and only those two upcards justify the extra bet. Nines split versus two through nine except seven: eighteen already beats a seven, and splitting nines against a seven is giving away a made total. Against eight or nine the two nines still improve on a push-prone eighteen.

Worked grid at a glance. Aces (11) always split. Eights (8) always split, including versus ace and ten: two eights are better than one sixteen against a strong dealer. Twos, threes, and sevens (2, 3, 7) versus dealer 2–7: split. Sixes (6) versus 2–6: split. Fours (4) versus 5–6: split. Nines (9) versus 2–9 except 7: split. Tens (10) never. Fives (5) never, and they are treated as a ten thereafter, so 5,5 versus 6 is a double on the hard-ten row and 5,5 versus an ace is a hit. Default 8 versus 6 is the teaching Yes. Rank 10 versus 6 is the teaching No, with stand as the follow-on total of twenty.

Assumptions that move close cells. American peek on ace and ten means you do not lose the extra split bets to an unrevealed dealer blackjack. No resplit value is added; the comparison is one split, not four tens. No count. European no-hole-card rules tax splits versus ace and ten and can flip a few of those cells; this table does not apply that tax. Removing doubling after split is what first flips fours versus five, then some twos and threes, in published charts. Expected-value gaps are large for aces and eights and small for fours versus five, which is why that cell is the canary.

History is Thorp and Braun after Baldwin. Pair splitting was treated as an ordinary total until those enumerations showed that eights and aces are mandatory in expectation and tens are a gift to the house. Recreational refusal of eight-eight versus ten is a variance preference, not an improvement of the mean. The table still says split. Bankroll size decides whether two units are comfortable; the comparison of means does not care about comfort.