The rule
You split a pair, draw to one of the new hands, and it becomes a total worth doubling. Whether you may is a house rule, abbreviated DAS on the felt.
Removing it costs +0.141 percentage points, from 0.512% to 0.653%.
Why it changes splitting decisions
This is the part that surprises people. Some pairs are worth splitting only because of what you can do afterwards.
4,4 is the clearest case. A pair of fours is a hard 8, a perfectly playable hand. Splitting it makes two hands starting on 4, each of which will often draw to 9, 10 or 11 and want to double. With DAS you split against a dealer 5 or 6; without DAS you do not split at all.
2,2, 3,3 and 6,6 shift the same way at the margins, and the full chart on our basic strategy page assumes DAS is allowed.
Reading the table
DAS is usually printed on the layout or noted on the table placard. It is worth about a tenth of what the blackjack payout is worth, so check the payout first and this second.
Frequently Asked Questions
- What does DAS mean?
- Double After Split: you may double a hand created by splitting a pair.
- How much is it worth?
- 0.14 percentage points. Losing it takes the edge from 0.51% to 0.65%.
- Which splits does it change?
- 4,4 most clearly: split it against a 5 or 6 with DAS, never without. 2,2, 3,3 and 6,6 shift at the margins.
- Is it more important than the blackjack payout?
- No. A 6:5 payout costs 9.6 times as much.
Corrected on 17 October 2026: with resplits valued as two hands the engine prices double after split at 0.14 points (was 0.13), so 6:5 costs 9.6 times as much, not 10.2; and the rule is worth about a tenth of the payout, not a fifth.
Corrected on 2 November 2026: this page said the model sits a few hundredths of a percent from an exact shoe analysis. It is about a tenth of a point: Wizard of Odds' simulations give about 0.41% for six decks under these rules, against the model's 0.5117%.
Every figure on this page is computed by an expected-value engine that plays the game out: it derives the dealer's final-total distribution by recursion, evaluates standing, hitting, doubling and splitting at each decision, and takes the best. Rule costs are measured by re-running the whole calculation with the rule changed, not by quoting a published delta. The model assumes independent card draws, which prices a six-deck game about a tenth of a point worse than a shoe does; the rule differences it produces match published values. Rules vary by table, so read the felt.