Why You Always Split a Pair of Aces or Eights
By Arnold Snyder
(First published in Player Magazine)
© 2005 Arnold Snyder
Aces and eights.
Poker players know this as the “Dead Man’s Hand.” Legend tells us that these are the cards Wild Bill Hickok was holding when he was shot down at a poker table in Deadwood, South Dakota, back in 1876.
To basic strategy blackjack players aces and eights have entirely different meanings. The common wisdom spouted by experienced players, dealers and pit bosses, is: “Always split aces and eights.”
Virtually all players would agree that splitting aces makes sense as blackjack basic strategy. Who wouldn’t trade one hand starting with a total of 12, for two hands starting with 11 each? Only a moron would keep the 12.
But eights?
Sure, it makes sense to split those eights in two when the dealer is showing a potential bust card. Who wants a blackjack hand totaling 16? No one. Ever. So if the dealer’s got a five up, or a deuce, or any other pitiful low card, I’d rather take my chances with two hands starting with 8, than one lousy, rotten 16.
But that’s as far as the “common wisdom” may seem to make sense. When the dealer’s got a scare card showing-any nine, ten, or ace-why on earth would I want to split my 8s? Sure, I know a 16 still looks like a loser. But two hands starting with 8 each against these scare cards just looks like two losers.
Or, so it seems…
In Beat the Dealer, Ed Thorp says that one of the hands that convinced the pit bosses that he was a complete fool back when he first hit Las Vegas as a blackjack card counter back in 1960 was a pair of eights. As per his computer analysis, he always split them. Before computers came along, it was not common wisdom to always split 8s. Is there an understandable logic behind this basic strategy play?
As a matter of fact, there is.
What if, instead of splitting 8s, casinos allowed us to “toss” one eight whenever we were dealt a pair, and take our chances instead with whatever card the dealer dealt us to replace it?
That’s a no-brainer. A total of 8 isn’t nearly so bad a start on a blackjack hand as a total of 16. With the 8, you have a pretty decent chance of drawing a ten for an 18 total. And although I’m not exactly thrilled with the idea of pitting my 18 against the dealer’s ten up I sure do like it better than a 16! And I might even draw an ace, 2, or 3 on my 8, giving me a chance at a much stronger hand than 16.
It’s not hard to see that a hand of 8 is a whole lot better than a 16 against a dealer ten. The 8 still looks like a loser, but nowhere near as bad of a loser. If I could toss one 8, I’d do it in a heartbeat. But casinos don’t give me the toss option.
The option they do give me–splitting–seems bad because I’ve got to put more money on the table on a loser. Whether or not two hands of 8 each is better than one hand of 16 is really a question of which option loses the least. And I can only answer that if I know exactly how much more a 16 total costs me–over time–than an 8 total. How do I figure that out?
In 1956, a small group of mathematicians (Baldwin, Cantey, Maisel, and McDermott–memorialized by blackjack pros ever since as “The Four Horsemen”) used old-fashioned adding machines to calculate the answer to this and every other blackjack basic strategy question by tediously running through the math of every possible outcome. Their conclusion: Split the 8s. Even with twice the money on the table, they found, you’ll fare better than you would taking your chances with that lousy 16.
For example, if the player has a sixteen versus a ten, with a $10 bet on the table, in the long run he will lose, on average, 40 cents on the hand. If he has a pair of 8s and he splits the 8s, so that he has two $10 hands starting with a total of 8, on average he will lose 16 cents on each of these hands, for a total loss of 32 cents. If you play the 16 against a ten instead of splitting the 8s, the long run cost is 8 cents for every ten dollars starting bet.
Splitting 8s versus dealer high cards is a classic defensive play. We know we will lose money on the play, but less than we would otherwise have lost.
Unfortunately, few players of that time believed the The Four Horsemen. Then, in 1962, Ed Thorp ran the same hand through his IBM computer and, in a fraction of the time it took those mathematicians, confirmed their answer. That 16 total is so bad, that when you see two 8s, you should throw the extra bet out there to split them with only one thought in your head: Thank heavens that sixteen isn’t composed of a 9 and a 7, or a ten and a 6! With two 8s, you still have a fighting chance.
So, painful as it is, you should always split those is. But if you ever find a casino that offers the “toss” option, don’t split, toss! You heard it here first. ♠
Never miss another post