Expected return spliting 10,10

[Baserken]

I know the correct decision in 10,10 vs any card is usually “Stand”, of course.

But, as a curiosity, I have a doubt with the player’s expected return in the case 10,10 vs 6 in case you split the 10s because I’ve found strange data for me. In this page:
http://wizardofodds.com/blackjack/appendix9-6ds17r4.html
indicates only 0.175 much less for example than 0.404 for 9,9 vs 6 when I think It should be more.

Really, in all cases 10,10 vs every card of the dealer the data don’t suit what I suppose. Am I reasoning something wrong? Thanks in advance.

 

[winr_winr_chicken_dinner!]

I would think it would be really close to splitting 9,9 myself, as you aren't playing your hand anymore so much as playing with an expectation of the dealer busting...

 

[FLASH1296]

• When you split 10's you never get to double.
  
• When you split 9's you can catch an ACE.
  
• If you do not see fit to split your 9's you are stuck with a sub-par 18.
  
• If you stand with your 20 you are remain a favorite no matter what.

 

[Brock Windsor]

Data error is my guess

I know the correct decision in 10,10 vs any card is usually “Stand”, of course.

But, as a curiosity, I have a doubt with the player’s expected return in the case 10,10 vs 6 in case you split the 10s because I’ve found strange data for me. In this page:
http://wizardofodds.com/blackjack/appendix9-6ds17r4.html
indicates only 0.175 much less for example than 0.404 for 9,9 vs 6 when I think It should be more.

Really, in all cases 10,10 vs every card of the dealer the data don’t suit what I suppose. Am I reasoning something wrong? Thanks in advance.

Even wizards can have output errors. Splitting 10's against a 6 should have an EV in the neighbourhood of 0.57+ I think. Not sure how that 0.175 got in there. Perhaps a more enlightened soul than I will take a look (Sonny?), but his 10 split numbers look wrong all the way through. Maybe it's not factoring the multiple hands properly?
-BW

 

[Sonny]

Based on Cacarulo's numbers (Archive copy), splitting tens vs. 6 should be somewhere between 0.32 and 0.44 depending on the number of decks and the rules. Nice catch Baserken!

-Sonny-

 

[Baserken]

Thanks everybody

Even Cacarulos data seems to me too much low. As Brock Windsor says (+0.57) my approach to the 10,10 vs 6 split probablity is: you receive one card in each 10, and then as you will never HIT OR double in each hand, the probability should be more or less the double of HITTING IN the case of 2 cards with total 10 (2,8 or 3,7 or 4,6) vs 6, more or less 0,29 x 2 = +0.58.

For example, with 9,9 vs 6 with no DAS split rule, you have in Wizardofodds an expected return of +0.40, more or less the double of HITTING IN the case of 2 cards with total 9, (+0.20, same data in Cacarulo and Wizardofodds). With DAS split allowed the return is a little bit more, +0.46 (same data in Cacarulo and Wizardofodds) because you DOUBLE in same case (receiving a 2).

I know two different sources wrong in the same point sounds too much. Is it possible?

 

[kewljason]

• When you split 10's you never get to double.
  
• When you split 9's you can catch an ACE.
  
• If you stand with your 20 you are remain a favorite no matter what.

Your first point is a good one Flash. With splitting 10's, you never get a chance to double. With splitting 9's you do. However, you mentioned 'what if you catch an ace'. Are you going to double on A9 vs 6?? I guess if the count is high enough, but the real advantage is what if you catch a 2 and can then double??? Anyway thats with 9's. In my opinion splitting 10's is a horse of a whole differnet color. (possible purple)

If you choose to break up a 20, which by definition of this game, is a pretty strong hand, it had better be at a very, very positive count.

My way of thinking is that if the count is positive enough to split 10's, then I already have my max bet out. And if I have my max bet out and get a 20 vs the dealer 6 in a game that by definition, 20, 21 wins, I am pretty damn good with that. Yes, simulation will show that at certain, rarely acheived high counts, it is better to split, but in real life, those counts happen so very rarely (yes this is redundant) that in most cases it seems you turn a sure winning hand into a split or possible double loss, which at max bet takes a while to overcome. Add to that the extra scutiny from splitting 10's and "to me" it's a play just not worth it.

Now A9 is even more of a gamble, IMO because there is no possiblity of a split. You are turning an almost assurred winning hand into either a double win or double loss, with no possiblility of a split. Now again, I know that simulations will show that there is a number that this becomes worthwile, but in the world of real play, give me an A9 with my max bet out and I'm good.

 

[Brock Windsor]

Thanks everybody

Even Cacarulos data seems to me too much low. As Brock Windsor says (+0.57) my approach to the 10,10 vs 6 split probablity is: you receive one card in each 10, and then as you will never HIT OR double in each hand, the probability should be more or less the double of HITTING IN the case of 2 cards with total 10 (2,8 or 3,7 or 4,6) vs 6, more or less 0,29 x 2 = +0.58.

For example, with 9,9 vs 6 with no DAS split rule, you have in Wizardofodds an expected return of +0.40, more or less the double of HITTING IN the case of 2 cards with total 9, (+0.20, same data in Cacarulo and Wizardofodds). With DAS split allowed the return is a little bit more, +0.46 (same data in Cacarulo and Wizardofodds) because you DOUBLE in same case (receiving a 2).

I know two different sources wrong in the same point sounds too much. Is it possible?

Cacarulos numbers most likely presume if a person drew an additional 10 (often) they would continue to split up to 4 times (thus further weakening the expectation). The numbers could also more accurately take into account the effect of removal. I am inclined to believe Cacarulos numbers are closest to exact. Shall we set an over/under on how many days until Dr. Shackleford's appendix9 more closely reflects Cacarulos results? The date today is March 19, 2010.
-Brock

 

[Cardcounter]

splitting 9's vs a 6 is not a close call at all it is a must split. You could catch a 2 on your first card and if the rules allow double down with an 11 vs a 6 but even if the rules don't allow a double down after a split it is still a mandatory split and the math isn't even close on the play. You will win far more often than you will lose the hand and will win both of your hands more than the one 18 you already made. But there will be those times when the dealer flips over a 5 under there 6 hits a 10 makes a 21 and beats both hands. a 6 up does not mean that the dealer will bust just that they have the best chance of busting with that card.

 

[Blue Efficacy]

Now here's another question, what if you reach the index for doubling soft 21, assuming the store you are in allows it? In my experience about half do.

Once when I was splitting 10's, despite all of the commotion fellow player made over it, the dealer practically insisted I doubled when I got an ace. I didn't. The next card was incidentally a ten, of course. Thankfully I have good enough cover to get away with splitting 10's at that specific store, as they will often, but not always, yell it out loudly and wait for a critter to acknowledge it.

They have also called out the (basic strategy) act of doubling soft 19 vs 6, but this one has only been done by dual rates.

Nevertheless, splitting nines is very powerful also. 4/13 you will get 19, a winner, 1/13 you will get 20, a bigger winner, and in the lower probability (depending on number of decks) that you get another 18, often a loser, most places let you split again a couple more times.

A question, how much is lost by splitting 9's vs 7 in neutral counts?

 

[Brock Windsor]

A question, how much is lost by splitting 9's vs 7 in neutral counts?

Will you accept "very little" as an answer? 8 decks you lose the least, 1 deck you lose the most but for the rarity of the play and the fact that it is close you won't notice the difference in your bottom line. Snyder suggests splitting 9's against an ace as cheap camo that is effective.
BW

 

[Blue Efficacy]

Will you accept "very little" as an answer? 8 decks you lose the least, 1 deck you lose the most but for the rarity of the play and the fact that it is close you won't notice the difference in your bottom line. Snyder even suggests always splitting 9's against every up card as cheap camo that is effective.
BW

Does this include against a face card? Against an ace is a cheap camo play, only 1 card gives dealer a dealt 20. but against face card, fully 1/3 will be 20. I am no expert but this seems like it wouldn't be so cheap.

 

[Renzey]

How much is lost by splitting 9/9 vs. 7? Will you accept "very little" as an answer? 8 decks you lose the least, 1 deck you lose the most but for the rarity of the play and the fact that it is close you won't notice the difference in your bottom line. Snyder even suggests always splitting 9's against every up card as cheap camo that is effective.
BW

If you stand with 9/9 vs. 7 (shoe game), you win the hand 70% of the time (counting pushes as a half win and half loss). If you split, you win on each new hand roughly 59% of the time. The loss in EV is about 3.7% -- no small margin since most good camo plays lose 1.5% or less.

Ironically, if somebody else is toying with the idea of splitting 9/9 vs. 7, your jumping in on it with pick you up a +18% EV.

 

[Brock Windsor]

Does this include against a face card? Against an ace is a cheap camo play, only 1 card gives dealer a dealt 20. but against face card, fully 1/3 will be 20. I am no expert but this seems like it wouldn't be so cheap.

Sorry I misquoted. It was splitting against the ace that was cheap camo.
-BW