Everything you really want to know about splitting nines in blackjack

Everything most blackjack buffs want to know about playing pairs of nines can be stated in a single sentence. Split 'em against a dealer's two through six, eight, or nine; otherwise stand.

Players with enquiring minds, who want to know, might be interested in delving deeper, however. In particular, why is it a good idea to break up what seems like a strong 18 and risk finishing with one or more bets riding on a flimsier total?

The criterion underlying this and all other dictates of Basic Strategy is statistical. The "correct" action maximizes the "expectation" (think of it like an edge) of a particular combination of player's hand and dealer's upcard. It's the option that yields the greatest average gain or lowest average loss per dollar bet at the start of the round, as determined by the values and probabilities all possible player and dealer outcomes.

For the case of a pair of nines, when the dealer's upcard is two through six, or eight, 18 is moderately strong. Standing in these situations will yield a profit averaged over large numbers of trials. An 18 isn't as good as most folks think, though. Of the five totals a dealer can reach without busting, three beat it.

An initial nine is also potent, and kicking off with two nines offers higher average profit than standing on 18 when facing the indicated upcards. This, even after matching the original bet by splitting and therefore being vulnerable to a double whammy.

Splitting 18 isn't the same as being dealt two totals of nine, to be sure. After splitting, you automatically get a second card on each side and play whatever total arises. If you pull a deuce on a split-off nine, you can double on the resulting 11. An ace or 10 puts you ahead of your original 18. Another nine, depending on house rules, lets you resplit or leaves you where you were at the outset. Anything else and you're weaker. Alternately, starting with a two-card nine, you can double-down against three- through six-up, and are favored by hitting against two, seven, or eight.

A pair of nines against a nine is a Basic strategy split, but for a different reason. You can improve on either or both sides of the split with a 10 or ace as your automatic card, sitting pretty with positive expectation by standing on the 19 or 20. The chance of such a draw is five out of 13. A deuce, one chance out of 13, sets you up to double with a strong 11. A nine leaves you where you were or with the chance to split again, according to house rules. The other seven chances out of 13 make matters worse. The bad news is that you're an underdog regardless of whether you split or stand. The good news is that splitting leads to less of an average loss over many occurrences of this combination.

Some solid citizens wonder why, if splitting nines is good, why splitting 10s is one of the game's most proscribed practices? Statistical calculations provide the precise answers.

Intuitively, the reasons are threefold. First, a 20 is a much stronger starting position than an 18 – only one rather than three dealer stopping points beat it. Second, no automatic second card forms a hand with a 10 on which you'd double under normal circumstances. And, third, the likelihood of a split-off 10 improving over the original 20 is small.

Of course, if you don't have the cash in reserve to be comfortable with the extra money needed to split nines or any other pairs, you needn't do it. In such cases, you might think about splitting only when doing so gives you the biggest benefit. All other things being equal, the upcards in the order yielding the most to the least gain in expectation, are five, six, four, eight, three, nine, two.

Anyway, who's to say you won't win by flouting the math and splitting or not based on your intuition or premonitions, rabbit's foot, doubt about "who wrote the 'book," or inability to think clearly under pressure? Well, some purists may say so. Here's how the poet, Sumner A Ingmark, described their attitudes:
It's fine if you work using special criteria,
But those who act blindly show signs of deliria.