Why Split Aces and 8s in Blackjack?

These are the two pairs basic strategy splits against almost every dealer upcard, with one narrow exception for 8s. Here are the actual numbers behind both.

Short answer

Split both. A pair of Aces is split because keeping them together caps the hand at a weak soft 12. A pair of 8s is split because playing them as a hard 16, the single worst total in blackjack, is worse than playing two hands starting from 8. Both hold against every dealer upcard under the site's default rules. The one exception: if the dealer hits soft 17 and late surrender is allowed, surrender 8s against a dealer Ace.

The EV of splitting vs the next-best play

These are the numbers the trainer shows, under the site's default rules: dealer stands on soft 17, double after split allowed, late surrender allowed. Split EV is the total for both hands, per $1 of your original bet. The values for a dealer 10 or Ace include the chance that the dealer has blackjack.

S = stand, H = hit, R = surrender. Splitting wins in every column, and usually by a wide margin. Against a dealer 9, 10 or Ace, splitting 8s still loses money on average (-0.405, -0.529, -0.581). The point isn't that it makes money. It loses less than any alternative.

Why Aces are split

A pair of Aces played together is a soft 12: one Ace has to count as 1, and a 12 is a weak total. Splitting gives two separate hands, each starting with an Ace and a strong chance of drawing into a total near 21. Most tables only allow one card per hand after splitting Aces, and the numbers above already assume that. Even so, splitting Aces is correct against every dealer upcard, and against a dealer 6 it's worth about 48¢ more per $1 than hitting.

Why 8s are split

A pair of 8s is a hard 16, too high to hit comfortably and too low to stand on with any confidence. Splitting turns that one bad hand into two hands that each start at 8, a meaningfully better starting point than 16 against every dealer upcard, including strong ones like 9 or 10. With late surrender available, surrendering a dealer 10 is close (-0.538 vs -0.529 for splitting) but still worse. Splitting also beats surrender against a dealer Ace when the dealer stands on soft 17.

The one exception: 8s vs Ace under H17

When the dealer hits soft 17, a dealer Ace becomes stronger. With late surrender allowed, surrendering 8,8 against an Ace (-0.654) then beats splitting (-0.672). Without surrender, you still split. This is the only spot where basic strategy doesn't split Aces or 8s.

About these numbers

The numbers come from the site's probability engine, which uses an infinite-deck approximation of the default 6-deck game. It models one split per pair (no resplitting), one card on each split Ace, and doubling after splitting 8s. Tables that allow resplitting are slightly better for the player than shown here, but the split decisions themselves don't change.

What makes these two different from other pairs

Most pair decisions depend heavily on the dealer's upcard: 9s split against some cards and stand against others, and 4s only split with Double After Split available. Aces and 8s are the exceptions. The starting hand they'd otherwise be stuck with (soft 12 and hard 16) is bad enough, and the split alternative good enough, that the dealer's upcard barely moves the conclusion. See the full basic strategy chart for every other pair, or read Splitting Pairs for the complete reasoning, including why 10s are the mirror-image case that's never split.

Practice it

The trainer shows the probability and expected value behind splitting Aces and 8s as you play, so you can see exactly how much value the split adds over keeping the pair together.

Practice Split Decisions