Next to a "16"

the worst hand you can be dealt is.....

a hard 17.

so why is it so many players "stand" on 17?

discuss.

If you simulate 100 Million hands of Blackjack, you will be better off by standing on 17 than you will by doing anything else.

I don't know about all games, but last night I found a $2 minimum 8D, H17, DA2, DAS, LS game. Since this was my first expierence with 8D and LS, I used the free tool at bjmath.com to look up the Basic Strategy. I was suprised to find that you're supposed to surrender hard 17 vs. A.

Towards the end of the night it was in a high count, around +6 KO, and I got a chance to try it out. I had my max bet out, which was only $4, and lost $2 on Insurance, then lost another $2 surrendering. Then he busted :P

Even though I lost 2 units on that hand, I'm still happy that I knew and made the right play.

Ray

but I'm afraid I can't agree that it was the "right play"
...to buy insurance
buying insurance is a natural born loser to begin with, but to
then surrender???%$!^^!??

umm, what book(s) have you been reading?

remaining deck is rich in 10 value cards. The insurance bet is only about whether the dealer has a blackjack, that is what the side bet of insurance is all about, nothing else. The decision to buy insurance has nothing to do with how large your bet on the hand is, or whether you have a bj or not. It is purely on the odds of the dealer having a blackjack. In fact, if you have a bj the dealer is actually less likely to have one.

For players using basic strategy only, then correct play is never buy insurance, it will be wrong play most of the time. You have to be an advantage player to know if the remaining deck is rich in high cards.

Knock-Out Blackjack is a book that teaches the KO card counting system.

It says to take insurance when the count is +3 or greater, reguardless of your hand. Since the count was +6, I feel confident that taking Insurance was the right play.

I don't know if Surrender Hard 17 vs. A has an associated playing index or not. I don't recall the KO book having insurance indexes. So I used the optimized basic strategy for that surrender. It was the right play considering the information I had. Until I hear somthing different from a reliable counting expert, I would make the same play again.

Ray

For example with the Uston+/- count you would surrender hard 17, below -4, with s17, and below +1, with h17. You don't mention which varriation of KO this was--which is why I say that unbalanced counts sneek up on you--or what IRC and Pivot you were using, or whether you were using TKO? Your best play was likely to play out the hand, in that you are dealing with late surrender, where you still lose the entire bet if the dealer has a natural, and where the surrender index use is conditional in the dealer not having a natural.

Then the proper play is to play the hand out, in that with a non ten in the hole, your chances of winning actually get better as the count increases. You have started knowing that the hole card is not a ten, but it can be ace thru 9. With a higher count, assuming you are using hi-lo or some similar count, there is a higher probability of tens and aces being in the remainder, including the unseen by you hole card,. So the card may be anything ace thru 9 with a usual 1/13 probability of 7 thru 9, (using hi-lo as an example; I used Uston +/- for the surrender number because I use that count), a lesser probability of 2 thru 6 (which are included in the hi-lo tagging), and a greater probability of the ace in the hole.

With the above your chances with a hard 17 actually increase with the the hi-lo true count going up! Even thought the basic strategy play is surrender with h17, you should not surrender 17 v. A when the count is high!

The key statement is that "you didn't have a KO 17 v. Ace surrender index." You lacked the tools to make the correct play, which was likely (not knowing your IRC, pivot and other KO variables for your statement on the KO count--see my other answer) take insurance and then stand!

Fifty %, care to explain it, with out STD ?

It (KO Book) says to take insurance when the count is +3 or greater, reguardless of your hand.

I don't recall the KO book having insurance indexes.

The first statement is the complete index for KO. Take insurance at RC> +3 no matter what. Many people think that you should insure your larger bets, but as has been said above, the term insurance is a mis-nomer. It is a side bet that pays 2:1 if the hole card under the dealer's ace up-card is a ten valued card.

Card counters do take insurance when they have their larger bets out, but not because the bet is large. They take insurance and have their big bets out for the same reason, a high count. The high count tells them first that they have the advantage due to lots of tens left in the undealt cards. Then, when the dealer gets the ace up card, the high count tells them that the insurance side bet is a good bet because of all those tens left in the undealt cards.

Stiff hands bust more easily on high counts, so unfortunately, frequently an advantage player may take insurance, lose the insurance bet, and then surrender. It sucks, but with the right act, can be great for cover.

-T-

or beaten on a very regular basis.

That makes sense to me. Since I still don't have the actual index though, I'll just stick with basic strategy for now.

I was reading the SBA page at sba21.com today and it says it can generate custom indexes like this, but didn't mention anything about unbalanced counts. Does SBA work with KO?

The only response by a professional player here giving real insight and accurate info on the 17 vs A and count dependence and of course ..... the post is busted.
What's new?
Now watch this post also get busted by a poser.
Once again, accurate info by pro players not welcome here because the posers don't want to lose the limelight and "expert" status

at least in later revisions. I personally stick to algebraic approximaton indexes. KO can be converted into an equivalent balanced count by subtracting 1/13 from each of its card values (I am stuborn about the old term), using this adjusted balanced count to find the surrender indexes, and then determining your average penetration (ie half the penetration in the game conditions you are aiming for, and then using this to estimate the running count for this balanced equivalent count at that point, and then adding back 4 points, for each deck up to that point, if it is necessary to use some program or method (algebraic approximation etc.) that requires a balanced count (style deliberate to emphacise just how intricate an unbalanced count can be: the price you pay for that initial no need to TC).

You are on your own at that point, knowing a KO running count index for a certain penetration, to convert it to your particular IRC and pivot count equivalents.

I used the Uston+/- as that counts the 7 and is the main count I use, but similar advice would come from hi-lo or KO, or TKO (true counted KO) users. You can find algebraic approximaton effects of removal for surrender on bjmath.com I recomend the original algebraic apprximation paper by Arnold Snyder (I hope rge still sells it) over the method outlined at bjmath.com, as Snyder's is much clearer. You can also use 6,7,8 Blackjack or Casino Verite for this. I would still recomend a look through the Snyder paper and the EORs, on bjmath.com and in Theory of Blackjack for the insight into indexes you will get from them. I also have a post on the beginer's board here (set the search date as far back as you can) on using Snyder's original method for multiparameter indexes, rather than the method given in TOB (an averaging method is used that may not be accurate for some counts and side counts) or on bjmath.com. Such side count methods are not needed but handy to know (just in case some new rule like Lucky Ladys comes up requiring a side count), and you might check with Igor for some review comments on my method by emailing him.

That said you should also not limit yourself to just one penetration point for your indexes either. Surrender is not as linear as other decisions. In the above I explained how you convert a true count index for an equivalent balanced count, into an unbalanced index for an average penetration. But the surrender index itself is not linear in the sense that true count indexes +1 TC point higher, and -1 TC point lower, are likely not going to have the same ev difference from the actual TC. Your simulator may find then some exceptions from usual assumptions where an unbalanced index may be better than a balanced one in general. So if at all possible try to get your simulator to work with the unbalanced index directly and compare.

I know that is a complex procedure, but by comparing the direct unbalanced count inputed index directly, with the index found the long way around, above, you can also setup your surrender indexes to need the least amount of pivot and IRC adjustment for different penetrations. Where the final index is little different, long or direct method, you have an index where an unbalanced count is slightly superior, and needs less pivot and IRC adjustment, where the imbalance of the unbalanced count, matches the imbalance, or non-linearity of the decision itself.

Simpler can be better--even though I am not an umbalanced count fan--but getting to that simpler situation can be a bumpy ride.

For now, I'll just assume that SBA would give me a unbalanced KO index, without me having to consider the cards in the discard tray. Since I'm still new to counting and still haven't even memorized all the standard I18 plays from the book, this issue will have to wait. Great response though.

Thanks,

Ray

should include the penetration assumptions--how many cards are presumed left when the dealer shuffles--for the basic indexes. That is all you need for running count use, other than following the IRC and pivot customizing advice. The extra work I outlined (and I said it was rough) was to advise on how you might check over the running count surrender indexes and find where some unbalanced count running count indexes actually work better with an unbalanced count than any true count indexes.