Even money Blackjack?

I just started "re" playing blackjack. I see now that when you have Blackjack and the dealer has an ace showing that you have the option of insurance or to take "even money" blackjack. When, if at all, would this be a wise decision? Any stats to go with it would be cool too. thanks.

The expected value (EV) for taking even money is 1.00000. The EV for not taking even money is 1.0384615, so that is the better bet. The reasoning and math is explained in the article Ask the Wizard: Blackjack - Basic Strategy (Specific Hands). There are two exceptions. If the money is a life-changing sum you could play it safe. If you are counting cards and know that the dealer's probability of getting blackjack has lowered the EV below 1.00000 then even money would be the best decision.

Edit: Added exception information.

For those who are NOT mathematically inclined, I generally offer this thought (with terms altered to fit the situation):

Which will make you more upset: Taking even money when the dealer didn't have BJ, or pushing when you could've had even money?

Quote: pork611

I just started "re" playing blackjack. I see now that when you have Blackjack and the dealer has an ace showing that you have the option of insurance or to take "even money" blackjack. When, if at all, would this be a wise decision? Any stats to go with it would be cool too. thanks.

Do you buy insurance on your non blackjack hands when the dealer shows an ace? Taking even money is just that - buying insurance. In the "old" days you actually had to put the money in the insurance circle when you had a blackjack versus the dealers ace if you wanted even money. The reason you don't now is to speed up the game as well as to plant that "better than nothing mentality" into the player's head.

Regardless of your hand, insurance is merely a side bet on whether the dealer's hole card is a 10. That's it. The dealer will average a 10 in the hole 4 out of 13 times or about 30% of the time. Card counters will buy insurance as dictated by the count and it is the most important of all the index plays.

The non counter will lose more money buying insurance and that includes taking even money. At many casinos the dealer is required to ask if a player wants even money because it's good for the house. Although they may mention it, they are not required to ask if you want to double your 11 against the dealer's 5 or surrender that 16 against the dealer's 10. There is a reason for that.

I'd also want to say that insuring a "good hand" - like a BJ or 20 - is worse than insuring a bad hand. Why? Because your chances are better for winning than with a bad hand. So don't take insurance unless you are counting.

Quote: BigJer

I'd also want to say that insuring a "good hand" - like a BJ or 20 - is worse than insuring a bad hand. Why? Because your chances are better for winning than with a bad hand. So don't take insurance unless you are counting.

Right you are, BigJer, counting only. This looks like a good time to tell this one since it happened just this morning. I insured my 16 against the dealer's ace not once but twice in my 45 minutes of play. There was no blackjack either time so I, of course, surrendered causing the dealer to think I was the biggest idiot in the whole place. Even the bus lady cheering me on from behind made a comment. I wouldn't have it any other way. :-)

Quote: BigJer

I'd also want to say that insuring a "good hand" - like a BJ or 20 - is worse than insuring a bad hand. Why? Because your chances are better for winning than with a bad hand. So don't take insurance unless you are counting.

That's not really true.

It is true that insuring a hand where you have 10s is worse than insuring a hand that doesn't have 10s. So insuring TT is worse than insuring a random hand. But insuring A9 is better than insuring a random hand, for the same reason.

If you want to consider variance and Kelly betting (and ignore hand composition) then insuring a good hand is slightly better than insuring a bad hand, but this is not a significant effect (unless you have a significant percentage of your bankroll at risk on the hand, which may be reasonable if you made the bet with additional information). Grosjean has a good article about this, including how much of your bankroll you should risk if you know that an ace is coming, given that you will play the hand sub-optimally in order to reduce variance (thus lowering your edge) It's not a trivial problem to solve, because as you lower variance you can bet more, but as you lower your edge you should bet less. In other words, there is not only an optimal Kelly bet, but also an optimal Kelly playing strategy :) It includes surrendering poor soft totals against strong dealer upcards.

I'm just talking in general. Without the composition dependent strategy.

Quote: BigJer

I'm just talking in general. Without the composition dependent strategy.

In that case why do you say it is "worse" to insure a good hand? If you ignore composition, count, and variance, then it is identical to insuring a bad hand.

If you take variance into account (and you want to lower it!) then it is better to insure a good hand than a bad one.

Quote: AxiomOfChoice

That's not really true.

It is true that insuring a hand where you have 10s is worse than insuring a hand that doesn't have 10s. So insuring TT is worse than insuring a random hand. But insuring A9 is better than insuring a random hand, for the same reason.

Well, a random 20 hand is 8 times more likely to be TT than to be A9. So if you want to be fair about "insuring 20 is worse than any other non-20 hand" is not that trivial to answer, it will depend on the deck size.

Quote: MangoJ

Well, a random 20 hand is 8 times more likely to be TT than to be A9. So if you want to be fair about "insuring 20 is worse than any other non-20 hand" is not that trivial to answer, it will depend on the deck size.

That is true, but he was not talking about composition when making his statement. I'm not entirely sure what he meant by the reason that he gave.