Online Blackjack question(s)

Hi,

------------------------------------------
there is a game of Online blackjack with the following rules:

6 Decks
Hole Card: No (Does not say if it is ENHC, but will assume it is this until I know otherwise)
Stand on all 17's
Surrender allowed: yes (except no surrender against A's)
Surrender type: Early
Double: Any two cards
Number of splits allowed: Once only
Hit to split A's: No
Double after split: yes

Game Limits (spreads): $5 - $100 and $10 - $200 (both spreads are 20 x table minimum)

and (if you can believe them) they have the rule below as well:

Games per shuffle: 20

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

The '20 games per shuffle' rule seems to be a good game for card counters (if true).

Question 1(a): How can you test the game to see if the above is true***?

True***: I understand that you can't prove it to be 100% (you can only get a 'guide' of how 'true' this rule is by doing tests, the more runs the better obviously)

Question 1(b): I plan to do 100 to 1000 tests, but how many should I do to be able to test that the '20 games per shuffle' is: 90% probable, 95% probable, and higher probabilities (99.9% probable or higher, if possible) ?

NB: I have only just started doing the tests (completed two so far), I plan to do 100 to 1000 tests.

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

On their website, they have the Theoretical RTP at ~98.9...%, yet when I put those parameters into a BJ analyser, I get a house edge of ~0.31% (or Theoretical RTP of ~ 99.69%).

Question 2: They are a fairly reputable sports-betting website, but can I trust their 'casino games section', if they seem to have the RTP calculation off by so much?

Update: There is one reason I can think of at the moment for the RTP discrepancy, maybe they posted the figures from their own 'Actual RTP figures instead of the 'Theoretical RTP' figures, (because the house edge of ~0.31% assumes correct basic strategy, and as we all know there are a lot of player's who play by 'gut decisions' or who don't know basic strategy, which will usually lower the 'actual RTP' when compared to the 'theoretical RTP')


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

See data from two tests below,

the table(s) show the cards that have been drawn from each rank and suit, (in other words, just do '6 minus the value shown', to work out how many cards are left in each rank and suit)

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

Test 1:

Card	Clubs	Diamonds	Hearts	Spades
A	2	2	2	3
2	0	2	2	4
3	1	1	0	1
4	2	1	2	3
5	1	2	1	3
6	1	3	2	2
7	3	3	2	2
8	2	2	3	2
9	2	2	5	1
10	3	2	3	4
J	2	2	2	3
Q	3	3	2	0
K	0	1	3	1

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

test 2:

Card	Clubs	Diamonds	Hearts	Spades
A	2	2	3	2
2	2	3	3	2
3	3	3	3	2
4	2	2	1	3
5	1	2	5	3
6	1	3	2	1
7	2	1	2	1
8	0	2	2	3
9	1	3	2	2
10	3	2	0	3
J	2	2	4	3
Q	1	3	3	4
K	1	1	1	1

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

Note 1: For test 1, the average cards per hand was 5.30 (106 cards were drawn from 20 hands)
Note 2: For test 2, the average cards per hand was 5.55 (111 cards were drawn)

Forgot to mention that the above game is: Single-handed (although most people would have eventually realised this from the 'average cards per hand' info)


They also have a 5-handed version, for this version the game rules are the same as the post above, except (see below)

8 Decks (instead of 6)

12 games per shuffle (instead of 20)

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

The figure below are just a guess (I will either update these figures /or/ completely remove this section of the post, when I know more /or/ if someone posts 'better figures/info' in regards to this section, respectively)

6 Deck version: ~4 decks left in the shoe (after the 19th game),

8 Deck version: ~4.7 decks left in the shoe (after the 11the game, playing 5 hands per game)

the above figures are just a guess^^^, but the 6 deck game looks better to me from a counting point of view, but I would be happy for any input on this


guess^^^, for my guess(es) i used the figures below:

Average Cards for Dealer per game: 3 cards

Average Cards for Player per game (per hand): 2.5 cards

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

again, I will update this area when I have better 'cards per game' figures (for dealer and player)

Quote: ksdjdj

Games per shuffle: 20...

The '20 games per shuffle' rule seems to be a good game for card counters (if true).

Hmmm, 20 rounds no matter the number of players sounds not 'too' bad. Figuring each player on average takes about 2.5 cards, this would be 5+dealer so 6*2.5 = 15 cards per round... 15x20 = 300 of the available 312 cards. Sounds like a great game!

Quote: ksdjdj

Forgot to mention that the above game is: Single-handed (although most people would have eventually realised this from the 'average cards per hand' info)

They also have a 5-handed version, for this version the game rules are the same as the post above, except (see below)

8 Decks (instead of 6)

12 games per shuffle (instead of 20)...

...aaaaaaaaaaaaaaaaaaannnnnnnnnnnndddddddddd the expected anomaly revealed. So we now know the original game is heads up, so what does this tell us:

Game 1
Average ~2.5 cards for you and dealer... 5 cards per round * 20 rounds = 100 cards of the 312 card shoe. This means they're cutting off 66% of the show, on average. This is UNGODLY AWFUL penetration and makes the game beyond NOT worth playing (especially with a 1-20 max spread). Penetration is actually THE MOST important factor in defining a playable/good game. The more penetration you have the more you'll see fluctuating counts and the more opportunities you'll have to bet more in favorable positions. Decent penetration for a regular 6 deck game is 1.5 decks being cut out or "not played." This yields only 25% of the shoe being cut off and is the bare minimum for most who scout games, i.e. they look for even less being cut off.

33% penetration is simply not playable, nor profitable (even with a 1-20 spread, which btw your variance would be out of this world and you'd need a $100,000 bankroll to play that $5 game).

Game 2
More of the same above, but let's look at some numbers at least. ~2.5 cards per player (and dealer) results in 6*2.5 = 15 cards per round * 12 rounds = 180 cards dealt out of the 416 cards in the shoe. This results in 43% PEN, with 57% of the shoe being cut off. This is again just awful and unplayable. You'll be sitting around for hours and hours waiting for a count that has any kind of positive EV to bet, then you wouldn't even have a spread/ramp, you'd more likely just go from min to max bet as soon as you got any kind of positive expectation. Again, this would drive your variance through the wall and require a bankroll much much larger than I'm sure you have (and if you did have $100k bankroll why on Earth wouldn't you just travel and play good live games).

Conclusions
Both games are not playable, nor profitable, unless you have a ridiculous bankroll and hours and hours to kill making below minimum wage per hour. Not to mention... Is this one of those "online live" dealer games? They could still disable your account, take your money, etc, if they see you spreading min to max (which you'd have to do). To follow that point up, you may say this is from a "reputable" site, but in my opinion there are NO online "reputable" sites I would trust 100%. Bovada is the closest thing I'd trust to being 'somewhat' reputable, and I still had a big issue with them 4-5 months ago. There's been many a horror story about online casinos refusing to pay winners, closing peoples accounts, taking bonuses back, etc, etc, etc... and the real problem is if they do, there's NOTHING you can do about it other than Name Shame them here. They're all hosted in a foreign country out of your reach, so just playing online alone is a 'gamble.'

It seems over the past month about once a week someone posts about some great 'online' game they find and ask "hey, why isn't this beatable?" Because most people don't understand just how important penetration is. They see S17, DAS, DA2, $5... this has to be beatable?!?! If you'd like to understand a bit more please feel free to check out my A-Z Counting Cards in Blackjack thread, where Penetration (PEN) is discussed as the most important factor in any blackjack game.

There's also an excellent idea by Eliot concerning the theoretical profit that could be gained by an accurate counter (usually it assumed the counter back bets $100 only when the circumstances are in the player's favour) and is measured by the profit obtained per 100 hands.

A profitable situation comes when there are more tens (or whatever, if it's a different game or side-bet) to overcome the inherent House Edge, and so is more likely to occur when the deck is getting smaller. In theory with (say) 33% penetration there could be advantageous situations, but it's not worth the effort. With 83% there certainly are, but again you need nerves of steel, a big bankroll and accurate counting.

I didn't do the calculation with 33% but have these figures from a previous investigation (UK rules, 6-decks, approx HE=0.48%)
(i) 83%: $27
(ii) 66% $18
(iii) CSM (16 cards) $2.

This is why online games can typically offer a lower House Edge as they can shuffle 6-decks every hand. They will also [probably] know, in this case, that while there is a profit possible, it is so small that no-one would bother.

Quote: charliepatrick

In theory with (say) 33% penetration there could be advantageous situations

This is all nonsense. Even if your patience is unlimited to wait for that advantageous situation, you can't sit out the game on all other situations.

Besides that, although it says "20 games per shuffle" it doesn't mean you can identify the shuffle point.

Quote: MangoJ

This is all nonsense. Even if your patience is unlimited to wait for that advantageous situation, you can't sit out the game on all other situations.

Besides that, although it says "20 games per shuffle" it doesn't mean you can identify the shuffle point.

You are right - it is a technical measure (I use it to help determine how much a [new] game [being designed] is vulnerable to counters). Also "20 games per shuffle" will be somewhere between 16 cards and 66%: the exact figure doesn't matter: you can make the broad assessment the profit gets lower as fewer cards are used per shoe before shuffling and, as you say, if you had to play $5 all the time, it is even more likely there is no practical profit to be garnished.