Friday, December 31, 2010

A minor mistake

Last night while playing blackjack I made a mistake regarding the taking of insurance. It was an understandable error, but one I have made before and vowed never to make again, which made it partcularly annoying.

Insurance is offered when the dealer's upcard is an ace. A player who thinks the dealer has blackjack can put up to half his bet on a line on the table, making a side bet on the dealer's hand. If the dealer has blackjack, the insurance bet pays two-to-one. The player loses his main bet but is made whole by the win on the insurance bet. (This assumes the player does not himself have a blackjack. If he doesn, he can choose to take "even money" rather than the higher blackjack payout, coming out twice as good as with the push he would have if he decline even money.)

Insurance and even money are both bad bets most of the time. Unless a player is counting cards, he should never take insurance or even money. But if the remaining deck is rich enough in 10-value cards, the bet is a good one. In my counting system, Knock-Out, insurance is taken when the cout reaches one below the number at which the maximum bet is put out.

Last night I was playing a double-deck game with the cards dealt face-down. I was sitting at first base. When the dealer offered insurance, the count, based on his upcard and my two cards, was where I had to take it. I put the chips on the line. But I hadn't looked to my left to see whether either of the other two players at the table had flipped over a blackjack. After I put the money out, I noticed that the player at third base had done just that, bring the count below the point at which I should have taken insurance.

I would have felt better about my error if the dealer had blackjack anyway, but he didn't.

A little mistake, but one with a cost.

Two blackjack rules my brother hates

My brother, who lives in the Northeast and plays most of his blackjack at one of the huge Indian casinos in Connecticut, recently visited Las Vegas and couldn't stop complaining about the rule requiring dealers to hit soft 17, which has become pervasive here. (A few games are still available, at higher limits, in which the dealer stands on 17. Examples are the double deck games at the Mirage and Aria -- lowest minimum $25 --and in the high-limit room of the M Resort -- $100.)

He also railed against the "no double after split" rule at the Caesars (formerly Harrah's) properties where he gets free rooms and plays. (This rule is also common downtown.)

The first rule, hitting soft 17, costs the basic strategy player 0.2 percent. The second, no double after split, costs 0.13 percent -- at least in theory. But there are key differences in the ways the rules work in practice. I believe that the first is here to stay and likely will spread to those parts of the blackjack world where it is not now the norm. I'm not happy about that -- it's costing me money -- but I understand the logic from the casinos' point of view. The same cannot be said about the second rule, which probably doesn't help the casinos nearly as much as they think and hurts their employees as well as players.

First, a little background: In blackjack, a "soft" hand is one in which an ace can function as either a one or an 11. So a soft 17 is a hand such as ace, six or ace, three, three. Traditionally, dealers would stand on such hands. Hitting the soft 17 gives the house a chance to improve its hand. If the dealer's first two cards are ace, six, the hand is immediately improved if another ace or a two, three or four is drawn. In addition, the dealer can draw a card that seems to hurt the hand -- say a seven, resulting in a total of 14 -- but then can come up with a result better than 17 by drawing one of an even bigger number of cards. Because 10-value cards make up 30 percent of the deck, the dealer will often turn a soft 17 into a hard 17. This might seem to have no effect on the players -- the dealer's total is unchanged -- but it removes a card valuable to them.

On the other hand, this rule gives the dealer an opportunity to bust, giving all the players at the table with hands of less than 17 an unexpected win. This doesn't happen often enough to offset the harm the rule does to the players, but it happens often enough to add to the suspense and excitement of the game. Because of this, the typical player, who probably has no idea of the statistical disadvantge of the rule, probaly does not mind this rule.

From the house's point of view, the rule is always applied perfectly and there is no way the statistical gain it offers can be lost. Except to the relatively rare knowledgeable player -- not a desirable customer to the casino anyway -- there is no apparent downside to this rule.

The story concerning no double after split is different in key ways. First, where it is in force, there is usually a sign to that effect on the table. Even the relatively unsophisticated player is being told he is being deprived of a privilege he knows he has elsewhere; who wants to be told that? Second, to some degree, the rule is self-defeating to the casinos in pure financial terms. The 0.13 percent edge is provides assumes all players use perfect basic strategy, which is far, far from true. The house gains when players split when they shouldn't and then double on those split hands. This rule deprives the casinos of that gain. By limiting doubling and reducing splitting -- fewer hands are worth splitting if doubling isn't possible -- casinos with this rule are taking the most exciting hands out of the game, reducing its entertainment value. And they are hurting the dealers they employ by depriving them of the the tips that result from big wins -- such as the ones that result when a dealer busts and a three- or four-way split with a double down or two pays off.

To the extent that I can take the casinos' perspective, hitting soft 17 makes sense. Not allowing doubling after split appears to be, at best, not worth the cost, and at worst totally self-defeating. And, fortunately, players still have plenty of opportunities to avoid these games. If this rule can't be eliminated, I believe it can be contained if players vote with their feet.