Orson: Side bets question

[Orson]

I was getting a figure 28.15%, so the house still has a tiny edge, but I am not sure how my software treats hands when dealer does not need to complete his hand (i.e. when all players busts or got BJs). Obviously, with this bet out dealer needs to complete every hand. It seems to me that SBA doesn't have a function for dealer bust probability depending on count, but I will check more carefully.

Regards,
Orson

[Zenfighter]

Asterix dixit:

?Like the man who wants a dog who?s shaggy, but not too shaggy, the ten counter is
also difficult to satisfy. He wants the deck to be rich in tens, but not too rich. Some
authors, who try to explain why an abundance of tens favors the player, state that the dealer will bust more stiffs with ten rich decks. This is true, but only up to a point.
The dealer?s probability of busting, as a function of ten density, appears to maximize
(.295) with about 41% tens in the deck. This compares with a normal .286 and a .185
with no tens at all.?

De Belli Gallici 10:147

These are obviously SD?s figures. From my archives and out of a 10 billion hands sampling with 6 decks I?ve this:

Dealer bust: 28.1820 (0.002%)

At the 99% confidence level you have then:

28.1820 +/- 2.58*.002 = between [28.17684 and 28.18716]

Let?s be optimistic here and take the upper boundary. So what?

1/ .2818716 equals 1 in 3.54771463 tries, which means obviously that you?re
playing at an average disadvantage of -0.05% aprox. if you insist in betting always this option.

Now you have to take Prof. Griffin?s figures and make an inference of the adequate
ones for the 6 deck game. After this and by elementary math interpolation you will
get the density of tens you?ll need in order to have a profitable bet here.

If you want a no brainer use the pivot for Thorp?s Ts vs. others, aka Insurance count.

pivot = n*4, that?s 24 for 6 decks.

Removing 24 others off the top of a 6 deck game yields:

96/ 288 = .3333 density of tens, the main reason why these counters buy insurance here.

I?m frankly pessimistic with your side bets here, the advantages if any, seems to be
meager at best. I?ll be mistaken, naturally.

More clear now?

Sincerely
Z

[Orson]

It seems that there is no way to gain a significant edge on that bet. At least it can be used sometimes for camouflage, because it costs almost nothing.

Thank you,
Orson

[Cacarulo]

> I was getting a figure 28.15%, so the house
> still has a tiny edge, but I am not sure how
> my software treats hands when dealer does
> not need to complete his hand (i.e. when all
> players busts or got BJs). Obviously, with
> this bet out dealer needs to complete every
> hand. It seems to me that SBA doesn't have a
> function for dealer bust probability
> depending on count, but I will check more
> carefully.

I did a little programming and found a solution which is in agreement with Zenfighter's answer:

Hi-Lo is not suited for this bet as you will see in the following sim:

 

  TC     Bust% 

  -------------- 

  10     27.076% 

   9     27.277% 

   8     27.521% 

   7     27.671% 

   6     27.799% 

   5     27.939% 

   4     28.037% 

   3     28.107% 

   2     28.161% 

   1     28.203% 

   0     28.227% 

  -1     28.243% 

  -2     28.255% <== Max 

  -3     28.255% 

  -4     28.244% 

  -5     28.204% 

  -6     28.145% 

  -7     28.084% 

  -8     28.048% 

  -9     27.911% 

 -10     27.753% 

 

Overall: 28.189% 

As you can see the point where the dealer busts the most is at -2 (floored). This is not enough to overcome 1/3.5=0.2857.

Now, if you use a Ten Count (1 1 1 1 1 1 1 1 1 -2) in RC mode starting at -24 (6 decks) you'll get the following distribution:

 

  RC     Bust% 

  -------------- 

  14     29.101% 

  13     29.052% 

  12     29.030% 

  11     28.979% 

  10     28.996% 

   9     28.991% 

   8     28.913% 

   7     28.907% 

   6     28.860% 

   5     28.838% 

   4     28.817% 

   3     28.761% 

   2     28.722% 

   1     28.685% 

   0     28.635% 

  -1     28.606% <== Index 

  -2     28.547% 

  -3     28.500% 

  -4     28.480% 

  -5     28.433% 

  -6     28.395% 

  -7     28.372% 

  -8     28.308% 

  -9     28.277% 

 -10     28.254% 

 -11     28.205% 

 -12     28.181% 

 -13     28.158% 

 -14     28.125% 

 -15     28.092% 

 -16     28.085% 

 -17     28.060% 

 -18     28.055% 

 -19     28.026% 

 -20     28.014% 

 -21     27.990% 

 -22     27.992% 

 -23     27.963% 

 

Overall: 28.192% 

So, if you bet at -1 or more you will get an advantage.

Hope this helps.

Sincerely,
Cacarulo

[Zenfighter]

Absolutely outstanding piece of work!
I always knew that Don's Domain subscription fee,
was really at a bargain price! :-)

Congrats

Sincerely
Z

[Orson]