Resplitting aces (RSA) is a favorable blackjack rule, but it is a small one. It matters only after a player is first dealt A-A, splits the pair, and then receives another ace on one of the new hands. If the table allows RSA, that new pair of aces can be split again. If RSA is prohibited, the player must continue under the table's split-ace restrictions.
Because the triggering sequence is rare, RSA usually changes the house edge by only a few hundredths of a percentage point in common multi-deck games. The exact value depends on deck count, the maximum number of hands, whether split aces receive only one card, and the rest of the blackjack rules.
That makes RSA useful as a tie-breaker between otherwise similar tables, not a reason to ignore larger rules such as a 6:5 blackjack payout.
What has to happen before RSA matters
The rule has no effect on an ordinary hand. It becomes relevant only through this chain:
- The player receives two aces as the original hand.
- The player splits them and places a second equal wager.
- At least one split ace receives another ace.
- The table either allows that new A-A hand to be split again or stops further ace splitting.
If the second card to each split ace is anything other than an ace, the RSA rule never enters the hand.
Why the trigger is rare
In a six-deck shoe there are 24 aces among 312 cards. The probability that the first two cards to one player are both aces is:
C(24,2) / C(312,2) = 276 / 48,516 ≈ 0.569%So the original A-A hand appears only about once in 176 two-card deals before considering dealer and other-player cards.
After those two aces are removed, 22 aces remain among 310 cards. If each split ace receives one card, the rough probability that at least one of those two cards is another ace is:
1 − (288/310 × 287/309) ≈ 13.7%Multiplying the two events gives an approximate trigger frequency around:
0.569% × 13.7% ≈ 0.078%That is roughly eight times per 10,000 original hands before adjusting for cards exposed elsewhere at the table. The exact frequency changes with composition and procedure, but the calculation explains why RSA can be valuable when it occurs while still having only a small effect on the overall house edge.
Why an extra ace split has value
Aces are unusually powerful starting cards because many one-card draws produce strong totals:
- A + 10-value card = 21;
- A + 9 = 20;
- A + 8 = 19;
- A + 7 = 18.
When a new ace lands on a split ace, leaving the hand as A-A under a one-card-only rule, the player loses the opportunity to turn those two aces into two separate starting hands. RSA restores that opportunity.
The benefit is not that the player receives “free hands.” Every resplit requires another wager equal to the original stake. The value comes from being allowed to place that additional wager in a favorable situation—starting a new hand with an ace instead of accepting a weak restricted A-A hand.
RSA does not remove other split-ace restrictions
This is one of the most important distinctions on the page.
A table can allow resplitting aces and still enforce all of the following:
- only one additional card to each split ace;
- no hitting after that card;
- no doubling on a split-ace hand;
- an ace plus a ten-value card after the split counting as ordinary 21 rather than a natural blackjack;
- a maximum of three or four total hands.
RSA changes only the permission to split a new pair of aces again. It does not automatically make split aces play like ordinary hands.
For the procedural rules themselves, see [resplitting aces](/blackjack/resplitting-aces/) and [blackjack splitting rules](/blackjack/splitting-rules/).
Turning a small rule change into expected dollars
Suppose two otherwise identical games differ only in whether RSA is allowed. For illustration, assume the no-RSA game has a 0.50% house edge and allowing RSA improves the game by 0.03 percentage points, lowering the edge to 0.47%.
At $25 average action over 1,000 original hands:
Total original-hand action = $25 × 1,000 = $25,000
Expected loss at 0.50% = $125.00
Expected loss at 0.47% = $117.50
Difference = $7.50The dollar difference is real, but it is small relative to ordinary blackjack session variance. A player can easily win or lose hundreds or thousands of dollars while the long-run RSA value over that sample is only a few dollars.
This is why RSA should be used to compare similar games rather than treated as a dramatic advantage.
Rule value is not the same as rule frequency
A rare rule can still matter. The correct question is not “How often do I see it?” but “How much expected value does it add when averaged across all hands?”
RSA is a good example:
- the triggering situation is rare;
- the conditional improvement when it occurs is meaningful;
- after averaging that improvement across all hands, the total house-edge reduction remains small.
This is the same reason surrender, resplitting, and some double rules can have measurable but modest effects. The house edge is the sum of many conditional decisions and settlement rules.
Current regulated rules show RSA is a separate option
The Massachusetts Gaming Commission's active blackjack rules, dated April 9, 2026, are a useful example of how ace resplitting is treated separately from ordinary pair splitting. The rules permit a licensee, with proper notice, to allow multiple pair splits while also permitting the casino to prohibit a player from splitting a pair of aces more than once. The same rules state that split aces receive only one additional card.
The current rule set is published at Blackjack Rules — 4.9.26.
That distinction is exactly what players should look for. “Resplit up to four hands” does not necessarily mean “resplit aces.” Ace treatment can have its own limitation.
RSA versus other blackjack rule changes
If two tables are otherwise equal, choose RSA. But if the tables differ in several ways, prioritize the larger effects first.
A practical comparison order is:
- Blackjack payout. A 3:2 game is usually much stronger than a comparable 6:5 game.
- Dealer soft-17 rule. S17 is normally better than H17 when other conditions match.
- Double and surrender rules. DAS and surrender can matter more than RSA.
- Split restrictions. General resplit rules and RSA add smaller incremental value.
The exact ordering can change with the full ruleset, but the main point remains: do not sacrifice a major favorable rule to gain RSA.
The [house edge by rules](/blackjack/house-edge-by-rules/) guide is designed for full-table comparisons.
What RSA changes for basic strategy
RSA does not rewrite the entire basic-strategy chart. The initial decision with A-A remains a split in standard blackjack conditions. The rule changes what happens later if another ace arrives after the split.
That means a player does not need a completely different strategy card just because RSA is available. The player needs to know the table's split limit and whether the newly formed A-A can be split again.
This is also why casinos must train dealers and floor supervisors on the exact local rule. A dealer who incorrectly refuses an allowed ace resplit—or allows one where the rule prohibits it—creates a real financial and dispute issue.
Maximum hands still matters
“RSA allowed” can mean different things depending on the maximum number of hands.
For example:
- Split once only → two hands total;
- Resplit once → up to three hands total;
- Resplit up to three times → up to four hands total.
If the table allows RSA but caps play at three hands, a later ace can still be blocked once the cap is reached. The useful question is therefore not merely “Can I resplit aces?” but “How many total hands can aces be split into?”
The practical value of RSA
Resplitting aces is a favorable rule because it gives the player another chance to turn two aces into separate strong starting hands. Its effect on the overall game is small because the required sequence appears infrequently.
When comparing two genuinely similar blackjack tables, RSA is worth taking. When comparing a strong 3:2 game without RSA against a weak 6:5 game with RSA, the larger rules dominate.
RSA improves blackjack, but it is a precision improvement—not a rescue rule.