July 13th, 2016 at 7:34:21 PM
permalink

My question involves the existence of simulators for Spanish 21, or Super Fun 21, or Ultimate Fun 21 (the game on Carnival). Does anyone know if this exists? In Excel form, or otherwise? I would certainly be willing to pay for this, or pay for the creation of a simulator.

As background, I've been playing Ultimate Fun 21 on many Carnival Cruise ships for nearly a decade, and have consistently won money (too long and too consistently to be considered short-term luck). I'm analytical by nature myself, having run a quantitative hedge fund for several years in the past. I'd like to simulate the game if possible, running 5-10 million hands per night, and tweaking strategies slightly each time to analyze sensitivity. But I'm not a programmer and am looking for the technology if it is out there.

I know that the house is always supposed to win, but my gut and experience is telling me this game may be long-term beatable (at least the version they have on the ships). I've read all about the small HE on this site, but I'm experiencing a player edge over many thousands of hands, and want to simulate the game before I invest big money at the table.

Thank you in advance for any guidance.

As background, I've been playing Ultimate Fun 21 on many Carnival Cruise ships for nearly a decade, and have consistently won money (too long and too consistently to be considered short-term luck). I'm analytical by nature myself, having run a quantitative hedge fund for several years in the past. I'd like to simulate the game if possible, running 5-10 million hands per night, and tweaking strategies slightly each time to analyze sensitivity. But I'm not a programmer and am looking for the technology if it is out there.

I know that the house is always supposed to win, but my gut and experience is telling me this game may be long-term beatable (at least the version they have on the ships). I've read all about the small HE on this site, but I'm experiencing a player edge over many thousands of hands, and want to simulate the game before I invest big money at the table.

Thank you in advance for any guidance.

July 13th, 2016 at 10:59:40 PM
permalink

Hello, Carnivore, and welcome.

Could you list as many of the rules of the particular version you're asking about as possible?

How many decks

What cards are removed

What hands automatically win

What hands pay a premium, and how much for each

What player options are there, such as splits and doubles; specifically, what hands can be doubled on, what hands can be split and how many times, that sort of thing.

With those types of details, the math guys can give you a good idea of what it would take to build a sim, or how an existing sim might be modified to work within the game.

Not my area of expertise, so I may be asking the wrong questions, but I think you'll get better answers if you provide that info.

Could you list as many of the rules of the particular version you're asking about as possible?

How many decks

What cards are removed

What hands automatically win

What hands pay a premium, and how much for each

What player options are there, such as splits and doubles; specifically, what hands can be doubled on, what hands can be split and how many times, that sort of thing.

With those types of details, the math guys can give you a good idea of what it would take to build a sim, or how an existing sim might be modified to work within the game.

Not my area of expertise, so I may be asking the wrong questions, but I think you'll get better answers if you provide that info.

If the House lost every hand, they wouldn't deal the game.

July 14th, 2016 at 12:18:21 PM
permalink

Hi Carnivore, and welcome to the forums!

So you've played for a decade and have consistently won? When you say that, do you mean you pretty much win 'almost' every session? How many hours would you estimate you've played? Do you only play on cruises? How often do you go on cruises? If you play 4 hours per day, with an average 5 day cruise, and you cruise 3 times per year, that would only be 60 hours per year... resulting in 600 hours after a decade. The long term for blackjack is usually around 1,000 hours.

Do you count cards, use index plays, or any other form of advantage play?

If you're playing basic strategy, we need to know the exact rules of the game. How many decks do they deal from, where do they place the cut card (i.e. how many decks are cut off per shoe). Is this out of a continuous shuffler? Can you double on any 2 cards? How about any number of cards, etc? We need to know the EXACT rules to derive a specific house edge, and then apply any special plays that you do past basic strategy (if any exist). That will drive all of our numbers for your Expected Value (EV).

So you've played for a decade and have consistently won? When you say that, do you mean you pretty much win 'almost' every session? How many hours would you estimate you've played? Do you only play on cruises? How often do you go on cruises? If you play 4 hours per day, with an average 5 day cruise, and you cruise 3 times per year, that would only be 60 hours per year... resulting in 600 hours after a decade. The long term for blackjack is usually around 1,000 hours.

Do you count cards, use index plays, or any other form of advantage play?

If you're playing basic strategy, we need to know the exact rules of the game. How many decks do they deal from, where do they place the cut card (i.e. how many decks are cut off per shoe). Is this out of a continuous shuffler? Can you double on any 2 cards? How about any number of cards, etc? We need to know the EXACT rules to derive a specific house edge, and then apply any special plays that you do past basic strategy (if any exist). That will drive all of our numbers for your Expected Value (EV).

Playing it correctly means you've already won.

July 14th, 2016 at 1:29:06 PM
permalink

I've won on every cruise, which is 8 or 9. And I've only had a handful of losing sessions of a couple of hours length. Either I am the world's luckiest player, which I doubt, or this game is beatable. I've probably spent 300-350 hours playing the game if I had to guess. So maybe 8,000-10,000 hands.

I am a casual counter ... the reason I am asking the question is I may become a serious counter and raise my bet size significantly if it makes sense.

Here are the exact rules:

Kings are removed, so 48 card decks.

8 decks in the shoe.

All decks shuffled by hand (no machine).

Cut card has usually been between 75-80% of the way back.

Double on any two cards.

Player BJ gets paid immediately, always beats dealer BJ.

Player 21 always gets paid immediately, even if dealer hits 21 later.

Player BJ gets paid 3:2.

7-7-7 gets paid 3:2.

6-7-8 gets paid 3:2.

You can surrender, and I do fairly frequently.

You can split and resplit (and double on those splits) as much as you like.

There is double down rescue, so if you don't like your double card, you can withdraw your double and forfeit only the original wager.

There is the Lucky Ladies sucker bet, which I do not factor in, because the odds/payout is atrociously bad.

Minimum table bet is $6, max is $500.

Thank you in advance for any help! I'd love to plug these rules into a simulator and run it until I get several million hands played to see some trends. I'm a Kelly criterion guy, so if there is a positive return, I'll bet using Kelly.

I am a casual counter ... the reason I am asking the question is I may become a serious counter and raise my bet size significantly if it makes sense.

Here are the exact rules:

Kings are removed, so 48 card decks.

8 decks in the shoe.

All decks shuffled by hand (no machine).

Cut card has usually been between 75-80% of the way back.

Double on any two cards.

Player BJ gets paid immediately, always beats dealer BJ.

Player 21 always gets paid immediately, even if dealer hits 21 later.

Player BJ gets paid 3:2.

7-7-7 gets paid 3:2.

6-7-8 gets paid 3:2.

You can surrender, and I do fairly frequently.

You can split and resplit (and double on those splits) as much as you like.

There is double down rescue, so if you don't like your double card, you can withdraw your double and forfeit only the original wager.

There is the Lucky Ladies sucker bet, which I do not factor in, because the odds/payout is atrociously bad.

Minimum table bet is $6, max is $500.

Thank you in advance for any help! I'd love to plug these rules into a simulator and run it until I get several million hands played to see some trends. I'm a Kelly criterion guy, so if there is a positive return, I'll bet using Kelly.

July 14th, 2016 at 1:55:50 PM
permalink

This game sounds fairly similar to Spanish 21 (although the bonuses for 777 and 678 are different). Here are a few pages worth looking at as you may be able to estimate the House Edge with a little detective work and clarifying the exact rules:Quote:TheCarnivore123...You can surrender...

Blackjack variants ( https://wizardofodds.com/games/#blackjack ) including

http://wizardofodds.com/games/pontoon/australian/

http://wizardofodds.com/games/spanish-21/

http://wizardofodds.com/games/super-fun-21/

https://wizardofodds.com/games/blackjack/calculator/

https://wizardofodds.com/games/blackjack/basics/#toc-RuleVariations

Thus you should get an estimate for various rules - note Spanish deck also affects the House Edge but in Spanish 21 etc. this is offset by the bonuses.

Questions include

Early or Late surrender - this means whether you can opt to lose only half before the dealer draws any cards or the whole bet loses to a Dealer's Blackjack instead.

Player 21 beats Blackjack (i.e. no hole card dealt until all players have acted)

Can hands formed from split Aces hit and/or double.

You can see that Spanish 21 and some of the variants have a low House Edge so any additional advantage (say allowing Early surrender 0 .54% in regular BJ) would swing it in the player's favour.

July 14th, 2016 at 1:58:50 PM
permalink

Ok, now we're talkin!

I'm going to start with a regular base game of 8D, H17, DAS, DA2, Split to 4 hands, Late Surrender. The penetration you gave is about the average.

House Edge Starting Point = .57% (using the Wizards Blackjack House Edge Calculator).

1) Kings are removed... so the TC starts at -4, in the 8 deck shoe. Remember, 4x8 = 32 kings removed =/. This alone ups the HE 2%.

- House Edge = 2.57%

2) Player 21 automatic winner (blackjack and 21) is worth .54% for the player.

- House Edge = 2.03%

3) If 777 paid 2-1 automatically, that's only worth .03%, so paying 3-2 automatically, it's probably about .2% for the player.

- House Edge = 2.01%

4) 6-7-8 paid 3-2 automatically isn't listed, but I'd be more than willing to wager it's worth approximately .05% (given it will happen "very barely slightly" more than 777) for the player. Again, I believe I'm being generous with the amount this is helping the player. I'd always rather over-estimate than under.

- House Edge = 1.96%

5) Player May Double on Any Number of Cards? I think that was what you're going for? This is worth .15% for the player.

- House Edge = 1.81

6) Double Down Rescue is .10% for the player.

- House Edge = 1.71%

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

Are there any other rules such as 5 card charlie or an automatic 5 card winner?

At this point the house edge is more like a 6:5 game of blackjack. It's a pretty terrible game. You say you count "casually." I assume that means if you notice a bunch of small cards come out you bet a little bigger on the following couple rounds? If you're not using a proven system (such as Hi/Low) and you're not 100% proficient, then I would pretty much assume you're playing "close" (I hope) to basic strategy... Which would net you a 1.71% expected loss on your action.

Min is $6, Max is $500... Let's say your "average" bet is $25. Using your own estimate of 350 hours and upwards of 10,000 hands, let's run some simple numbers...

Expected Loss (EL) = TotalWagered * HouseEdge = (NumHands*AvgBet)*HouseEdge = (10,000*25)*(-1.71) = -$4275

However, you must also look at your standard deviations to see your range of variance (luck).

AvgBet = $25

AvgAdvantage = -1.71%

OriginalSD = 1.15*AvgBet = 28.75

SD(x hands) = Sqrt(NumHands) * OriginalSD

Thus, SD(10,000 hands) = Sqrt(10,000) * 28.75 = $2875... This is just for 1 Standard Deviation though, which comes with 68% confidence. If we want to say your EV +/- SD with 99% confidence we need to use 3SD. 3SD = $8625.

So, for the example above, with your given hours/hands... You could expect to LOSE $4,275 +/- $8625... So yes, it IS quite possible that you're UP money but still playing a losing game at this point.

If you want more finite numbers for "your game" then just let me know what you think your average bet is. Also let me know if there are any additional rules. Sorry to tell you that it doesn't seem like this is a winning game, and that you are in fact getting lucky =/. No one likes hearing that and I hope I don't sound like a dick saying that, it's just what the math says is the truth.

Oddly enough, the Lucky Ladies side bet is a beatable side bet and it sounds like the only beatable thing on this table lol. What's the minimum and maximum you can play on the lucky ladies side bet?

I'm going to start with a regular base game of 8D, H17, DAS, DA2, Split to 4 hands, Late Surrender. The penetration you gave is about the average.

House Edge Starting Point = .57% (using the Wizards Blackjack House Edge Calculator).

1) Kings are removed... so the TC starts at -4, in the 8 deck shoe. Remember, 4x8 = 32 kings removed =/. This alone ups the HE 2%.

- House Edge = 2.57%

2) Player 21 automatic winner (blackjack and 21) is worth .54% for the player.

- House Edge = 2.03%

3) If 777 paid 2-1 automatically, that's only worth .03%, so paying 3-2 automatically, it's probably about .2% for the player.

- House Edge = 2.01%

4) 6-7-8 paid 3-2 automatically isn't listed, but I'd be more than willing to wager it's worth approximately .05% (given it will happen "very barely slightly" more than 777) for the player. Again, I believe I'm being generous with the amount this is helping the player. I'd always rather over-estimate than under.

- House Edge = 1.96%

5) Player May Double on Any Number of Cards? I think that was what you're going for? This is worth .15% for the player.

- House Edge = 1.81

6) Double Down Rescue is .10% for the player.

- House Edge = 1.71%

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

Are there any other rules such as 5 card charlie or an automatic 5 card winner?

At this point the house edge is more like a 6:5 game of blackjack. It's a pretty terrible game. You say you count "casually." I assume that means if you notice a bunch of small cards come out you bet a little bigger on the following couple rounds? If you're not using a proven system (such as Hi/Low) and you're not 100% proficient, then I would pretty much assume you're playing "close" (I hope) to basic strategy... Which would net you a 1.71% expected loss on your action.

Min is $6, Max is $500... Let's say your "average" bet is $25. Using your own estimate of 350 hours and upwards of 10,000 hands, let's run some simple numbers...

Expected Loss (EL) = TotalWagered * HouseEdge = (NumHands*AvgBet)*HouseEdge = (10,000*25)*(-1.71) = -$4275

However, you must also look at your standard deviations to see your range of variance (luck).

AvgBet = $25

AvgAdvantage = -1.71%

OriginalSD = 1.15*AvgBet = 28.75

SD(x hands) = Sqrt(NumHands) * OriginalSD

Thus, SD(10,000 hands) = Sqrt(10,000) * 28.75 = $2875... This is just for 1 Standard Deviation though, which comes with 68% confidence. If we want to say your EV +/- SD with 99% confidence we need to use 3SD. 3SD = $8625.

So, for the example above, with your given hours/hands... You could expect to LOSE $4,275 +/- $8625... So yes, it IS quite possible that you're UP money but still playing a losing game at this point.

If you want more finite numbers for "your game" then just let me know what you think your average bet is. Also let me know if there are any additional rules. Sorry to tell you that it doesn't seem like this is a winning game, and that you are in fact getting lucky =/. No one likes hearing that and I hope I don't sound like a dick saying that, it's just what the math says is the truth.

Oddly enough, the Lucky Ladies side bet is a beatable side bet and it sounds like the only beatable thing on this table lol. What's the minimum and maximum you can play on the lucky ladies side bet?

Playing it correctly means you've already won.

July 14th, 2016 at 2:20:39 PM
permalink

My average bet is probably only about $10. I don't want to commit significant money to anything unless I know I can win at it. Willing to take the bet up to $100 for a winnable game.

For the Lucky Ladies, the minimum is $1 and the max is $10.

Also, I forgot to add Dealer must hit soft 17. And yes, the payout on a 5-card 21 is 3:2 also.

Some thoughts ... I am surprised that automatic 21 (BJ and 21) is only 0.54% for player. I would think it's a much bigger advantage than that. This payout is very common ... with seven players at the table, someone has it every other hand at least. Same thing with doubling after 3 or 4 cards ... I would think the edge is more than 0.15%, given the frequency.

This is why I would like the simulator, if it exists. I understand the concept of house edge, but I want to test it out empirically. I used to have a poker program play 10 million hands a night. I want to do the same thing with this game, if possible.

Thanks for the continued input!

For the Lucky Ladies, the minimum is $1 and the max is $10.

Also, I forgot to add Dealer must hit soft 17. And yes, the payout on a 5-card 21 is 3:2 also.

Some thoughts ... I am surprised that automatic 21 (BJ and 21) is only 0.54% for player. I would think it's a much bigger advantage than that. This payout is very common ... with seven players at the table, someone has it every other hand at least. Same thing with doubling after 3 or 4 cards ... I would think the edge is more than 0.15%, given the frequency.

This is why I would like the simulator, if it exists. I understand the concept of house edge, but I want to test it out empirically. I used to have a poker program play 10 million hands a night. I want to do the same thing with this game, if possible.

Thanks for the continued input!

July 14th, 2016 at 2:24:29 PM
permalink

One more thought ... Surrender isn't mentioned. I surrender A LOT ... to the point the other players at the table look at me funny. I am thinking this might be a part of the success? I always surrender 14, 15, 16 to an ace. I surrender 15 and 16 to a nine or a ten. On the other hand, I've doubled 12 vs. a dealer 6 with good success.

July 14th, 2016 at 3:11:15 PM
permalink

Assuming you're allowed Early surrender then I'm guessing it is correct to do so against an Ace - it certainly is with a full deck.

As to doubling 12 vs 6, this is usually a bad move so I am surprised you say this has "good success". Are you allowed to rescue one bet and only lose the other if you bust with a 10? Otherwise I can only imagine you were plain lucky.

As to doubling 12 vs 6, this is usually a bad move so I am surprised you say this has "good success". Are you allowed to rescue one bet and only lose the other if you bust with a 10? Otherwise I can only imagine you were plain lucky.

July 14th, 2016 at 4:41:43 PM
permalink

Quote:charliepatrickAssuming you're allowed Early surrender then I'm guessing it is correct to do so against an Ace - it certainly is with a full deck.

As to doubling 12 vs 6, this is usually a bad move so I am surprised you say this has "good success". Are you allowed to rescue one bet and only lose the other if you bust with a 10? Otherwise I can only imagine you were plain lucky.

Curious, Charlie; with 1/4 of the tens removed, the risk of busting on a doubled 12 has gone down some amount. Still not enough to make this a close EV choice to hitting, and better than standing?

If the House lost every hand, they wouldn't deal the game.