Simple math question regarding RoyalMatch -

zengrifter

zengrifter

Banned
Simple math question regarding RoyalMatch -

Grojean demonstrates that a simple red/black +1/-1 RC will posiEV for 3-10 RM at +7/-7. He notes that this imprecise approach will only glean about 50% of the opportunity.

Given the above, how many non-x suit must I see to have an advantage?

For example, if the first cards out of the deck contain no hearts, how many must I see - 5? 6? 8? 10?

Thanks! zg
 
T

The Mayor

Well-Known Member
Proposition

Give me 10% of your first year earnings playing the RM and I'll write you a sim in C++ to answer this question, and other questions you have, about RM.

--Mayor
 
T

The Mayor

Well-Known Member
The only way I know to answer it...

OK,

So you can see the delicacy of this, give specific parameters that you want an exact solution for. That means:

1) What are the exact payoffs for the RM game? (There are two flavors)
2) How many decks?
3) Exactly which cards are removed? (Not just "6 non-hearts" removed).

Then I will give you the exact answer for that situation.

--Mayor
 
zengrifter

zengrifter

Banned
**my responses-

1) What are the exact payoffs for the RM game? (There are two flavors)

**3/10 payoff

2) How many decks?

**1D

3) Exactly which cards are removed? (Not just "6 non-hearts" removed).

**5,6,7 etc. 'non-x' suit depletion off-the-top. How many depletion of non-x are required? zg
 
T

The Mayor

Well-Known Member
Not enough info

when you say non-x that is very complicated...

for example, if 5 non-hearts are gone, then those may be

KQ of S
KQ of D
K of C

for which there would be a considerably different closed result than if

2,3,4,5,6 of S were gone.

Do you see? You want a closed result, so you have to be exact! If you want a result that averages over all possible depleted cards, you need a simulation.

In the first case, the EV is even more negative. In the second case, it may be almost an even game (or better). So which do you want?

Sorry, even at this level, you are beyond the scope of a closed result. I could sim this easily, but I cannot and will not compute the separate result for each of the 576,000 (or so) cases that can occur by removing 5 random non-hearts and then average them.
 
zengrifter

zengrifter

Banned
Ok, I get your point...

"Do you see? You want a closed result, so you have to be exact! If you want a result that averages over all possible depleted cards, you need a simulation."
----------------

... but the "simple" question was to be a derivitive of Grossjean's sim:

+7/-7 RC red-black count = posiEV

Based on the above, letys go beyond red/black count - how many non-x suits must I see off the top? zg
 
zengrifter

zengrifter

Banned
Can't you just mentally extrapolate from...

...Grossjean's "+7/-7 red-black" parameter for a down and dirty estimate?
- zenignorance
 
T

The Mayor

Well-Known Member
No better than you

Again, the two ways I know how to do this are

1) An exact computation for an exact composition of the cards
2) A C++ program to simulate a specific situation in all the ways it arises.

You pick...

--Mayor
 
zengrifter

zengrifter

Banned
OK! - This one -

This one - "A C++ program to simulate a specific situation in all the ways it arises."

You get 10% of my lifetime RM earnings! Congradulations! zg
 
Automatic Monkey

Automatic Monkey

Banned
How about a BASIC program?

Hey ZG, if you want to work on sidebets it's easy enough to write your own routines with a little practice.

http://www.libertybasic.com It's a free download and you can become an expert in 2 weeks. I used it to analyze the Match the Dealer sidebet, turned out to be not worth it.

The first thing you do is generate an array to represent a shoe full of cards, then you randomize it, then flipping through the cards one at a time represents the deal. That's the heart of the program. Then you can modify it to keep count using whatever count you want, and check for a win on any sidebet you want. I can rework my Match the Dealer program for any count and any sidebet in about an hour.
 
T

The Mayor

Well-Known Member
Suited KQ

There are (52 choose 2) = 52*51/2 = 1326 possible pairs of cards you can be dealt from a deck for your first 2 cards. Out of these, 8 of them are suited KQ.

So, for the initial pair you are dealt, there are 1318 non-suited KQ and 8 Suited KQ, so the odds are 1318-to-8, or about 165-to-1. And they pay 10-to-1 for it.

Also, I didn't ask for 10% lifetime, only 10% of your first-year earnings from anything I produce. I trust you with this one.

Write me private email with any questions you have, I certainly won't post any information on beating side bets here, though I may use it!

--Mayor
 
Sonny

Sonny

Well-Known Member
How about for FREE!!!!!

Hey ZG, I've got just what you're looking for. It's an Excell spreadsheet that calculates the House/Player advantage for any given deck composition. You just type in how many of each card are in the deck(s) and it instantly shows you the % advantage/disadvantage - no sim required! You can input any number of decks and cards, and even remove one card at a time to see exactly how each card influences the bet.

I would be glad to email it to you if you want. Do you still use the Yahoo address that I used before?

-Sonny-

P.S. - The only stipulation would be that you have to share your results with me. =)
 
T

The Mayor

Well-Known Member
A little taste of research

ZG, I wrote a program, and essentially you can forget about the "non-x" idea. I haven't error checked my program, but it is in the range of seeing 18-20 non-x before you get an edge.

What is much more significant towards giving you an edge is "all-x". That is, if the first few cards are all the same suit (or nearly), your edge grows much more quickly.

Again, all of this needs to be simmed, along with many other ideas.
 
zengrifter

zengrifter

Banned
ahhh!

"What is much more significant towards giving you an edge is "all-x". That is, if the first few cards are all the same suit (or nearly), your edge grows much more quickly."
 
You must log in or register to reply here.
Top