DigiDeal e-table "buster" side bet?
These machines offer two side bets. One is the "blackjack buster" side bet, which pays when the dealer busts. A 3 or 4 card bust pays 2x, up to 250x for an 8 card bust.
My question is whether anyone knows of any way to calculate the dealer bust frequency, or can direct me to any table showing percentages. I'm also curious to know how the count impacts the dealer's likelihood of busting.
Given that a low count means there are more low cards in the deck, it seems more likely that the dealer will draw more cards before busting. Acting on this theory, I up my bet on this side bet when the count is low. Thus far I've had good results betting like this, but I'm also aware it could just as easily be due to random variance rather than valid math. And, english major that I am, I have no clue how to evaluate this.
Anyone run across this before?
No counting info on that page, but I suspect a) you're right that a low count improves your odds, but b) not enough to overcome the edge. Even if it did overcome the edge, your main BJ bet is playing at a larger disadvantage, so I doubt you could gain an overall edge.
1) I'm pretty sure Buster has been analyzed for countability and found not practically beatable, and
2) If the e-table shuffles after every hand you can't count anyway. There's no good way to tell if an e-table is or isn't shuffling every hand unless it explicitly tells you. I think the DigiDeal tables are configurable to do either, though.
Quote: MathExtremistTwo more comments:
1) I'm pretty sure Buster has been analyzed for countability and found not practically beatable, and
2) If the e-table shuffles after every hand you can't count anyway. There's no good way to tell if an e-table is or isn't shuffling every hand unless it explicitly tells you. I think the DigiDeal tables are configurable to do either, though.
The tables I play don't shuffle after every hand - they are configured to have a "virtual shoe" that shows cards remaining. I'm almost embarassed to tell you the penetration, though. One establishment deals from six decks, shuffling after three. The other isn't much better, dealing from seven and shuffling after four.
Quote: ProfessorApeThe tables I play don't shuffle after every hand - they are configured to have a "virtual shoe" that shows cards remaining. I'm almost embarassed to tell you the penetration, though. One establishment deals from six decks, shuffling after three. The other isn't much better, dealing from seven and shuffling after four.
Bear in mind that such graphical treatment may not reflect the actual math. There's nothing preventing the game from displaying what looks like a virtual shoe and a cut card but, internally, shuffling a six-deck shoe every hand. It's a near-instantaneous process in software. I'm not saying this is what happens -- I don't actually know -- but one of the touted benefits of e-table BJ games is their invulnerability to card counting due to instant reshuffling. Not using that would seem odd: even with 50% penetration and even if you weren't spreading min/max, your hourly profit from counting a countable e-table game is at least 2x a physical table due to the rapidity of game play. And if the casino doesn't bother with heat on e-tables, or if they're fully-automatic (dealerless), then you could spread min-max and really clean up. Strange.
I just ran the EOR's on the 6D, H17 version. The EOR's were strange:Quote: MathExtremistTwo more comments:
1) I'm pretty sure Buster has been analyzed for countability and found not practically beatable
| A | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | T |
|---|---|---|---|---|---|---|---|---|---|
| 0.118% | -0.286% | -0.120% | -0.001% | 0.066% | -0.069% | 0.176% | 0.112% | 0.040% | -0.009% |
Note that a 5 makes every stiff hand for the dealer, but a 6 busts a 16, hence the difference between the EOR for a 5 and a 6. My opinion is that this bet is not significantly vulnerable to card counting. Certainly, the High-Low system would be highly ineffective. The counter needs to overcome a house edge of more than 6.2%, so it would be a long haul to get to a big enough true count to get an edge.
I didn't double-check this, so use it at your own risk.
Quote: MathExtremistBear in mind that such graphical treatment may not reflect the actual math. There's nothing preventing the game from displaying what looks like a virtual shoe and a cut card but, internally, shuffling a six-deck shoe every hand. It's a near-instantaneous process in software. I'm not saying this is what happens -- I don't actually know -- but one of the touted benefits of e-table BJ games is their invulnerability to card counting due to instant reshuffling. Not using that would seem odd: even with 50% penetration and even if you weren't spreading min/max, your hourly profit from counting a countable e-table game is at least 2x a physical table due to the rapidity of game play. And if the casino doesn't bother with heat on e-tables, or if they're fully-automatic (dealerless), then you could spread min-max and really clean up. Strange.
I've been suspicious of that as well - heck, I'm suspicious of anything that relies on a computer. If they are shuffling after every hand, though, they are opening themselves up to potential trouble. The dealers all say it shuffles after three or four decks (depending on which shop you're in). The "shoe" screen also states "Shuffle Pending" when the cut card comes up, and runs through a graphic animation of cards being shuffled.
Of course, I don't know how one could ever prove it. In my limited playing time and with my rudimentary counting, I've had some pretty good runs when the count is high but have nowhere near enough experience to draw any conclusions. I'd like to think it's my smart play instead of variance, but then again I have pretty good denial skills.
Quote: teliotI just ran the EOR's on the 6D, H17 version. The EOR's were strange:
A 2 3 4 5 6 7 8 9 T
0.118% -0.286% -0.120% -0.001% 0.066% -0.069% 0.176% 0.112% 0.040% -0.009%
Note that a 5 makes every stiff hand for the dealer, but a 6 busts a 16, hence the difference between the EOR for a 5 and a 6. My opinion is that this bet is not significantly vulnerable to card counting. Certainly, the High-Low system would be highly ineffective. The counter needs to overcome a house edge of more than 6.2%, so it would be a long haul to get to a big enough true count to get an edge.
I didn't double-check this, so use it at your own risk.
Sorry to be a newbie, but what doe this show, the effect of removal on the dealer's likelihood of going bust? I have a basic understanding of EOR, but my liberal arts brain doesn't get this one ...
This shows the net change in house edge by removing the selected card from the 6 deck shoe and then recomputing the house edge. So, removing an Ace shifts the edge 0.118% towards the player. Removing a deuces shifts the edge 0.286% towards the house. Etc. These sorts of computations form the basis for developing and evaluating card counting systems.Quote: ProfessorApeSorry to be a newbie, but what doe this show, the effect of removal on the dealer's likelihood of going bust? I have a basic understanding of EOR, but my liberal arts brain doesn't get this one ...
Seems odd - one might think the ten would have a bigger impact.
Thanks for your input, it's quite illuminating.
Quote: ProfessorApeSo this is based on the house edge of 6.2% on the side bet, right? If so, I think I understand. So, even if you were to remove all 24 aces from a 6 deck shoe, the house edge would still be around 3.5%?
Seems odd - one might think the ten would have a bigger impact.
Thanks for your input, it's quite illuminating.
Sort of -- the cumulative results of multiple card removals are not linear, but it's a rough approximation. And it's also about as much as our puny human brains can handle from a counting standpoint. Doing an EOR analysis based on combinations of multiple cards would produce a counting scheme too complicated for anyone to use.
Well said. For the non-linearity part, I get a house edge of 0.3097% dealing from a six-deck shoe devoid of aces.Quote: MathExtremistSort of -- the cumulative results of multiple card removals are not linear, but it's a rough approximation. And it's also about as much as our puny human brains can handle from a counting standpoint. Doing an EOR analysis based on combinations of multiple cards would produce a counting scheme too complicated for anyone to use.
Card Counting the Buster Blackjack Side Bet, 6 Decks
And this one on the two-deck version:
Card Counting the Buster Blackjack Side Bet, 2 Decks
