In Blackjack, Taking Insurance Is Usually A Waste of Money

Casinos don't usually urge you to exploit situations where they're vulnerable. Like splitting sevens in blackjack when the dealer has a six showing or doubling on a 10 against a dealer five. The only bet they print right on the blackjack layout, that the dealer always asks if you want, is insurance. Nobody hypes insurance to do you a favor. It's high-profit for the house... except under certain circumstances.

What is it?

If the dealer draws ace-up, you can "take insurance" after the deal, before any hands are played. Bet up to half your original wager and win 2-to-1 if the dealer has blackjack.

Six situations can occur when you "hedge" a $10 bet with $5 insurance:

1) You and the dealer have blackjacks. You win $10 on insurance and tie your original bet, $10 net gain.

2) You have blackjack and the dealer doesn't. You lose $5 on insurance and win $15 on your original bet, $10 net gain.

3) You don't have blackjack but the dealer does. You win $10 on insurance and lose your original $10 bet, a push.

4) Neither you nor the dealer have blackjack and you win the hand. You lose the $5 insurance but win $10 on your original bet, $5 net gain.

5) Neither you nor the dealer have blackjack and you lose the hand. You lose the insurance and original bet, $15 net loss.

6) Neither you nor the dealer have blackjack and you tie the hand. You lose $5 on insurance and tie the original bet, $5 net loss.

Some players, viewing these possibilities, conclude insurance is a good bet because three win, two lose, and one pushes. Others like insurance when they have blackjack to lock up a win, rather than win one way and tie the other. Both analyses are what statisticians call "anecdotal" and the conclusions defy the laws of probability.

What's wrong with it?

Insurance, pure and simple, is a bet the hole card has a 10-value. Using insurance to protect other bets doesn't alter this fact. So, in evaluating insurance, nothing else matters.

If cards came from an infinite shoe, you'd average four 10-values for every nine others. This makes odds nine-to-four against the hole card being a 10. For most purposes, the six and eight deck shoes dealt at most blackjack tables might as well be infinite.

To simplify the math, say that in a year, you take $5 in insurance 1300 times. If the cards are distributed statistically, you'd win 400 of these bets, raking in 400 x $10 = $4000. But you'd lose 900 bets, giving up 900 x $5 = $4500. You'd be $500 behind.

Is it ever good?

Paid at 2-to-1, insurance would be a "fair" bet if for every four 10-value cards there were eight others. You'd have the edge if there were fewer than eight others.

Card counters occasionally find shoes with high enough proportions of 10-values to make insurance a good bet. Experts say this is one of the big benefits of counting, and typically recommend insurance when "true counts" exceed +3.

Some non-counters in games with over five hands being dealt take insurance when only a few of the cards exposed in the deal are 10-valued. The chance of a 10 in the hole is then above normal.

Sumner A Ingmark, whose amusing schmoozing many players are using to keep from losing, put it thus: